diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
index d2d28048efb2..72730e9ca06c 100644
--- a/drivers/gpu/drm/nouveau/Kconfig
+++ b/drivers/gpu/drm/nouveau/Kconfig
@@ -10,6 +10,7 @@ config DRM_NOUVEAU
 	select FB
 	select FRAMEBUFFER_CONSOLE if !EMBEDDED
 	select FB_BACKLIGHT if DRM_NOUVEAU_BACKLIGHT
+	select ACPI_VIDEO if ACPI
 	help
 	  Choose this option for open-source nVidia support.
 
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index e9b06e4ef2a2..23fa82d667d6 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -9,7 +9,8 @@ nouveau-y := nouveau_drv.o nouveau_state.o nouveau_channel.o nouveau_mem.o \
              nouveau_bo.o nouveau_fence.o nouveau_gem.o nouveau_ttm.o \
              nouveau_hw.o nouveau_calc.o nouveau_bios.o nouveau_i2c.o \
              nouveau_display.o nouveau_connector.o nouveau_fbcon.o \
-             nouveau_dp.o \
+             nouveau_dp.o nouveau_ramht.o \
+	     nouveau_pm.o nouveau_volt.o nouveau_perf.o nouveau_temp.o \
              nv04_timer.o \
              nv04_mc.o nv40_mc.o nv50_mc.o \
              nv04_fb.o nv10_fb.o nv30_fb.o nv40_fb.o nv50_fb.o nvc0_fb.o \
@@ -23,7 +24,8 @@ nouveau-y := nouveau_drv.o nouveau_state.o nouveau_channel.o nouveau_mem.o \
              nv04_dac.o nv04_dfp.o nv04_tv.o nv17_tv.o nv17_tv_modes.o \
              nv04_crtc.o nv04_display.o nv04_cursor.o nv04_fbcon.o \
              nv10_gpio.o nv50_gpio.o \
-	     nv50_calc.o
+	     nv50_calc.o \
+	     nv04_pm.o nv50_pm.o nva3_pm.o
 
 nouveau-$(CONFIG_DRM_NOUVEAU_DEBUG) += nouveau_debugfs.o
 nouveau-$(CONFIG_COMPAT) += nouveau_ioc32.o
diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c
index c17a055ee3e5..119152606e4c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_acpi.c
+++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c
@@ -292,6 +292,6 @@ nouveau_acpi_edid(struct drm_device *dev, struct drm_connector *connector)
 	if (ret < 0)
 		return ret;
 
-	nv_connector->edid = edid;
+	nv_connector->edid = kmemdup(edid, EDID_LENGTH, GFP_KERNEL);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 974b0f8ae048..53f4eba65cb9 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -43,9 +43,6 @@
 #define BIOSLOG(sip, fmt, arg...) NV_DEBUG(sip->dev, fmt, ##arg)
 #define LOG_OLD_VALUE(x)
 
-#define ROM16(x) le16_to_cpu(*(uint16_t *)&(x))
-#define ROM32(x) le32_to_cpu(*(uint32_t *)&(x))
-
 struct init_exec {
 	bool execute;
 	bool repeat;
@@ -272,12 +269,6 @@ struct init_tbl_entry {
 	int (*handler)(struct nvbios *, uint16_t, struct init_exec *);
 };
 
-struct bit_entry {
-	uint8_t id[2];
-	uint16_t length;
-	uint16_t offset;
-};
-
 static int parse_init_table(struct nvbios *, unsigned int, struct init_exec *);
 
 #define MACRO_INDEX_SIZE	2
@@ -1231,7 +1222,7 @@ init_dp_condition(struct nvbios *bios, uint16_t offset, struct init_exec *iexec)
 			return 3;
 		}
 
-		if (cond & 1)
+		if (!(cond & 1))
 			iexec->execute = false;
 	}
 		break;
@@ -4675,6 +4666,92 @@ int run_tmds_table(struct drm_device *dev, struct dcb_entry *dcbent, int head, i
 	return 0;
 }
 
+struct pll_mapping {
+	u8  type;
+	u32 reg;
+};
+
+static struct pll_mapping nv04_pll_mapping[] = {
+	{ PLL_CORE  , NV_PRAMDAC_NVPLL_COEFF },
+	{ PLL_MEMORY, NV_PRAMDAC_MPLL_COEFF },
+	{ PLL_VPLL0 , NV_PRAMDAC_VPLL_COEFF },
+	{ PLL_VPLL1 , NV_RAMDAC_VPLL2 },
+	{}
+};
+
+static struct pll_mapping nv40_pll_mapping[] = {
+	{ PLL_CORE  , 0x004000 },
+	{ PLL_MEMORY, 0x004020 },
+	{ PLL_VPLL0 , NV_PRAMDAC_VPLL_COEFF },
+	{ PLL_VPLL1 , NV_RAMDAC_VPLL2 },
+	{}
+};
+
+static struct pll_mapping nv50_pll_mapping[] = {
+	{ PLL_CORE  , 0x004028 },
+	{ PLL_SHADER, 0x004020 },
+	{ PLL_UNK03 , 0x004000 },
+	{ PLL_MEMORY, 0x004008 },
+	{ PLL_UNK40 , 0x00e810 },
+	{ PLL_UNK41 , 0x00e818 },
+	{ PLL_UNK42 , 0x00e824 },
+	{ PLL_VPLL0 , 0x614100 },
+	{ PLL_VPLL1 , 0x614900 },
+	{}
+};
+
+static struct pll_mapping nv84_pll_mapping[] = {
+	{ PLL_CORE  , 0x004028 },
+	{ PLL_SHADER, 0x004020 },
+	{ PLL_MEMORY, 0x004008 },
+	{ PLL_UNK05 , 0x004030 },
+	{ PLL_UNK41 , 0x00e818 },
+	{ PLL_VPLL0 , 0x614100 },
+	{ PLL_VPLL1 , 0x614900 },
+	{}
+};
+
+u32
+get_pll_register(struct drm_device *dev, enum pll_types type)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct pll_mapping *map;
+	int i;
+
+	if (dev_priv->card_type < NV_40)
+		map = nv04_pll_mapping;
+	else
+	if (dev_priv->card_type < NV_50)
+		map = nv40_pll_mapping;
+	else {
+		u8 *plim = &bios->data[bios->pll_limit_tbl_ptr];
+
+		if (plim[0] >= 0x30) {
+			u8 *entry = plim + plim[1];
+			for (i = 0; i < plim[3]; i++, entry += plim[2]) {
+				if (entry[0] == type)
+					return ROM32(entry[3]);
+			}
+
+			return 0;
+		}
+
+		if (dev_priv->chipset == 0x50)
+			map = nv50_pll_mapping;
+		else
+			map = nv84_pll_mapping;
+	}
+
+	while (map->reg) {
+		if (map->type == type)
+			return map->reg;
+		map++;
+	}
+
+	return 0;
+}
+
 int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims *pll_lim)
 {
 	/*
@@ -4750,6 +4827,17 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims
 	/* initialize all members to zero */
 	memset(pll_lim, 0, sizeof(struct pll_lims));
 
+	/* if we were passed a type rather than a register, figure
+	 * out the register and store it
+	 */
+	if (limit_match > PLL_MAX)
+		pll_lim->reg = limit_match;
+	else {
+		pll_lim->reg = get_pll_register(dev, limit_match);
+		if (!pll_lim->reg)
+			return -ENOENT;
+	}
+
 	if (pll_lim_ver == 0x10 || pll_lim_ver == 0x11) {
 		uint8_t *pll_rec = &bios->data[bios->pll_limit_tbl_ptr + headerlen + recordlen * pllindex];
 
@@ -4785,7 +4873,6 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims
 		pll_lim->max_usable_log2p = 0x6;
 	} else if (pll_lim_ver == 0x20 || pll_lim_ver == 0x21) {
 		uint16_t plloffs = bios->pll_limit_tbl_ptr + headerlen;
-		uint32_t reg = 0; /* default match */
 		uint8_t *pll_rec;
 		int i;
 
@@ -4797,37 +4884,22 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims
 			NV_WARN(dev, "Default PLL limit entry has non-zero "
 				       "register field\n");
 
-		if (limit_match > MAX_PLL_TYPES)
-			/* we've been passed a reg as the match */
-			reg = limit_match;
-		else /* limit match is a pll type */
-			for (i = 1; i < entries && !reg; i++) {
-				uint32_t cmpreg = ROM32(bios->data[plloffs + recordlen * i]);
-
-				if (limit_match == NVPLL &&
-				    (cmpreg == NV_PRAMDAC_NVPLL_COEFF || cmpreg == 0x4000))
-					reg = cmpreg;
-				if (limit_match == MPLL &&
-				    (cmpreg == NV_PRAMDAC_MPLL_COEFF || cmpreg == 0x4020))
-					reg = cmpreg;
-				if (limit_match == VPLL1 &&
-				    (cmpreg == NV_PRAMDAC_VPLL_COEFF || cmpreg == 0x4010))
-					reg = cmpreg;
-				if (limit_match == VPLL2 &&
-				    (cmpreg == NV_RAMDAC_VPLL2 || cmpreg == 0x4018))
-					reg = cmpreg;
-			}
-
 		for (i = 1; i < entries; i++)
-			if (ROM32(bios->data[plloffs + recordlen * i]) == reg) {
+			if (ROM32(bios->data[plloffs + recordlen * i]) == pll_lim->reg) {
 				pllindex = i;
 				break;
 			}
 
+		if ((dev_priv->card_type >= NV_50) && (pllindex == 0)) {
+			NV_ERROR(dev, "Register 0x%08x not found in PLL "
+				 "limits table", pll_lim->reg);
+			return -ENOENT;
+		}
+
 		pll_rec = &bios->data[plloffs + recordlen * pllindex];
 
 		BIOSLOG(bios, "Loading PLL limits for reg 0x%08x\n",
-			pllindex ? reg : 0);
+			pllindex ? pll_lim->reg : 0);
 
 		/*
 		 * Frequencies are stored in tables in MHz, kHz are more
@@ -4877,8 +4949,8 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims
 		if (cv == 0x51 && !pll_lim->refclk) {
 			uint32_t sel_clk = bios_rd32(bios, NV_PRAMDAC_SEL_CLK);
 
-			if (((limit_match == NV_PRAMDAC_VPLL_COEFF || limit_match == VPLL1) && sel_clk & 0x20) ||
-			    ((limit_match == NV_RAMDAC_VPLL2 || limit_match == VPLL2) && sel_clk & 0x80)) {
+			if ((pll_lim->reg == NV_PRAMDAC_VPLL_COEFF && sel_clk & 0x20) ||
+			    (pll_lim->reg == NV_RAMDAC_VPLL2 && sel_clk & 0x80)) {
 				if (bios_idxprt_rd(bios, NV_CIO_CRX__COLOR, NV_CIO_CRE_CHIP_ID_INDEX) < 0xa3)
 					pll_lim->refclk = 200000;
 				else
@@ -4891,10 +4963,10 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims
 		int i;
 
 		BIOSLOG(bios, "Loading PLL limits for register 0x%08x\n",
-			limit_match);
+			pll_lim->reg);
 
 		for (i = 0; i < entries; i++, entry += recordlen) {
-			if (ROM32(entry[3]) == limit_match) {
+			if (ROM32(entry[3]) == pll_lim->reg) {
 				record = &bios->data[ROM16(entry[1])];
 				break;
 			}
@@ -4902,7 +4974,7 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims
 
 		if (!record) {
 			NV_ERROR(dev, "Register 0x%08x not found in PLL "
-				 "limits table", limit_match);
+				 "limits table", pll_lim->reg);
 			return -ENOENT;
 		}
 
@@ -4931,10 +5003,10 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims
 		int i;
 
 		BIOSLOG(bios, "Loading PLL limits for register 0x%08x\n",
-			limit_match);
+			pll_lim->reg);
 
 		for (i = 0; i < entries; i++, entry += recordlen) {
-			if (ROM32(entry[3]) == limit_match) {
+			if (ROM32(entry[3]) == pll_lim->reg) {
 				record = &bios->data[ROM16(entry[1])];
 				break;
 			}
@@ -4942,7 +5014,7 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims
 
 		if (!record) {
 			NV_ERROR(dev, "Register 0x%08x not found in PLL "
-				 "limits table", limit_match);
+				 "limits table", pll_lim->reg);
 			return -ENOENT;
 		}
 
@@ -5293,7 +5365,7 @@ parse_bit_M_tbl_entry(struct drm_device *dev, struct nvbios *bios,
 	if (bitentry->length < 0x5)
 		return 0;
 
-	if (bitentry->id[1] < 2) {
+	if (bitentry->version < 2) {
 		bios->ram_restrict_group_count = bios->data[bitentry->offset + 2];
 		bios->ram_restrict_tbl_ptr = ROM16(bios->data[bitentry->offset + 3]);
 	} else {
@@ -5403,27 +5475,40 @@ struct bit_table {
 
 #define BIT_TABLE(id, funcid) ((struct bit_table){ id, parse_bit_##funcid##_tbl_entry })
 
+int
+bit_table(struct drm_device *dev, u8 id, struct bit_entry *bit)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	u8 entries, *entry;
+
+	entries = bios->data[bios->offset + 10];
+	entry   = &bios->data[bios->offset + 12];
+	while (entries--) {
+		if (entry[0] == id) {
+			bit->id = entry[0];
+			bit->version = entry[1];
+			bit->length = ROM16(entry[2]);
+			bit->offset = ROM16(entry[4]);
+			bit->data = ROMPTR(bios, entry[4]);
+			return 0;
+		}
+
+		entry += bios->data[bios->offset + 9];
+	}
+
+	return -ENOENT;
+}
+
 static int
 parse_bit_table(struct nvbios *bios, const uint16_t bitoffset,
 		struct bit_table *table)
 {
 	struct drm_device *dev = bios->dev;
-	uint8_t maxentries = bios->data[bitoffset + 4];
-	int i, offset;
 	struct bit_entry bitentry;
 
-	for (i = 0, offset = bitoffset + 6; i < maxentries; i++, offset += 6) {
-		bitentry.id[0] = bios->data[offset];
-
-		if (bitentry.id[0] != table->id)
-			continue;
-
-		bitentry.id[1] = bios->data[offset + 1];
-		bitentry.length = ROM16(bios->data[offset + 2]);
-		bitentry.offset = ROM16(bios->data[offset + 4]);
-
+	if (bit_table(dev, table->id, &bitentry) == 0)
 		return table->parse_fn(dev, bios, &bitentry);
-	}
 
 	NV_INFO(dev, "BIT table '%c' not found\n", table->id);
 	return -ENOSYS;
@@ -5683,8 +5768,14 @@ static uint16_t findstr(uint8_t *data, int n, const uint8_t *str, int len)
 static struct dcb_gpio_entry *
 new_gpio_entry(struct nvbios *bios)
 {
+	struct drm_device *dev = bios->dev;
 	struct dcb_gpio_table *gpio = &bios->dcb.gpio;
 
+	if (gpio->entries >= DCB_MAX_NUM_GPIO_ENTRIES) {
+		NV_ERROR(dev, "exceeded maximum number of gpio entries!!\n");
+		return NULL;
+	}
+
 	return &gpio->entry[gpio->entries++];
 }
 
@@ -5705,114 +5796,91 @@ nouveau_bios_gpio_entry(struct drm_device *dev, enum dcb_gpio_tag tag)
 	return NULL;
 }
 
-static void
-parse_dcb30_gpio_entry(struct nvbios *bios, uint16_t offset)
-{
-	struct dcb_gpio_entry *gpio;
-	uint16_t ent = ROM16(bios->data[offset]);
-	uint8_t line = ent & 0x1f,
-		tag = ent >> 5 & 0x3f,
-		flags = ent >> 11 & 0x1f;
-
-	if (tag == 0x3f)
-		return;
-
-	gpio = new_gpio_entry(bios);
-
-	gpio->tag = tag;
-	gpio->line = line;
-	gpio->invert = flags != 4;
-	gpio->entry = ent;
-}
-
-static void
-parse_dcb40_gpio_entry(struct nvbios *bios, uint16_t offset)
-{
-	uint32_t entry = ROM32(bios->data[offset]);
-	struct dcb_gpio_entry *gpio;
-
-	if ((entry & 0x0000ff00) == 0x0000ff00)
-		return;
-
-	gpio = new_gpio_entry(bios);
-	gpio->tag = (entry & 0x0000ff00) >> 8;
-	gpio->line = (entry & 0x0000001f) >> 0;
-	gpio->state_default = (entry & 0x01000000) >> 24;
-	gpio->state[0] = (entry & 0x18000000) >> 27;
-	gpio->state[1] = (entry & 0x60000000) >> 29;
-	gpio->entry = entry;
-}
-
 static void
 parse_dcb_gpio_table(struct nvbios *bios)
 {
 	struct drm_device *dev = bios->dev;
-	uint16_t gpio_table_ptr = bios->dcb.gpio_table_ptr;
-	uint8_t *gpio_table = &bios->data[gpio_table_ptr];
-	int header_len = gpio_table[1],
-	    entries = gpio_table[2],
-	    entry_len = gpio_table[3];
-	void (*parse_entry)(struct nvbios *, uint16_t) = NULL;
+	struct dcb_gpio_entry *e;
+	u8 headerlen, entries, recordlen;
+	u8 *dcb, *gpio = NULL, *entry;
 	int i;
 
-	if (bios->dcb.version >= 0x40) {
-		if (gpio_table_ptr && entry_len != 4) {
-			NV_WARN(dev, "Invalid DCB GPIO table entry length.\n");
-			return;
-		}
+	dcb = ROMPTR(bios, bios->data[0x36]);
+	if (dcb[0] >= 0x30) {
+		gpio = ROMPTR(bios, dcb[10]);
+		if (!gpio)
+			goto no_table;
 
-		parse_entry = parse_dcb40_gpio_entry;
+		headerlen = gpio[1];
+		entries   = gpio[2];
+		recordlen = gpio[3];
+	} else
+	if (dcb[0] >= 0x22 && dcb[-1] >= 0x13) {
+		gpio = ROMPTR(bios, dcb[-15]);
+		if (!gpio)
+			goto no_table;
 
-	} else if (bios->dcb.version >= 0x30) {
-		if (gpio_table_ptr && entry_len != 2) {
-			NV_WARN(dev, "Invalid DCB GPIO table entry length.\n");
-			return;
-		}
-
-		parse_entry = parse_dcb30_gpio_entry;
-
-	} else if (bios->dcb.version >= 0x22) {
-		/*
-		 * DCBs older than v3.0 don't really have a GPIO
-		 * table, instead they keep some GPIO info at fixed
-		 * locations.
-		 */
-		uint16_t dcbptr = ROM16(bios->data[0x36]);
-		uint8_t *tvdac_gpio = &bios->data[dcbptr - 5];
+		headerlen = 3;
+		entries   = gpio[2];
+		recordlen = gpio[1];
+	} else
+	if (dcb[0] >= 0x22) {
+		/* No GPIO table present, parse the TVDAC GPIO data. */
+		uint8_t *tvdac_gpio = &dcb[-5];
 
 		if (tvdac_gpio[0] & 1) {
-			struct dcb_gpio_entry *gpio = new_gpio_entry(bios);
-
-			gpio->tag = DCB_GPIO_TVDAC0;
-			gpio->line = tvdac_gpio[1] >> 4;
-			gpio->invert = tvdac_gpio[0] & 2;
+			e = new_gpio_entry(bios);
+			e->tag = DCB_GPIO_TVDAC0;
+			e->line = tvdac_gpio[1] >> 4;
+			e->invert = tvdac_gpio[0] & 2;
 		}
+
+		goto no_table;
 	} else {
-		/*
-		 * No systematic way to store GPIO info on pre-v2.2
-		 * DCBs, try to match the PCI device IDs.
-		 */
+		NV_DEBUG(dev, "no/unknown gpio table on DCB 0x%02x\n", dcb[0]);
+		goto no_table;
+	}
 
-		/* Apple iMac G4 NV18 */
-		if (nv_match_device(dev, 0x0189, 0x10de, 0x0010)) {
-			struct dcb_gpio_entry *gpio = new_gpio_entry(bios);
+	entry = gpio + headerlen;
+	for (i = 0; i < entries; i++, entry += recordlen) {
+		e = new_gpio_entry(bios);
+		if (!e)
+			break;
 
-			gpio->tag = DCB_GPIO_TVDAC0;
-			gpio->line = 4;
+		if (gpio[0] < 0x40) {
+			e->entry = ROM16(entry[0]);
+			e->tag = (e->entry & 0x07e0) >> 5;
+			if (e->tag == 0x3f) {
+				bios->dcb.gpio.entries--;
+				continue;
+			}
+
+			e->line = (e->entry & 0x001f);
+			e->invert = ((e->entry & 0xf800) >> 11) != 4;
+		} else {
+			e->entry = ROM32(entry[0]);
+			e->tag = (e->entry & 0x0000ff00) >> 8;
+			if (e->tag == 0xff) {
+				bios->dcb.gpio.entries--;
+				continue;
+			}
+
+			e->line = (e->entry & 0x0000001f) >> 0;
+			e->state_default = (e->entry & 0x01000000) >> 24;
+			e->state[0] = (e->entry & 0x18000000) >> 27;
+			e->state[1] = (e->entry & 0x60000000) >> 29;
 		}
-
 	}
 
-	if (!gpio_table_ptr)
-		return;
-
-	if (entries > DCB_MAX_NUM_GPIO_ENTRIES) {
-		NV_WARN(dev, "Too many entries in the DCB GPIO table.\n");
-		entries = DCB_MAX_NUM_GPIO_ENTRIES;
+no_table:
+	/* Apple iMac G4 NV18 */
+	if (nv_match_device(dev, 0x0189, 0x10de, 0x0010)) {
+		e = new_gpio_entry(bios);
+		if (e) {
+			e->tag = DCB_GPIO_TVDAC0;
+			e->line = 4;
+		}
 	}
-
-	for (i = 0; i < entries; i++)
-		parse_entry(bios, gpio_table_ptr + header_len + entry_len * i);
 }
 
 struct dcb_connector_table_entry *
@@ -6680,6 +6748,8 @@ static int nouveau_parse_vbios_struct(struct drm_device *dev)
 					bit_signature, sizeof(bit_signature));
 	if (offset) {
 		NV_TRACE(dev, "BIT BIOS found\n");
+		bios->type = NVBIOS_BIT;
+		bios->offset = offset;
 		return parse_bit_structure(bios, offset + 6);
 	}
 
@@ -6687,6 +6757,8 @@ static int nouveau_parse_vbios_struct(struct drm_device *dev)
 					bmp_signature, sizeof(bmp_signature));
 	if (offset) {
 		NV_TRACE(dev, "BMP BIOS found\n");
+		bios->type = NVBIOS_BMP;
+		bios->offset = offset;
 		return parse_bmp_structure(dev, bios, offset);
 	}
 
@@ -6806,6 +6878,8 @@ nouveau_bios_init(struct drm_device *dev)
 			"running VBIOS init tables.\n");
 		bios->execute = true;
 	}
+	if (nouveau_force_post)
+		bios->execute = true;
 
 	ret = nouveau_run_vbios_init(dev);
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.h b/drivers/gpu/drm/nouveau/nouveau_bios.h
index c1de2f3fcb0e..50a648e01c49 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.h
@@ -34,6 +34,20 @@
 
 #define DCB_LOC_ON_CHIP 0
 
+#define ROM16(x) le16_to_cpu(*(uint16_t *)&(x))
+#define ROM32(x) le32_to_cpu(*(uint32_t *)&(x))
+#define ROMPTR(bios, x) (ROM16(x) ? &(bios)->data[ROM16(x)] : NULL)
+
+struct bit_entry {
+	uint8_t  id;
+	uint8_t  version;
+	uint16_t length;
+	uint16_t offset;
+	uint8_t *data;
+};
+
+int bit_table(struct drm_device *, u8 id, struct bit_entry *);
+
 struct dcb_i2c_entry {
 	uint32_t entry;
 	uint8_t port_type;
@@ -170,16 +184,28 @@ enum LVDS_script {
 	LVDS_PANEL_OFF
 };
 
-/* changing these requires matching changes to reg tables in nv_get_clock */
-#define MAX_PLL_TYPES	4
+/* these match types in pll limits table version 0x40,
+ * nouveau uses them on all chipsets internally where a
+ * specific pll needs to be referenced, but the exact
+ * register isn't known.
+ */
 enum pll_types {
-	NVPLL,
-	MPLL,
-	VPLL1,
-	VPLL2
+	PLL_CORE   = 0x01,
+	PLL_SHADER = 0x02,
+	PLL_UNK03  = 0x03,
+	PLL_MEMORY = 0x04,
+	PLL_UNK05  = 0x05,
+	PLL_UNK40  = 0x40,
+	PLL_UNK41  = 0x41,
+	PLL_UNK42  = 0x42,
+	PLL_VPLL0  = 0x80,
+	PLL_VPLL1  = 0x81,
+	PLL_MAX    = 0xff
 };
 
 struct pll_lims {
+	u32 reg;
+
 	struct {
 		int minfreq;
 		int maxfreq;
@@ -212,6 +238,11 @@ struct pll_lims {
 
 struct nvbios {
 	struct drm_device *dev;
+	enum {
+		NVBIOS_BMP,
+		NVBIOS_BIT
+	} type;
+	uint16_t offset;
 
 	uint8_t chip_version;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index f6f44779d82f..80353e2b8409 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -36,21 +36,6 @@
 #include <linux/log2.h>
 #include <linux/slab.h>
 
-int
-nouveau_bo_sync_gpu(struct nouveau_bo *nvbo, struct nouveau_channel *chan)
-{
-	struct nouveau_fence *prev_fence = nvbo->bo.sync_obj;
-	int ret;
-
-	if (!prev_fence || nouveau_fence_channel(prev_fence) == chan)
-		return 0;
-
-	spin_lock(&nvbo->bo.lock);
-	ret = ttm_bo_wait(&nvbo->bo, false, false, false);
-	spin_unlock(&nvbo->bo.lock);
-	return ret;
-}
-
 static void
 nouveau_bo_del_ttm(struct ttm_buffer_object *bo)
 {
@@ -58,8 +43,6 @@ nouveau_bo_del_ttm(struct ttm_buffer_object *bo)
 	struct drm_device *dev = dev_priv->dev;
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 
-	ttm_bo_kunmap(&nvbo->kmap);
-
 	if (unlikely(nvbo->gem))
 		DRM_ERROR("bo %p still attached to GEM object\n", bo);
 
@@ -164,8 +147,6 @@ nouveau_bo_new(struct drm_device *dev, struct nouveau_channel *chan,
 	nouveau_bo_fixup_align(dev, tile_mode, tile_flags, &align, &size);
 	align >>= PAGE_SHIFT;
 
-	nvbo->placement.fpfn = 0;
-	nvbo->placement.lpfn = mappable ? dev_priv->fb_mappable_pages : 0;
 	nouveau_bo_placement_set(nvbo, flags, 0);
 
 	nvbo->channel = chan;
@@ -305,7 +286,8 @@ nouveau_bo_map(struct nouveau_bo *nvbo)
 void
 nouveau_bo_unmap(struct nouveau_bo *nvbo)
 {
-	ttm_bo_kunmap(&nvbo->kmap);
+	if (nvbo)
+		ttm_bo_kunmap(&nvbo->kmap);
 }
 
 u16
@@ -399,14 +381,19 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 		man->default_caching = TTM_PL_FLAG_CACHED;
 		break;
 	case TTM_PL_VRAM:
+		man->func = &ttm_bo_manager_func;
 		man->flags = TTM_MEMTYPE_FLAG_FIXED |
 			     TTM_MEMTYPE_FLAG_MAPPABLE;
 		man->available_caching = TTM_PL_FLAG_UNCACHED |
 					 TTM_PL_FLAG_WC;
 		man->default_caching = TTM_PL_FLAG_WC;
-		man->gpu_offset = dev_priv->vm_vram_base;
+		if (dev_priv->card_type == NV_50)
+			man->gpu_offset = 0x40000000;
+		else
+			man->gpu_offset = 0;
 		break;
 	case TTM_PL_TT:
+		man->func = &ttm_bo_manager_func;
 		switch (dev_priv->gart_info.type) {
 		case NOUVEAU_GART_AGP:
 			man->flags = TTM_MEMTYPE_FLAG_MAPPABLE;
@@ -469,19 +456,26 @@ nouveau_bo_move_accel_cleanup(struct nouveau_channel *chan,
 	if (ret)
 		return ret;
 
-	ret = ttm_bo_move_accel_cleanup(&nvbo->bo, fence, NULL,
-					evict || (nvbo->channel &&
-						  nvbo->channel != chan),
+	if (nvbo->channel) {
+		ret = nouveau_fence_sync(fence, nvbo->channel);
+		if (ret)
+			goto out;
+	}
+
+	ret = ttm_bo_move_accel_cleanup(&nvbo->bo, fence, NULL, evict,
 					no_wait_reserve, no_wait_gpu, new_mem);
+out:
 	nouveau_fence_unref((void *)&fence);
 	return ret;
 }
 
 static inline uint32_t
-nouveau_bo_mem_ctxdma(struct nouveau_bo *nvbo, struct nouveau_channel *chan,
-		      struct ttm_mem_reg *mem)
+nouveau_bo_mem_ctxdma(struct ttm_buffer_object *bo,
+		      struct nouveau_channel *chan, struct ttm_mem_reg *mem)
 {
-	if (chan == nouveau_bdev(nvbo->bo.bdev)->channel) {
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
+
+	if (nvbo->no_vm) {
 		if (mem->mem_type == TTM_PL_TT)
 			return NvDmaGART;
 		return NvDmaVRAM;
@@ -493,86 +487,181 @@ nouveau_bo_mem_ctxdma(struct nouveau_bo *nvbo, struct nouveau_channel *chan,
 }
 
 static int
-nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
-		     bool no_wait_reserve, bool no_wait_gpu,
-		     struct ttm_mem_reg *new_mem)
+nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
+		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
 {
-	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev);
-	struct ttm_mem_reg *old_mem = &bo->mem;
-	struct nouveau_channel *chan;
-	uint64_t src_offset, dst_offset;
-	uint32_t page_count;
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
+	u64 length = (new_mem->num_pages << PAGE_SHIFT);
+	u64 src_offset, dst_offset;
 	int ret;
 
-	chan = nvbo->channel;
-	if (!chan || nvbo->tile_flags || nvbo->no_vm)
-		chan = dev_priv->channel;
-
-	src_offset = old_mem->mm_node->start << PAGE_SHIFT;
-	dst_offset = new_mem->mm_node->start << PAGE_SHIFT;
-	if (chan != dev_priv->channel) {
-		if (old_mem->mem_type == TTM_PL_TT)
-			src_offset += dev_priv->vm_gart_base;
-		else
+	src_offset = old_mem->start << PAGE_SHIFT;
+	dst_offset = new_mem->start << PAGE_SHIFT;
+	if (!nvbo->no_vm) {
+		if (old_mem->mem_type == TTM_PL_VRAM)
 			src_offset += dev_priv->vm_vram_base;
-
-		if (new_mem->mem_type == TTM_PL_TT)
-			dst_offset += dev_priv->vm_gart_base;
 		else
+			src_offset += dev_priv->vm_gart_base;
+
+		if (new_mem->mem_type == TTM_PL_VRAM)
 			dst_offset += dev_priv->vm_vram_base;
+		else
+			dst_offset += dev_priv->vm_gart_base;
 	}
 
 	ret = RING_SPACE(chan, 3);
 	if (ret)
 		return ret;
-	BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_DMA_SOURCE, 2);
-	OUT_RING(chan, nouveau_bo_mem_ctxdma(nvbo, chan, old_mem));
-	OUT_RING(chan, nouveau_bo_mem_ctxdma(nvbo, chan, new_mem));
 
-	if (dev_priv->card_type >= NV_50) {
-		ret = RING_SPACE(chan, 4);
+	BEGIN_RING(chan, NvSubM2MF, 0x0184, 2);
+	OUT_RING  (chan, nouveau_bo_mem_ctxdma(bo, chan, old_mem));
+	OUT_RING  (chan, nouveau_bo_mem_ctxdma(bo, chan, new_mem));
+
+	while (length) {
+		u32 amount, stride, height;
+
+		amount  = min(length, (u64)(4 * 1024 * 1024));
+		stride  = 16 * 4;
+		height  = amount / stride;
+
+		if (new_mem->mem_type == TTM_PL_VRAM && nvbo->tile_flags) {
+			ret = RING_SPACE(chan, 8);
+			if (ret)
+				return ret;
+
+			BEGIN_RING(chan, NvSubM2MF, 0x0200, 7);
+			OUT_RING  (chan, 0);
+			OUT_RING  (chan, 0);
+			OUT_RING  (chan, stride);
+			OUT_RING  (chan, height);
+			OUT_RING  (chan, 1);
+			OUT_RING  (chan, 0);
+			OUT_RING  (chan, 0);
+		} else {
+			ret = RING_SPACE(chan, 2);
+			if (ret)
+				return ret;
+
+			BEGIN_RING(chan, NvSubM2MF, 0x0200, 1);
+			OUT_RING  (chan, 1);
+		}
+		if (old_mem->mem_type == TTM_PL_VRAM && nvbo->tile_flags) {
+			ret = RING_SPACE(chan, 8);
+			if (ret)
+				return ret;
+
+			BEGIN_RING(chan, NvSubM2MF, 0x021c, 7);
+			OUT_RING  (chan, 0);
+			OUT_RING  (chan, 0);
+			OUT_RING  (chan, stride);
+			OUT_RING  (chan, height);
+			OUT_RING  (chan, 1);
+			OUT_RING  (chan, 0);
+			OUT_RING  (chan, 0);
+		} else {
+			ret = RING_SPACE(chan, 2);
+			if (ret)
+				return ret;
+
+			BEGIN_RING(chan, NvSubM2MF, 0x021c, 1);
+			OUT_RING  (chan, 1);
+		}
+
+		ret = RING_SPACE(chan, 14);
 		if (ret)
 			return ret;
-		BEGIN_RING(chan, NvSubM2MF, 0x0200, 1);
-		OUT_RING(chan, 1);
-		BEGIN_RING(chan, NvSubM2MF, 0x021c, 1);
-		OUT_RING(chan, 1);
+
+		BEGIN_RING(chan, NvSubM2MF, 0x0238, 2);
+		OUT_RING  (chan, upper_32_bits(src_offset));
+		OUT_RING  (chan, upper_32_bits(dst_offset));
+		BEGIN_RING(chan, NvSubM2MF, 0x030c, 8);
+		OUT_RING  (chan, lower_32_bits(src_offset));
+		OUT_RING  (chan, lower_32_bits(dst_offset));
+		OUT_RING  (chan, stride);
+		OUT_RING  (chan, stride);
+		OUT_RING  (chan, stride);
+		OUT_RING  (chan, height);
+		OUT_RING  (chan, 0x00000101);
+		OUT_RING  (chan, 0x00000000);
+		BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_NOP, 1);
+		OUT_RING  (chan, 0);
+
+		length -= amount;
+		src_offset += amount;
+		dst_offset += amount;
 	}
 
+	return 0;
+}
+
+static int
+nv04_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
+		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
+{
+	u32 src_offset = old_mem->start << PAGE_SHIFT;
+	u32 dst_offset = new_mem->start << PAGE_SHIFT;
+	u32 page_count = new_mem->num_pages;
+	int ret;
+
+	ret = RING_SPACE(chan, 3);
+	if (ret)
+		return ret;
+
+	BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_DMA_SOURCE, 2);
+	OUT_RING  (chan, nouveau_bo_mem_ctxdma(bo, chan, old_mem));
+	OUT_RING  (chan, nouveau_bo_mem_ctxdma(bo, chan, new_mem));
+
 	page_count = new_mem->num_pages;
 	while (page_count) {
 		int line_count = (page_count > 2047) ? 2047 : page_count;
 
-		if (dev_priv->card_type >= NV_50) {
-			ret = RING_SPACE(chan, 3);
-			if (ret)
-				return ret;
-			BEGIN_RING(chan, NvSubM2MF, 0x0238, 2);
-			OUT_RING(chan, upper_32_bits(src_offset));
-			OUT_RING(chan, upper_32_bits(dst_offset));
-		}
 		ret = RING_SPACE(chan, 11);
 		if (ret)
 			return ret;
+
 		BEGIN_RING(chan, NvSubM2MF,
 				 NV_MEMORY_TO_MEMORY_FORMAT_OFFSET_IN, 8);
-		OUT_RING(chan, lower_32_bits(src_offset));
-		OUT_RING(chan, lower_32_bits(dst_offset));
-		OUT_RING(chan, PAGE_SIZE); /* src_pitch */
-		OUT_RING(chan, PAGE_SIZE); /* dst_pitch */
-		OUT_RING(chan, PAGE_SIZE); /* line_length */
-		OUT_RING(chan, line_count);
-		OUT_RING(chan, (1<<8)|(1<<0));
-		OUT_RING(chan, 0);
+		OUT_RING  (chan, src_offset);
+		OUT_RING  (chan, dst_offset);
+		OUT_RING  (chan, PAGE_SIZE); /* src_pitch */
+		OUT_RING  (chan, PAGE_SIZE); /* dst_pitch */
+		OUT_RING  (chan, PAGE_SIZE); /* line_length */
+		OUT_RING  (chan, line_count);
+		OUT_RING  (chan, 0x00000101);
+		OUT_RING  (chan, 0x00000000);
 		BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_NOP, 1);
-		OUT_RING(chan, 0);
+		OUT_RING  (chan, 0);
 
 		page_count -= line_count;
 		src_offset += (PAGE_SIZE * line_count);
 		dst_offset += (PAGE_SIZE * line_count);
 	}
 
+	return 0;
+}
+
+static int
+nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
+		     bool no_wait_reserve, bool no_wait_gpu,
+		     struct ttm_mem_reg *new_mem)
+{
+	struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev);
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
+	struct nouveau_channel *chan;
+	int ret;
+
+	chan = nvbo->channel;
+	if (!chan || nvbo->no_vm)
+		chan = dev_priv->channel;
+
+	if (dev_priv->card_type < NV_50)
+		ret = nv04_bo_move_m2mf(chan, bo, &bo->mem, new_mem);
+	else
+		ret = nv50_bo_move_m2mf(chan, bo, &bo->mem, new_mem);
+	if (ret)
+		return ret;
+
 	return nouveau_bo_move_accel_cleanup(chan, nvbo, evict, no_wait_reserve, no_wait_gpu, new_mem);
 }
 
@@ -606,12 +695,7 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
 
 	ret = ttm_bo_move_ttm(bo, evict, no_wait_reserve, no_wait_gpu, new_mem);
 out:
-	if (tmp_mem.mm_node) {
-		spin_lock(&bo->bdev->glob->lru_lock);
-		drm_mm_put_block(tmp_mem.mm_node);
-		spin_unlock(&bo->bdev->glob->lru_lock);
-	}
-
+	ttm_bo_mem_put(bo, &tmp_mem);
 	return ret;
 }
 
@@ -644,12 +728,7 @@ nouveau_bo_move_flips(struct ttm_buffer_object *bo, bool evict, bool intr,
 		goto out;
 
 out:
-	if (tmp_mem.mm_node) {
-		spin_lock(&bo->bdev->glob->lru_lock);
-		drm_mm_put_block(tmp_mem.mm_node);
-		spin_unlock(&bo->bdev->glob->lru_lock);
-	}
-
+	ttm_bo_mem_put(bo, &tmp_mem);
 	return ret;
 }
 
@@ -669,7 +748,7 @@ nouveau_bo_vm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *new_mem,
 		return 0;
 	}
 
-	offset = new_mem->mm_node->start << PAGE_SHIFT;
+	offset = new_mem->start << PAGE_SHIFT;
 
 	if (dev_priv->card_type == NV_50) {
 		ret = nv50_mem_vm_bind_linear(dev,
@@ -719,12 +798,6 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
 	if (ret)
 		return ret;
 
-	/* Software copy if the card isn't up and running yet. */
-	if (!dev_priv->channel) {
-		ret = ttm_bo_move_memcpy(bo, evict, no_wait_reserve, no_wait_gpu, new_mem);
-		goto out;
-	}
-
 	/* Fake bo copy. */
 	if (old_mem->mem_type == TTM_PL_SYSTEM && !bo->ttm) {
 		BUG_ON(bo->mem.mm_node != NULL);
@@ -733,6 +806,12 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr,
 		goto out;
 	}
 
+	/* Software copy if the card isn't up and running yet. */
+	if (!dev_priv->channel) {
+		ret = ttm_bo_move_memcpy(bo, evict, no_wait_reserve, no_wait_gpu, new_mem);
+		goto out;
+	}
+
 	/* Hardware assisted copy. */
 	if (new_mem->mem_type == TTM_PL_SYSTEM)
 		ret = nouveau_bo_move_flipd(bo, evict, intr, no_wait_reserve, no_wait_gpu, new_mem);
@@ -783,14 +862,14 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
 	case TTM_PL_TT:
 #if __OS_HAS_AGP
 		if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) {
-			mem->bus.offset = mem->mm_node->start << PAGE_SHIFT;
+			mem->bus.offset = mem->start << PAGE_SHIFT;
 			mem->bus.base = dev_priv->gart_info.aper_base;
 			mem->bus.is_iomem = true;
 		}
 #endif
 		break;
 	case TTM_PL_VRAM:
-		mem->bus.offset = mem->mm_node->start << PAGE_SHIFT;
+		mem->bus.offset = mem->start << PAGE_SHIFT;
 		mem->bus.base = pci_resource_start(dev->pdev, 1);
 		mem->bus.is_iomem = true;
 		break;
@@ -808,7 +887,26 @@ nouveau_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
 static int
 nouveau_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 {
-	return 0;
+	struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev);
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
+
+	/* as long as the bo isn't in vram, and isn't tiled, we've got
+	 * nothing to do here.
+	 */
+	if (bo->mem.mem_type != TTM_PL_VRAM) {
+		if (dev_priv->card_type < NV_50 || !nvbo->tile_flags)
+			return 0;
+	}
+
+	/* make sure bo is in mappable vram */
+	if (bo->mem.start + bo->mem.num_pages < dev_priv->fb_mappable_pages)
+		return 0;
+
+
+	nvbo->placement.fpfn = 0;
+	nvbo->placement.lpfn = dev_priv->fb_mappable_pages;
+	nouveau_bo_placement_set(nvbo, TTM_PL_VRAM, 0);
+	return ttm_bo_validate(bo, &nvbo->placement, false, true, false);
 }
 
 struct ttm_bo_driver nouveau_bo_driver = {
diff --git a/drivers/gpu/drm/nouveau/nouveau_calc.c b/drivers/gpu/drm/nouveau/nouveau_calc.c
index ca85da784846..dad96cce5e39 100644
--- a/drivers/gpu/drm/nouveau/nouveau_calc.c
+++ b/drivers/gpu/drm/nouveau/nouveau_calc.c
@@ -198,8 +198,8 @@ nv04_update_arb(struct drm_device *dev, int VClk, int bpp,
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv_fifo_info fifo_data;
 	struct nv_sim_state sim_data;
-	int MClk = nouveau_hw_get_clock(dev, MPLL);
-	int NVClk = nouveau_hw_get_clock(dev, NVPLL);
+	int MClk = nouveau_hw_get_clock(dev, PLL_MEMORY);
+	int NVClk = nouveau_hw_get_clock(dev, PLL_CORE);
 	uint32_t cfg1 = nvReadFB(dev, NV04_PFB_CFG1);
 
 	sim_data.pclk_khz = VClk;
@@ -234,7 +234,7 @@ nv04_update_arb(struct drm_device *dev, int VClk, int bpp,
 }
 
 static void
-nv30_update_arb(int *burst, int *lwm)
+nv20_update_arb(int *burst, int *lwm)
 {
 	unsigned int fifo_size, burst_size, graphics_lwm;
 
@@ -251,14 +251,14 @@ nouveau_calc_arb(struct drm_device *dev, int vclk, int bpp, int *burst, int *lwm
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 
-	if (dev_priv->card_type < NV_30)
+	if (dev_priv->card_type < NV_20)
 		nv04_update_arb(dev, vclk, bpp, burst, lwm);
 	else if ((dev->pci_device & 0xfff0) == 0x0240 /*CHIPSET_C51*/ ||
 		 (dev->pci_device & 0xfff0) == 0x03d0 /*CHIPSET_C512*/) {
 		*burst = 128;
 		*lwm = 0x0480;
 	} else
-		nv30_update_arb(burst, lwm);
+		nv20_update_arb(burst, lwm);
 }
 
 static int
diff --git a/drivers/gpu/drm/nouveau/nouveau_channel.c b/drivers/gpu/drm/nouveau/nouveau_channel.c
index 0480f064f2c1..373950e34814 100644
--- a/drivers/gpu/drm/nouveau/nouveau_channel.c
+++ b/drivers/gpu/drm/nouveau/nouveau_channel.c
@@ -48,14 +48,14 @@ nouveau_channel_pushbuf_ctxdma_init(struct nouveau_channel *chan)
 						  dev_priv->gart_info.aper_size,
 						  NV_DMA_ACCESS_RO, &pushbuf,
 						  NULL);
-		chan->pushbuf_base = pb->bo.mem.mm_node->start << PAGE_SHIFT;
+		chan->pushbuf_base = pb->bo.mem.start << PAGE_SHIFT;
 	} else
 	if (dev_priv->card_type != NV_04) {
 		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY, 0,
 					     dev_priv->fb_available_size,
 					     NV_DMA_ACCESS_RO,
 					     NV_DMA_TARGET_VIDMEM, &pushbuf);
-		chan->pushbuf_base = pb->bo.mem.mm_node->start << PAGE_SHIFT;
+		chan->pushbuf_base = pb->bo.mem.start << PAGE_SHIFT;
 	} else {
 		/* NV04 cmdbuf hack, from original ddx.. not sure of it's
 		 * exact reason for existing :)  PCI access to cmdbuf in
@@ -67,17 +67,11 @@ nouveau_channel_pushbuf_ctxdma_init(struct nouveau_channel *chan)
 					     dev_priv->fb_available_size,
 					     NV_DMA_ACCESS_RO,
 					     NV_DMA_TARGET_PCI, &pushbuf);
-		chan->pushbuf_base = pb->bo.mem.mm_node->start << PAGE_SHIFT;
-	}
-
-	ret = nouveau_gpuobj_ref_add(dev, chan, 0, pushbuf, &chan->pushbuf);
-	if (ret) {
-		NV_ERROR(dev, "Error referencing pushbuf ctxdma: %d\n", ret);
-		if (pushbuf != dev_priv->gart_info.sg_ctxdma)
-			nouveau_gpuobj_del(dev, &pushbuf);
-		return ret;
+		chan->pushbuf_base = pb->bo.mem.start << PAGE_SHIFT;
 	}
 
+	nouveau_gpuobj_ref(pushbuf, &chan->pushbuf);
+	nouveau_gpuobj_ref(NULL, &pushbuf);
 	return 0;
 }
 
@@ -229,7 +223,7 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret,
 
 	ret = nouveau_dma_init(chan);
 	if (!ret)
-		ret = nouveau_fence_init(chan);
+		ret = nouveau_fence_channel_init(chan);
 	if (ret) {
 		nouveau_channel_free(chan);
 		return ret;
@@ -276,7 +270,7 @@ nouveau_channel_free(struct nouveau_channel *chan)
 	 * above attempts at idling were OK, but if we failed this'll tell TTM
 	 * we're done with the buffers.
 	 */
-	nouveau_fence_fini(chan);
+	nouveau_fence_channel_fini(chan);
 
 	/* This will prevent pfifo from switching channels. */
 	pfifo->reassign(dev, false);
@@ -308,8 +302,9 @@ nouveau_channel_free(struct nouveau_channel *chan)
 	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
 
 	/* Release the channel's resources */
-	nouveau_gpuobj_ref_del(dev, &chan->pushbuf);
+	nouveau_gpuobj_ref(NULL, &chan->pushbuf);
 	if (chan->pushbuf_bo) {
+		nouveau_bo_unmap(chan->pushbuf_bo);
 		nouveau_bo_unpin(chan->pushbuf_bo);
 		nouveau_bo_ref(NULL, &chan->pushbuf_bo);
 	}
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index fc737037f751..0871495096fa 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -76,6 +76,22 @@ nouveau_encoder_connector_get(struct nouveau_encoder *encoder)
 	return NULL;
 }
 
+/*TODO: This could use improvement, and learn to handle the fixed
+ *      BIOS tables etc.  It's fine currently, for its only user.
+ */
+int
+nouveau_connector_bpp(struct drm_connector *connector)
+{
+	struct nouveau_connector *nv_connector = nouveau_connector(connector);
+
+	if (nv_connector->edid && nv_connector->edid->revision >= 4) {
+		u8 bpc = ((nv_connector->edid->input & 0x70) >> 3) + 4;
+		if (bpc > 4)
+			return bpc;
+	}
+
+	return 18;
+}
 
 static void
 nouveau_connector_destroy(struct drm_connector *drm_connector)
@@ -130,6 +146,36 @@ nouveau_connector_ddc_detect(struct drm_connector *connector,
 	return NULL;
 }
 
+static struct nouveau_encoder *
+nouveau_connector_of_detect(struct drm_connector *connector)
+{
+#ifdef __powerpc__
+	struct drm_device *dev = connector->dev;
+	struct nouveau_connector *nv_connector = nouveau_connector(connector);
+	struct nouveau_encoder *nv_encoder;
+	struct device_node *cn, *dn = pci_device_to_OF_node(dev->pdev);
+
+	if (!dn ||
+	    !((nv_encoder = find_encoder_by_type(connector, OUTPUT_TMDS)) ||
+	      (nv_encoder = find_encoder_by_type(connector, OUTPUT_ANALOG))))
+		return NULL;
+
+	for_each_child_of_node(dn, cn) {
+		const char *name = of_get_property(cn, "name", NULL);
+		const void *edid = of_get_property(cn, "EDID", NULL);
+		int idx = name ? name[strlen(name) - 1] - 'A' : 0;
+
+		if (nv_encoder->dcb->i2c_index == idx && edid) {
+			nv_connector->edid =
+				kmemdup(edid, EDID_LENGTH, GFP_KERNEL);
+			of_node_put(cn);
+			return nv_encoder;
+		}
+	}
+#endif
+	return NULL;
+}
+
 static void
 nouveau_connector_set_encoder(struct drm_connector *connector,
 			      struct nouveau_encoder *nv_encoder)
@@ -225,6 +271,12 @@ nouveau_connector_detect(struct drm_connector *connector, bool force)
 		return connector_status_connected;
 	}
 
+	nv_encoder = nouveau_connector_of_detect(connector);
+	if (nv_encoder) {
+		nouveau_connector_set_encoder(connector, nv_encoder);
+		return connector_status_connected;
+	}
+
 detect_analog:
 	nv_encoder = find_encoder_by_type(connector, OUTPUT_ANALOG);
 	if (!nv_encoder && !nouveau_tv_disable)
@@ -630,7 +682,7 @@ nouveau_connector_mode_valid(struct drm_connector *connector,
 		else
 			max_clock = nv_encoder->dp.link_nr * 162000;
 
-		clock *= 3;
+		clock = clock * nouveau_connector_bpp(connector) / 8;
 		break;
 	default:
 		BUG_ON(1);
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h
index 0d2e668ccfe5..c21ed6b16f88 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.h
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.h
@@ -55,4 +55,7 @@ nouveau_connector_create(struct drm_device *, int index);
 void
 nouveau_connector_set_polling(struct drm_connector *);
 
+int
+nouveau_connector_bpp(struct drm_connector *);
+
 #endif /* __NOUVEAU_CONNECTOR_H__ */
diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.c b/drivers/gpu/drm/nouveau/nouveau_debugfs.c
index 7933de4aff2e..8e1592368cce 100644
--- a/drivers/gpu/drm/nouveau/nouveau_debugfs.c
+++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.c
@@ -157,7 +157,23 @@ nouveau_debugfs_vbios_image(struct seq_file *m, void *data)
 	return 0;
 }
 
+static int
+nouveau_debugfs_evict_vram(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_nouveau_private *dev_priv = node->minor->dev->dev_private;
+	int ret;
+
+	ret = ttm_bo_evict_mm(&dev_priv->ttm.bdev, TTM_PL_VRAM);
+	if (ret)
+		seq_printf(m, "failed: %d", ret);
+	else
+		seq_printf(m, "succeeded\n");
+	return 0;
+}
+
 static struct drm_info_list nouveau_debugfs_list[] = {
+	{ "evict_vram", nouveau_debugfs_evict_vram, 0, NULL },
 	{ "chipset", nouveau_debugfs_chipset_info, 0, NULL },
 	{ "memory", nouveau_debugfs_memory_info, 0, NULL },
 	{ "vbios.rom", nouveau_debugfs_vbios_image, 0, NULL },
diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c
index 2e3c6caa97ee..82581e600dcd 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dma.c
@@ -28,6 +28,7 @@
 #include "drm.h"
 #include "nouveau_drv.h"
 #include "nouveau_dma.h"
+#include "nouveau_ramht.h"
 
 void
 nouveau_dma_pre_init(struct nouveau_channel *chan)
@@ -58,26 +59,17 @@ nouveau_dma_init(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *m2mf = NULL;
-	struct nouveau_gpuobj *nvsw = NULL;
+	struct nouveau_gpuobj *obj = NULL;
 	int ret, i;
 
 	/* Create NV_MEMORY_TO_MEMORY_FORMAT for buffer moves */
 	ret = nouveau_gpuobj_gr_new(chan, dev_priv->card_type < NV_50 ?
-				    0x0039 : 0x5039, &m2mf);
+				    0x0039 : 0x5039, &obj);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_ref_add(dev, chan, NvM2MF, m2mf, NULL);
-	if (ret)
-		return ret;
-
-	/* Create an NV_SW object for various sync purposes */
-	ret = nouveau_gpuobj_sw_new(chan, NV_SW, &nvsw);
-	if (ret)
-		return ret;
-
-	ret = nouveau_gpuobj_ref_add(dev, chan, NvSw, nvsw, NULL);
+	ret = nouveau_ramht_insert(chan, NvM2MF, obj);
+	nouveau_gpuobj_ref(NULL, &obj);
 	if (ret)
 		return ret;
 
@@ -91,11 +83,6 @@ nouveau_dma_init(struct nouveau_channel *chan)
 	if (ret)
 		return ret;
 
-	/* Map M2MF notifier object - fbcon. */
-	ret = nouveau_bo_map(chan->notifier_bo);
-	if (ret)
-		return ret;
-
 	/* Insert NOPS for NOUVEAU_DMA_SKIPS */
 	ret = RING_SPACE(chan, NOUVEAU_DMA_SKIPS);
 	if (ret)
@@ -113,13 +100,6 @@ nouveau_dma_init(struct nouveau_channel *chan)
 	BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_DMA_NOTIFY, 1);
 	OUT_RING(chan, NvNotify0);
 
-	/* Initialise NV_SW */
-	ret = RING_SPACE(chan, 2);
-	if (ret)
-		return ret;
-	BEGIN_RING(chan, NvSubSw, 0, 1);
-	OUT_RING(chan, NvSw);
-
 	/* Sit back and pray the channel works.. */
 	FIRE_RING(chan);
 
@@ -217,7 +197,7 @@ nv50_dma_push_wait(struct nouveau_channel *chan, int count)
 
 		chan->dma.ib_free = get - chan->dma.ib_put;
 		if (chan->dma.ib_free <= 0)
-			chan->dma.ib_free += chan->dma.ib_max + 1;
+			chan->dma.ib_free += chan->dma.ib_max;
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.h b/drivers/gpu/drm/nouveau/nouveau_dma.h
index 8b05c15866d5..d578c21d3c8d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dma.h
+++ b/drivers/gpu/drm/nouveau/nouveau_dma.h
@@ -72,6 +72,7 @@ enum {
 	NvGdiRect	= 0x8000000c,
 	NvImageBlit	= 0x8000000d,
 	NvSw		= 0x8000000e,
+	NvSema		= 0x8000000f,
 
 	/* G80+ display objects */
 	NvEvoVRAM	= 0x01000000,
diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c
index 8a1b188b4cd1..4562f309ae3d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dp.c
@@ -317,7 +317,8 @@ train:
 		return false;
 
 	config[0] = nv_encoder->dp.link_nr;
-	if (nv_encoder->dp.dpcd_version >= 0x11)
+	if (nv_encoder->dp.dpcd_version >= 0x11 &&
+	    nv_encoder->dp.enhanced_frame)
 		config[0] |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
 
 	ret = nouveau_dp_lane_count_set(encoder, config[0]);
@@ -468,10 +469,12 @@ nouveau_dp_detect(struct drm_encoder *encoder)
 	    !nv_encoder->dcb->dpconf.link_bw)
 		nv_encoder->dp.link_bw = DP_LINK_BW_1_62;
 
-	nv_encoder->dp.link_nr = dpcd[2] & 0xf;
+	nv_encoder->dp.link_nr = dpcd[2] & DP_MAX_LANE_COUNT_MASK;
 	if (nv_encoder->dp.link_nr > nv_encoder->dcb->dpconf.link_nr)
 		nv_encoder->dp.link_nr = nv_encoder->dcb->dpconf.link_nr;
 
+	nv_encoder->dp.enhanced_frame = (dpcd[2] & DP_ENHANCED_FRAME_CAP);
+
 	return true;
 }
 
@@ -524,7 +527,8 @@ nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr,
 		nv_wr32(dev, NV50_AUXCH_CTRL(index), ctrl | 0x80000000);
 		nv_wr32(dev, NV50_AUXCH_CTRL(index), ctrl);
 		nv_wr32(dev, NV50_AUXCH_CTRL(index), ctrl | 0x00010000);
-		if (!nv_wait(NV50_AUXCH_CTRL(index), 0x00010000, 0x00000000)) {
+		if (!nv_wait(dev, NV50_AUXCH_CTRL(index),
+			     0x00010000, 0x00000000)) {
 			NV_ERROR(dev, "expected bit 16 == 0, got 0x%08x\n",
 				 nv_rd32(dev, NV50_AUXCH_CTRL(index)));
 			ret = -EBUSY;
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.c b/drivers/gpu/drm/nouveau/nouveau_drv.c
index 209912a1b7a5..edc4a9ab28d1 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.c
@@ -31,13 +31,14 @@
 #include "nouveau_hw.h"
 #include "nouveau_fb.h"
 #include "nouveau_fbcon.h"
+#include "nouveau_pm.h"
 #include "nv50_display.h"
 
 #include "drm_pciids.h"
 
-MODULE_PARM_DESC(noagp, "Disable AGP");
-int nouveau_noagp;
-module_param_named(noagp, nouveau_noagp, int, 0400);
+MODULE_PARM_DESC(agpmode, "AGP mode (0 to disable AGP)");
+int nouveau_agpmode = -1;
+module_param_named(agpmode, nouveau_agpmode, int, 0400);
 
 MODULE_PARM_DESC(modeset, "Enable kernel modesetting");
 static int nouveau_modeset = -1; /* kms */
@@ -79,6 +80,10 @@ MODULE_PARM_DESC(nofbaccel, "Disable fbcon acceleration");
 int nouveau_nofbaccel = 0;
 module_param_named(nofbaccel, nouveau_nofbaccel, int, 0400);
 
+MODULE_PARM_DESC(force_post, "Force POST");
+int nouveau_force_post = 0;
+module_param_named(force_post, nouveau_force_post, int, 0400);
+
 MODULE_PARM_DESC(override_conntype, "Ignore DCB connector type");
 int nouveau_override_conntype = 0;
 module_param_named(override_conntype, nouveau_override_conntype, int, 0400);
@@ -102,6 +107,14 @@ MODULE_PARM_DESC(reg_debug, "Register access debug bitmask:\n"
 int nouveau_reg_debug;
 module_param_named(reg_debug, nouveau_reg_debug, int, 0600);
 
+MODULE_PARM_DESC(perflvl, "Performance level (default: boot)\n");
+char *nouveau_perflvl;
+module_param_named(perflvl, nouveau_perflvl, charp, 0400);
+
+MODULE_PARM_DESC(perflvl_wr, "Allow perflvl changes (warning: dangerous!)\n");
+int nouveau_perflvl_wr;
+module_param_named(perflvl_wr, nouveau_perflvl_wr, int, 0400);
+
 int nouveau_fbpercrtc;
 #if 0
 module_param_named(fbpercrtc, nouveau_fbpercrtc, int, 0400);
@@ -271,6 +284,8 @@ nouveau_pci_resume(struct pci_dev *pdev)
 	if (ret)
 		return ret;
 
+	nouveau_pm_resume(dev);
+
 	if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) {
 		ret = nouveau_mem_init_agp(dev);
 		if (ret) {
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h
index b1be617373b6..3a07e580d27a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.h
@@ -133,22 +133,24 @@ enum nouveau_flags {
 #define NVOBJ_ENGINE_DISPLAY	2
 #define NVOBJ_ENGINE_INT	0xdeadbeef
 
-#define NVOBJ_FLAG_ALLOW_NO_REFS	(1 << 0)
 #define NVOBJ_FLAG_ZERO_ALLOC		(1 << 1)
 #define NVOBJ_FLAG_ZERO_FREE		(1 << 2)
-#define NVOBJ_FLAG_FAKE			(1 << 3)
 struct nouveau_gpuobj {
+	struct drm_device *dev;
+	struct kref refcount;
 	struct list_head list;
 
-	struct nouveau_channel *im_channel;
 	struct drm_mm_node *im_pramin;
 	struct nouveau_bo *im_backing;
-	uint32_t im_backing_start;
 	uint32_t *im_backing_suspend;
 	int im_bound;
 
 	uint32_t flags;
-	int refcount;
+
+	u32 size;
+	u32 pinst;
+	u32 cinst;
+	u64 vinst;
 
 	uint32_t engine;
 	uint32_t class;
@@ -157,16 +159,6 @@ struct nouveau_gpuobj {
 	void *priv;
 };
 
-struct nouveau_gpuobj_ref {
-	struct list_head list;
-
-	struct nouveau_gpuobj *gpuobj;
-	uint32_t instance;
-
-	struct nouveau_channel *channel;
-	int handle;
-};
-
 struct nouveau_channel {
 	struct drm_device *dev;
 	int id;
@@ -192,33 +184,32 @@ struct nouveau_channel {
 	} fence;
 
 	/* DMA push buffer */
-	struct nouveau_gpuobj_ref *pushbuf;
-	struct nouveau_bo         *pushbuf_bo;
-	uint32_t                   pushbuf_base;
+	struct nouveau_gpuobj *pushbuf;
+	struct nouveau_bo     *pushbuf_bo;
+	uint32_t               pushbuf_base;
 
 	/* Notifier memory */
 	struct nouveau_bo *notifier_bo;
 	struct drm_mm notifier_heap;
 
 	/* PFIFO context */
-	struct nouveau_gpuobj_ref *ramfc;
-	struct nouveau_gpuobj_ref *cache;
+	struct nouveau_gpuobj *ramfc;
+	struct nouveau_gpuobj *cache;
 
 	/* PGRAPH context */
 	/* XXX may be merge 2 pointers as private data ??? */
-	struct nouveau_gpuobj_ref *ramin_grctx;
+	struct nouveau_gpuobj *ramin_grctx;
 	void *pgraph_ctx;
 
 	/* NV50 VM */
-	struct nouveau_gpuobj     *vm_pd;
-	struct nouveau_gpuobj_ref *vm_gart_pt;
-	struct nouveau_gpuobj_ref *vm_vram_pt[NV50_VM_VRAM_NR];
+	struct nouveau_gpuobj *vm_pd;
+	struct nouveau_gpuobj *vm_gart_pt;
+	struct nouveau_gpuobj *vm_vram_pt[NV50_VM_VRAM_NR];
 
 	/* Objects */
-	struct nouveau_gpuobj_ref *ramin; /* Private instmem */
-	struct drm_mm              ramin_heap; /* Private PRAMIN heap */
-	struct nouveau_gpuobj_ref *ramht; /* Hash table */
-	struct list_head           ramht_refs; /* Objects referenced by RAMHT */
+	struct nouveau_gpuobj *ramin; /* Private instmem */
+	struct drm_mm          ramin_heap; /* Private PRAMIN heap */
+	struct nouveau_ramht  *ramht; /* Hash table */
 
 	/* GPU object info for stuff used in-kernel (mm_enabled) */
 	uint32_t m2mf_ntfy;
@@ -296,7 +287,7 @@ struct nouveau_fb_engine {
 struct nouveau_fifo_engine {
 	int  channels;
 
-	struct nouveau_gpuobj_ref *playlist[2];
+	struct nouveau_gpuobj *playlist[2];
 	int cur_playlist;
 
 	int  (*init)(struct drm_device *);
@@ -305,7 +296,6 @@ struct nouveau_fifo_engine {
 	void (*disable)(struct drm_device *);
 	void (*enable)(struct drm_device *);
 	bool (*reassign)(struct drm_device *, bool enable);
-	bool (*cache_flush)(struct drm_device *dev);
 	bool (*cache_pull)(struct drm_device *dev, bool enable);
 
 	int  (*channel_id)(struct drm_device *);
@@ -334,7 +324,7 @@ struct nouveau_pgraph_engine {
 	int grctx_size;
 
 	/* NV2x/NV3x context table (0x400780) */
-	struct nouveau_gpuobj_ref *ctx_table;
+	struct nouveau_gpuobj *ctx_table;
 
 	int  (*init)(struct drm_device *);
 	void (*takedown)(struct drm_device *);
@@ -369,6 +359,91 @@ struct nouveau_gpio_engine {
 	void (*irq_enable)(struct drm_device *, enum dcb_gpio_tag, bool on);
 };
 
+struct nouveau_pm_voltage_level {
+	u8 voltage;
+	u8 vid;
+};
+
+struct nouveau_pm_voltage {
+	bool supported;
+	u8 vid_mask;
+
+	struct nouveau_pm_voltage_level *level;
+	int nr_level;
+};
+
+#define NOUVEAU_PM_MAX_LEVEL 8
+struct nouveau_pm_level {
+	struct device_attribute dev_attr;
+	char name[32];
+	int id;
+
+	u32 core;
+	u32 memory;
+	u32 shader;
+	u32 unk05;
+
+	u8 voltage;
+	u8 fanspeed;
+
+	u16 memscript;
+};
+
+struct nouveau_pm_temp_sensor_constants {
+	u16 offset_constant;
+	s16 offset_mult;
+	u16 offset_div;
+	u16 slope_mult;
+	u16 slope_div;
+};
+
+struct nouveau_pm_threshold_temp {
+	s16 critical;
+	s16 down_clock;
+	s16 fan_boost;
+};
+
+struct nouveau_pm_memtiming {
+	u32 reg_100220;
+	u32 reg_100224;
+	u32 reg_100228;
+	u32 reg_10022c;
+	u32 reg_100230;
+	u32 reg_100234;
+	u32 reg_100238;
+	u32 reg_10023c;
+};
+
+struct nouveau_pm_memtimings {
+	bool supported;
+	struct nouveau_pm_memtiming *timing;
+	int nr_timing;
+};
+
+struct nouveau_pm_engine {
+	struct nouveau_pm_voltage voltage;
+	struct nouveau_pm_level perflvl[NOUVEAU_PM_MAX_LEVEL];
+	int nr_perflvl;
+	struct nouveau_pm_memtimings memtimings;
+	struct nouveau_pm_temp_sensor_constants sensor_constants;
+	struct nouveau_pm_threshold_temp threshold_temp;
+
+	struct nouveau_pm_level boot;
+	struct nouveau_pm_level *cur;
+
+	struct device *hwmon;
+
+	int (*clock_get)(struct drm_device *, u32 id);
+	void *(*clock_pre)(struct drm_device *, struct nouveau_pm_level *,
+			   u32 id, int khz);
+	void (*clock_set)(struct drm_device *, void *);
+	int (*voltage_get)(struct drm_device *);
+	int (*voltage_set)(struct drm_device *, int voltage);
+	int (*fanspeed_get)(struct drm_device *);
+	int (*fanspeed_set)(struct drm_device *, int fanspeed);
+	int (*temp_get)(struct drm_device *);
+};
+
 struct nouveau_engine {
 	struct nouveau_instmem_engine instmem;
 	struct nouveau_mc_engine      mc;
@@ -378,6 +453,7 @@ struct nouveau_engine {
 	struct nouveau_fifo_engine    fifo;
 	struct nouveau_display_engine display;
 	struct nouveau_gpio_engine    gpio;
+	struct nouveau_pm_engine      pm;
 };
 
 struct nouveau_pll_vals {
@@ -522,8 +598,14 @@ struct drm_nouveau_private {
 	int flags;
 
 	void __iomem *mmio;
+
+	spinlock_t ramin_lock;
 	void __iomem *ramin;
-	uint32_t ramin_size;
+	u32 ramin_size;
+	u32 ramin_base;
+	bool ramin_available;
+	struct drm_mm ramin_heap;
+	struct list_head gpuobj_list;
 
 	struct nouveau_bo *vga_ram;
 
@@ -540,6 +622,12 @@ struct drm_nouveau_private {
 		atomic_t validate_sequence;
 	} ttm;
 
+	struct {
+		spinlock_t lock;
+		struct drm_mm heap;
+		struct nouveau_bo *bo;
+	} fence;
+
 	int fifo_alloc_count;
 	struct nouveau_channel *fifos[NOUVEAU_MAX_CHANNEL_NR];
 
@@ -550,15 +638,11 @@ struct drm_nouveau_private {
 	spinlock_t context_switch_lock;
 
 	/* RAMIN configuration, RAMFC, RAMHT and RAMRO offsets */
-	struct nouveau_gpuobj *ramht;
+	struct nouveau_ramht  *ramht;
+	struct nouveau_gpuobj *ramfc;
+	struct nouveau_gpuobj *ramro;
+
 	uint32_t ramin_rsvd_vram;
-	uint32_t ramht_offset;
-	uint32_t ramht_size;
-	uint32_t ramht_bits;
-	uint32_t ramfc_offset;
-	uint32_t ramfc_size;
-	uint32_t ramro_offset;
-	uint32_t ramro_size;
 
 	struct {
 		enum {
@@ -576,14 +660,12 @@ struct drm_nouveau_private {
 	} gart_info;
 
 	/* nv10-nv40 tiling regions */
-	struct {
-		struct nouveau_tile_reg reg[NOUVEAU_MAX_TILE_NR];
-		spinlock_t lock;
-	} tile;
+	struct nouveau_tile_reg tile[NOUVEAU_MAX_TILE_NR];
 
 	/* VRAM/fb configuration */
 	uint64_t vram_size;
 	uint64_t vram_sys_base;
+	u32 vram_rblock_size;
 
 	uint64_t fb_phys;
 	uint64_t fb_available_size;
@@ -600,10 +682,6 @@ struct drm_nouveau_private {
 	struct nouveau_gpuobj *vm_vram_pt[NV50_VM_VRAM_NR];
 	int vm_vram_pt_nr;
 
-	struct drm_mm ramin_heap;
-
-	struct list_head gpuobj_list;
-
 	struct nvbios vbios;
 
 	struct nv04_mode_state mode_reg;
@@ -633,6 +711,12 @@ struct drm_nouveau_private {
 	struct apertures_struct *apertures;
 };
 
+static inline struct drm_nouveau_private *
+nouveau_private(struct drm_device *dev)
+{
+	return dev->dev_private;
+}
+
 static inline struct drm_nouveau_private *
 nouveau_bdev(struct ttm_bo_device *bd)
 {
@@ -669,7 +753,7 @@ nouveau_bo_ref(struct nouveau_bo *ref, struct nouveau_bo **pnvbo)
 } while (0)
 
 /* nouveau_drv.c */
-extern int nouveau_noagp;
+extern int nouveau_agpmode;
 extern int nouveau_duallink;
 extern int nouveau_uscript_lvds;
 extern int nouveau_uscript_tmds;
@@ -683,7 +767,10 @@ extern char *nouveau_vbios;
 extern int nouveau_ignorelid;
 extern int nouveau_nofbaccel;
 extern int nouveau_noaccel;
+extern int nouveau_force_post;
 extern int nouveau_override_conntype;
+extern char *nouveau_perflvl;
+extern int nouveau_perflvl_wr;
 
 extern int nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state);
 extern int nouveau_pci_resume(struct pci_dev *pdev);
@@ -704,8 +791,10 @@ extern bool nouveau_wait_for_idle(struct drm_device *);
 extern int  nouveau_card_init(struct drm_device *);
 
 /* nouveau_mem.c */
-extern int  nouveau_mem_detect(struct drm_device *dev);
-extern int  nouveau_mem_init(struct drm_device *);
+extern int  nouveau_mem_vram_init(struct drm_device *);
+extern void nouveau_mem_vram_fini(struct drm_device *);
+extern int  nouveau_mem_gart_init(struct drm_device *);
+extern void nouveau_mem_gart_fini(struct drm_device *);
 extern int  nouveau_mem_init_agp(struct drm_device *);
 extern int  nouveau_mem_reset_agp(struct drm_device *);
 extern void nouveau_mem_close(struct drm_device *);
@@ -749,7 +838,6 @@ extern void nouveau_channel_free(struct nouveau_channel *);
 extern int  nouveau_gpuobj_early_init(struct drm_device *);
 extern int  nouveau_gpuobj_init(struct drm_device *);
 extern void nouveau_gpuobj_takedown(struct drm_device *);
-extern void nouveau_gpuobj_late_takedown(struct drm_device *);
 extern int  nouveau_gpuobj_suspend(struct drm_device *dev);
 extern void nouveau_gpuobj_suspend_cleanup(struct drm_device *dev);
 extern void nouveau_gpuobj_resume(struct drm_device *dev);
@@ -759,24 +847,11 @@ extern void nouveau_gpuobj_channel_takedown(struct nouveau_channel *);
 extern int nouveau_gpuobj_new(struct drm_device *, struct nouveau_channel *,
 			      uint32_t size, int align, uint32_t flags,
 			      struct nouveau_gpuobj **);
-extern int nouveau_gpuobj_del(struct drm_device *, struct nouveau_gpuobj **);
-extern int nouveau_gpuobj_ref_add(struct drm_device *, struct nouveau_channel *,
-				  uint32_t handle, struct nouveau_gpuobj *,
-				  struct nouveau_gpuobj_ref **);
-extern int nouveau_gpuobj_ref_del(struct drm_device *,
-				  struct nouveau_gpuobj_ref **);
-extern int nouveau_gpuobj_ref_find(struct nouveau_channel *, uint32_t handle,
-				   struct nouveau_gpuobj_ref **ref_ret);
-extern int nouveau_gpuobj_new_ref(struct drm_device *,
-				  struct nouveau_channel *alloc_chan,
-				  struct nouveau_channel *ref_chan,
-				  uint32_t handle, uint32_t size, int align,
-				  uint32_t flags, struct nouveau_gpuobj_ref **);
-extern int nouveau_gpuobj_new_fake(struct drm_device *,
-				   uint32_t p_offset, uint32_t b_offset,
-				   uint32_t size, uint32_t flags,
-				   struct nouveau_gpuobj **,
-				   struct nouveau_gpuobj_ref**);
+extern void nouveau_gpuobj_ref(struct nouveau_gpuobj *,
+			       struct nouveau_gpuobj **);
+extern int nouveau_gpuobj_new_fake(struct drm_device *, u32 pinst, u64 vinst,
+				   u32 size, u32 flags,
+				   struct nouveau_gpuobj **);
 extern int nouveau_gpuobj_dma_new(struct nouveau_channel *, int class,
 				  uint64_t offset, uint64_t size, int access,
 				  int target, struct nouveau_gpuobj **);
@@ -879,6 +954,7 @@ extern struct dcb_gpio_entry *nouveau_bios_gpio_entry(struct drm_device *,
 						      enum dcb_gpio_tag);
 extern struct dcb_connector_table_entry *
 nouveau_bios_connector_entry(struct drm_device *, int index);
+extern u32 get_pll_register(struct drm_device *, enum pll_types);
 extern int get_pll_limits(struct drm_device *, uint32_t limit_match,
 			  struct pll_lims *);
 extern int nouveau_bios_run_display_table(struct drm_device *,
@@ -925,10 +1001,10 @@ extern int  nv40_fb_init(struct drm_device *);
 extern void nv40_fb_takedown(struct drm_device *);
 extern void nv40_fb_set_region_tiling(struct drm_device *, int, uint32_t,
 				      uint32_t, uint32_t);
-
 /* nv50_fb.c */
 extern int  nv50_fb_init(struct drm_device *);
 extern void nv50_fb_takedown(struct drm_device *);
+extern void nv50_fb_vm_trap(struct drm_device *, int display, const char *);
 
 /* nvc0_fb.c */
 extern int  nvc0_fb_init(struct drm_device *);
@@ -939,7 +1015,6 @@ extern int  nv04_fifo_init(struct drm_device *);
 extern void nv04_fifo_disable(struct drm_device *);
 extern void nv04_fifo_enable(struct drm_device *);
 extern bool nv04_fifo_reassign(struct drm_device *, bool);
-extern bool nv04_fifo_cache_flush(struct drm_device *);
 extern bool nv04_fifo_cache_pull(struct drm_device *, bool);
 extern int  nv04_fifo_channel_id(struct drm_device *);
 extern int  nv04_fifo_create_context(struct nouveau_channel *);
@@ -977,7 +1052,6 @@ extern void nvc0_fifo_takedown(struct drm_device *);
 extern void nvc0_fifo_disable(struct drm_device *);
 extern void nvc0_fifo_enable(struct drm_device *);
 extern bool nvc0_fifo_reassign(struct drm_device *, bool);
-extern bool nvc0_fifo_cache_flush(struct drm_device *);
 extern bool nvc0_fifo_cache_pull(struct drm_device *, bool);
 extern int  nvc0_fifo_channel_id(struct drm_device *);
 extern int  nvc0_fifo_create_context(struct nouveau_channel *);
@@ -1169,15 +1243,21 @@ extern int nouveau_bo_sync_gpu(struct nouveau_bo *, struct nouveau_channel *);
 
 /* nouveau_fence.c */
 struct nouveau_fence;
-extern int nouveau_fence_init(struct nouveau_channel *);
-extern void nouveau_fence_fini(struct nouveau_channel *);
+extern int nouveau_fence_init(struct drm_device *);
+extern void nouveau_fence_fini(struct drm_device *);
+extern int nouveau_fence_channel_init(struct nouveau_channel *);
+extern void nouveau_fence_channel_fini(struct nouveau_channel *);
 extern void nouveau_fence_update(struct nouveau_channel *);
 extern int nouveau_fence_new(struct nouveau_channel *, struct nouveau_fence **,
 			     bool emit);
 extern int nouveau_fence_emit(struct nouveau_fence *);
+extern void nouveau_fence_work(struct nouveau_fence *fence,
+			       void (*work)(void *priv, bool signalled),
+			       void *priv);
 struct nouveau_channel *nouveau_fence_channel(struct nouveau_fence *);
 extern bool nouveau_fence_signalled(void *obj, void *arg);
 extern int nouveau_fence_wait(void *obj, void *arg, bool lazy, bool intr);
+extern int nouveau_fence_sync(struct nouveau_fence *, struct nouveau_channel *);
 extern int nouveau_fence_flush(void *obj, void *arg);
 extern void nouveau_fence_unref(void **obj);
 extern void *nouveau_fence_ref(void *obj);
@@ -1255,12 +1335,11 @@ static inline void nv_wr32(struct drm_device *dev, unsigned reg, u32 val)
 	iowrite32_native(val, dev_priv->mmio + reg);
 }
 
-static inline void nv_mask(struct drm_device *dev, u32 reg, u32 mask, u32 val)
+static inline u32 nv_mask(struct drm_device *dev, u32 reg, u32 mask, u32 val)
 {
 	u32 tmp = nv_rd32(dev, reg);
-	tmp &= ~mask;
-	tmp |= val;
-	nv_wr32(dev, reg, tmp);
+	nv_wr32(dev, reg, (tmp & ~mask) | val);
+	return tmp;
 }
 
 static inline u8 nv_rd08(struct drm_device *dev, unsigned reg)
@@ -1275,7 +1354,7 @@ static inline void nv_wr08(struct drm_device *dev, unsigned reg, u8 val)
 	iowrite8(val, dev_priv->mmio + reg);
 }
 
-#define nv_wait(reg, mask, val) \
+#define nv_wait(dev, reg, mask, val) \
 	nouveau_wait_until(dev, 2000000000ULL, (reg), (mask), (val))
 
 /* PRAMIN access */
@@ -1292,17 +1371,8 @@ static inline void nv_wi32(struct drm_device *dev, unsigned offset, u32 val)
 }
 
 /* object access */
-static inline u32 nv_ro32(struct drm_device *dev, struct nouveau_gpuobj *obj,
-				unsigned index)
-{
-	return nv_ri32(dev, obj->im_pramin->start + index * 4);
-}
-
-static inline void nv_wo32(struct drm_device *dev, struct nouveau_gpuobj *obj,
-				unsigned index, u32 val)
-{
-	nv_wi32(dev, obj->im_pramin->start + index * 4, val);
-}
+extern u32 nv_ro32(struct nouveau_gpuobj *, u32 offset);
+extern void nv_wo32(struct nouveau_gpuobj *, u32 offset, u32 val);
 
 /*
  * Logging
@@ -1403,6 +1473,7 @@ nv_match_device(struct drm_device *dev, unsigned device,
 #define NV_SW_SEMAPHORE_OFFSET                                       0x00000064
 #define NV_SW_SEMAPHORE_ACQUIRE                                      0x00000068
 #define NV_SW_SEMAPHORE_RELEASE                                      0x0000006c
+#define NV_SW_YIELD                                                  0x00000080
 #define NV_SW_DMA_VBLSEM                                             0x0000018c
 #define NV_SW_VBLSEM_OFFSET                                          0x00000400
 #define NV_SW_VBLSEM_RELEASE_VALUE                                   0x00000404
diff --git a/drivers/gpu/drm/nouveau/nouveau_encoder.h b/drivers/gpu/drm/nouveau/nouveau_encoder.h
index 7c82d68bc155..ae69b61d93db 100644
--- a/drivers/gpu/drm/nouveau/nouveau_encoder.h
+++ b/drivers/gpu/drm/nouveau/nouveau_encoder.h
@@ -55,6 +55,7 @@ struct nouveau_encoder {
 			int dpcd_version;
 			int link_nr;
 			int link_bw;
+			bool enhanced_frame;
 		} dp;
 	};
 };
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 87ac21ec23d2..441b12420bb1 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -28,9 +28,11 @@
 #include "drm.h"
 
 #include "nouveau_drv.h"
+#include "nouveau_ramht.h"
 #include "nouveau_dma.h"
 
-#define USE_REFCNT (dev_priv->card_type >= NV_10)
+#define USE_REFCNT(dev) (nouveau_private(dev)->chipset >= 0x10)
+#define USE_SEMA(dev) (nouveau_private(dev)->chipset >= 0x17)
 
 struct nouveau_fence {
 	struct nouveau_channel *channel;
@@ -39,6 +41,15 @@ struct nouveau_fence {
 
 	uint32_t sequence;
 	bool signalled;
+
+	void (*work)(void *priv, bool signalled);
+	void *priv;
+};
+
+struct nouveau_semaphore {
+	struct kref ref;
+	struct drm_device *dev;
+	struct drm_mm_node *mem;
 };
 
 static inline struct nouveau_fence *
@@ -59,14 +70,13 @@ nouveau_fence_del(struct kref *ref)
 void
 nouveau_fence_update(struct nouveau_channel *chan)
 {
-	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
-	struct list_head *entry, *tmp;
-	struct nouveau_fence *fence;
+	struct drm_device *dev = chan->dev;
+	struct nouveau_fence *tmp, *fence;
 	uint32_t sequence;
 
 	spin_lock(&chan->fence.lock);
 
-	if (USE_REFCNT)
+	if (USE_REFCNT(dev))
 		sequence = nvchan_rd32(chan, 0x48);
 	else
 		sequence = atomic_read(&chan->fence.last_sequence_irq);
@@ -75,12 +85,14 @@ nouveau_fence_update(struct nouveau_channel *chan)
 		goto out;
 	chan->fence.sequence_ack = sequence;
 
-	list_for_each_safe(entry, tmp, &chan->fence.pending) {
-		fence = list_entry(entry, struct nouveau_fence, entry);
-
+	list_for_each_entry_safe(fence, tmp, &chan->fence.pending, entry) {
 		sequence = fence->sequence;
 		fence->signalled = true;
 		list_del(&fence->entry);
+
+		if (unlikely(fence->work))
+			fence->work(fence->priv, true);
+
 		kref_put(&fence->refcount, nouveau_fence_del);
 
 		if (sequence == chan->fence.sequence_ack)
@@ -121,8 +133,8 @@ nouveau_fence_channel(struct nouveau_fence *fence)
 int
 nouveau_fence_emit(struct nouveau_fence *fence)
 {
-	struct drm_nouveau_private *dev_priv = fence->channel->dev->dev_private;
 	struct nouveau_channel *chan = fence->channel;
+	struct drm_device *dev = chan->dev;
 	int ret;
 
 	ret = RING_SPACE(chan, 2);
@@ -143,13 +155,32 @@ nouveau_fence_emit(struct nouveau_fence *fence)
 	list_add_tail(&fence->entry, &chan->fence.pending);
 	spin_unlock(&chan->fence.lock);
 
-	BEGIN_RING(chan, NvSubSw, USE_REFCNT ? 0x0050 : 0x0150, 1);
+	BEGIN_RING(chan, NvSubSw, USE_REFCNT(dev) ? 0x0050 : 0x0150, 1);
 	OUT_RING(chan, fence->sequence);
 	FIRE_RING(chan);
 
 	return 0;
 }
 
+void
+nouveau_fence_work(struct nouveau_fence *fence,
+		   void (*work)(void *priv, bool signalled),
+		   void *priv)
+{
+	BUG_ON(fence->work);
+
+	spin_lock(&fence->channel->fence.lock);
+
+	if (fence->signalled) {
+		work(priv, true);
+	} else {
+		fence->work = work;
+		fence->priv = priv;
+	}
+
+	spin_unlock(&fence->channel->fence.lock);
+}
+
 void
 nouveau_fence_unref(void **sync_obj)
 {
@@ -213,6 +244,162 @@ nouveau_fence_wait(void *sync_obj, void *sync_arg, bool lazy, bool intr)
 	return ret;
 }
 
+static struct nouveau_semaphore *
+alloc_semaphore(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_semaphore *sema;
+
+	if (!USE_SEMA(dev))
+		return NULL;
+
+	sema = kmalloc(sizeof(*sema), GFP_KERNEL);
+	if (!sema)
+		goto fail;
+
+	spin_lock(&dev_priv->fence.lock);
+	sema->mem = drm_mm_search_free(&dev_priv->fence.heap, 4, 0, 0);
+	if (sema->mem)
+		sema->mem = drm_mm_get_block(sema->mem, 4, 0);
+	spin_unlock(&dev_priv->fence.lock);
+
+	if (!sema->mem)
+		goto fail;
+
+	kref_init(&sema->ref);
+	sema->dev = dev;
+	nouveau_bo_wr32(dev_priv->fence.bo, sema->mem->start / 4, 0);
+
+	return sema;
+fail:
+	kfree(sema);
+	return NULL;
+}
+
+static void
+free_semaphore(struct kref *ref)
+{
+	struct nouveau_semaphore *sema =
+		container_of(ref, struct nouveau_semaphore, ref);
+	struct drm_nouveau_private *dev_priv = sema->dev->dev_private;
+
+	spin_lock(&dev_priv->fence.lock);
+	drm_mm_put_block(sema->mem);
+	spin_unlock(&dev_priv->fence.lock);
+
+	kfree(sema);
+}
+
+static void
+semaphore_work(void *priv, bool signalled)
+{
+	struct nouveau_semaphore *sema = priv;
+	struct drm_nouveau_private *dev_priv = sema->dev->dev_private;
+
+	if (unlikely(!signalled))
+		nouveau_bo_wr32(dev_priv->fence.bo, sema->mem->start / 4, 1);
+
+	kref_put(&sema->ref, free_semaphore);
+}
+
+static int
+emit_semaphore(struct nouveau_channel *chan, int method,
+	       struct nouveau_semaphore *sema)
+{
+	struct drm_nouveau_private *dev_priv = sema->dev->dev_private;
+	struct nouveau_fence *fence;
+	bool smart = (dev_priv->card_type >= NV_50);
+	int ret;
+
+	ret = RING_SPACE(chan, smart ? 8 : 4);
+	if (ret)
+		return ret;
+
+	if (smart) {
+		BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 1);
+		OUT_RING(chan, NvSema);
+	}
+	BEGIN_RING(chan, NvSubSw, NV_SW_SEMAPHORE_OFFSET, 1);
+	OUT_RING(chan, sema->mem->start);
+
+	if (smart && method == NV_SW_SEMAPHORE_ACQUIRE) {
+		/*
+		 * NV50 tries to be too smart and context-switch
+		 * between semaphores instead of doing a "first come,
+		 * first served" strategy like previous cards
+		 * do.
+		 *
+		 * That's bad because the ACQUIRE latency can get as
+		 * large as the PFIFO context time slice in the
+		 * typical DRI2 case where you have several
+		 * outstanding semaphores at the same moment.
+		 *
+		 * If we're going to ACQUIRE, force the card to
+		 * context switch before, just in case the matching
+		 * RELEASE is already scheduled to be executed in
+		 * another channel.
+		 */
+		BEGIN_RING(chan, NvSubSw, NV_SW_YIELD, 1);
+		OUT_RING(chan, 0);
+	}
+
+	BEGIN_RING(chan, NvSubSw, method, 1);
+	OUT_RING(chan, 1);
+
+	if (smart && method == NV_SW_SEMAPHORE_RELEASE) {
+		/*
+		 * Force the card to context switch, there may be
+		 * another channel waiting for the semaphore we just
+		 * released.
+		 */
+		BEGIN_RING(chan, NvSubSw, NV_SW_YIELD, 1);
+		OUT_RING(chan, 0);
+	}
+
+	/* Delay semaphore destruction until its work is done */
+	ret = nouveau_fence_new(chan, &fence, true);
+	if (ret)
+		return ret;
+
+	kref_get(&sema->ref);
+	nouveau_fence_work(fence, semaphore_work, sema);
+	nouveau_fence_unref((void *)&fence);
+
+	return 0;
+}
+
+int
+nouveau_fence_sync(struct nouveau_fence *fence,
+		   struct nouveau_channel *wchan)
+{
+	struct nouveau_channel *chan = nouveau_fence_channel(fence);
+	struct drm_device *dev = wchan->dev;
+	struct nouveau_semaphore *sema;
+	int ret;
+
+	if (likely(!fence || chan == wchan ||
+		   nouveau_fence_signalled(fence, NULL)))
+		return 0;
+
+	sema = alloc_semaphore(dev);
+	if (!sema) {
+		/* Early card or broken userspace, fall back to
+		 * software sync. */
+		return nouveau_fence_wait(fence, NULL, false, false);
+	}
+
+	/* Make wchan wait until it gets signalled */
+	ret = emit_semaphore(wchan, NV_SW_SEMAPHORE_ACQUIRE, sema);
+	if (ret)
+		goto out;
+
+	/* Signal the semaphore from chan */
+	ret = emit_semaphore(chan, NV_SW_SEMAPHORE_RELEASE, sema);
+out:
+	kref_put(&sema->ref, free_semaphore);
+	return ret;
+}
+
 int
 nouveau_fence_flush(void *sync_obj, void *sync_arg)
 {
@@ -220,26 +407,123 @@ nouveau_fence_flush(void *sync_obj, void *sync_arg)
 }
 
 int
-nouveau_fence_init(struct nouveau_channel *chan)
+nouveau_fence_channel_init(struct nouveau_channel *chan)
 {
+	struct drm_device *dev = chan->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpuobj *obj = NULL;
+	int ret;
+
+	/* Create an NV_SW object for various sync purposes */
+	ret = nouveau_gpuobj_sw_new(chan, NV_SW, &obj);
+	if (ret)
+		return ret;
+
+	ret = nouveau_ramht_insert(chan, NvSw, obj);
+	nouveau_gpuobj_ref(NULL, &obj);
+	if (ret)
+		return ret;
+
+	ret = RING_SPACE(chan, 2);
+	if (ret)
+		return ret;
+	BEGIN_RING(chan, NvSubSw, 0, 1);
+	OUT_RING(chan, NvSw);
+
+	/* Create a DMA object for the shared cross-channel sync area. */
+	if (USE_SEMA(dev)) {
+		struct drm_mm_node *mem = dev_priv->fence.bo->bo.mem.mm_node;
+
+		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY,
+					     mem->start << PAGE_SHIFT,
+					     mem->size << PAGE_SHIFT,
+					     NV_DMA_ACCESS_RW,
+					     NV_DMA_TARGET_VIDMEM, &obj);
+		if (ret)
+			return ret;
+
+		ret = nouveau_ramht_insert(chan, NvSema, obj);
+		nouveau_gpuobj_ref(NULL, &obj);
+		if (ret)
+			return ret;
+
+		ret = RING_SPACE(chan, 2);
+		if (ret)
+			return ret;
+		BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 1);
+		OUT_RING(chan, NvSema);
+	}
+
+	FIRE_RING(chan);
+
 	INIT_LIST_HEAD(&chan->fence.pending);
 	spin_lock_init(&chan->fence.lock);
 	atomic_set(&chan->fence.last_sequence_irq, 0);
+
 	return 0;
 }
 
 void
-nouveau_fence_fini(struct nouveau_channel *chan)
+nouveau_fence_channel_fini(struct nouveau_channel *chan)
 {
-	struct list_head *entry, *tmp;
-	struct nouveau_fence *fence;
-
-	list_for_each_safe(entry, tmp, &chan->fence.pending) {
-		fence = list_entry(entry, struct nouveau_fence, entry);
+	struct nouveau_fence *tmp, *fence;
 
+	list_for_each_entry_safe(fence, tmp, &chan->fence.pending, entry) {
 		fence->signalled = true;
 		list_del(&fence->entry);
+
+		if (unlikely(fence->work))
+			fence->work(fence->priv, false);
+
 		kref_put(&fence->refcount, nouveau_fence_del);
 	}
 }
 
+int
+nouveau_fence_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	int ret;
+
+	/* Create a shared VRAM heap for cross-channel sync. */
+	if (USE_SEMA(dev)) {
+		ret = nouveau_bo_new(dev, NULL, 4096, 0, TTM_PL_FLAG_VRAM,
+				     0, 0, false, true, &dev_priv->fence.bo);
+		if (ret)
+			return ret;
+
+		ret = nouveau_bo_pin(dev_priv->fence.bo, TTM_PL_FLAG_VRAM);
+		if (ret)
+			goto fail;
+
+		ret = nouveau_bo_map(dev_priv->fence.bo);
+		if (ret)
+			goto fail;
+
+		ret = drm_mm_init(&dev_priv->fence.heap, 0,
+				  dev_priv->fence.bo->bo.mem.size);
+		if (ret)
+			goto fail;
+
+		spin_lock_init(&dev_priv->fence.lock);
+	}
+
+	return 0;
+fail:
+	nouveau_bo_unmap(dev_priv->fence.bo);
+	nouveau_bo_ref(NULL, &dev_priv->fence.bo);
+	return ret;
+}
+
+void
+nouveau_fence_fini(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	if (USE_SEMA(dev)) {
+		drm_mm_takedown(&dev_priv->fence.heap);
+		nouveau_bo_unmap(dev_priv->fence.bo);
+		nouveau_bo_unpin(dev_priv->fence.bo);
+		nouveau_bo_ref(NULL, &dev_priv->fence.bo);
+	}
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 19620a6709f5..5c4c929d7f74 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -362,7 +362,7 @@ validate_list(struct nouveau_channel *chan, struct list_head *list,
 	list_for_each_entry(nvbo, list, entry) {
 		struct drm_nouveau_gem_pushbuf_bo *b = &pbbo[nvbo->pbbo_index];
 
-		ret = nouveau_bo_sync_gpu(nvbo, chan);
+		ret = nouveau_fence_sync(nvbo->bo.sync_obj, chan);
 		if (unlikely(ret)) {
 			NV_ERROR(dev, "fail pre-validate sync\n");
 			return ret;
@@ -385,7 +385,7 @@ validate_list(struct nouveau_channel *chan, struct list_head *list,
 			return ret;
 		}
 
-		ret = nouveau_bo_sync_gpu(nvbo, chan);
+		ret = nouveau_fence_sync(nvbo->bo.sync_obj, chan);
 		if (unlikely(ret)) {
 			NV_ERROR(dev, "fail post-validate sync\n");
 			return ret;
diff --git a/drivers/gpu/drm/nouveau/nouveau_grctx.h b/drivers/gpu/drm/nouveau/nouveau_grctx.h
index 5d39c4ce8006..4a8ad1307fa4 100644
--- a/drivers/gpu/drm/nouveau/nouveau_grctx.h
+++ b/drivers/gpu/drm/nouveau/nouveau_grctx.h
@@ -126,7 +126,7 @@ gr_def(struct nouveau_grctx *ctx, uint32_t reg, uint32_t val)
 	reg = (reg - 0x00400000) / 4;
 	reg = (reg - ctx->ctxprog_reg) + ctx->ctxvals_base;
 
-	nv_wo32(ctx->dev, ctx->data, reg, val);
+	nv_wo32(ctx->data, reg * 4, val);
 }
 #endif
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_hw.c b/drivers/gpu/drm/nouveau/nouveau_hw.c
index 7b613682e400..bed669a54a2d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_hw.c
+++ b/drivers/gpu/drm/nouveau/nouveau_hw.c
@@ -305,7 +305,7 @@ setPLL_double_lowregs(struct drm_device *dev, uint32_t NMNMreg,
 	bool mpll = Preg == 0x4020;
 	uint32_t oldPval = nvReadMC(dev, Preg);
 	uint32_t NMNM = pv->NM2 << 16 | pv->NM1;
-	uint32_t Pval = (oldPval & (mpll ? ~(0x11 << 16) : ~(1 << 16))) |
+	uint32_t Pval = (oldPval & (mpll ? ~(0x77 << 16) : ~(7 << 16))) |
 			0xc << 28 | pv->log2P << 16;
 	uint32_t saved4600 = 0;
 	/* some cards have different maskc040s */
@@ -427,22 +427,12 @@ nouveau_hw_get_pllvals(struct drm_device *dev, enum pll_types plltype,
 		       struct nouveau_pll_vals *pllvals)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	const uint32_t nv04_regs[MAX_PLL_TYPES] = { NV_PRAMDAC_NVPLL_COEFF,
-						    NV_PRAMDAC_MPLL_COEFF,
-						    NV_PRAMDAC_VPLL_COEFF,
-						    NV_RAMDAC_VPLL2 };
-	const uint32_t nv40_regs[MAX_PLL_TYPES] = { 0x4000,
-						    0x4020,
-						    NV_PRAMDAC_VPLL_COEFF,
-						    NV_RAMDAC_VPLL2 };
-	uint32_t reg1, pll1, pll2 = 0;
+	uint32_t reg1 = get_pll_register(dev, plltype), pll1, pll2 = 0;
 	struct pll_lims pll_lim;
 	int ret;
 
-	if (dev_priv->card_type < NV_40)
-		reg1 = nv04_regs[plltype];
-	else
-		reg1 = nv40_regs[plltype];
+	if (reg1 == 0)
+		return -ENOENT;
 
 	pll1 = nvReadMC(dev, reg1);
 
@@ -491,8 +481,10 @@ int
 nouveau_hw_get_clock(struct drm_device *dev, enum pll_types plltype)
 {
 	struct nouveau_pll_vals pllvals;
+	int ret;
 
-	if (plltype == MPLL && (dev->pci_device & 0x0ff0) == CHIPSET_NFORCE) {
+	if (plltype == PLL_MEMORY &&
+	    (dev->pci_device & 0x0ff0) == CHIPSET_NFORCE) {
 		uint32_t mpllP;
 
 		pci_read_config_dword(pci_get_bus_and_slot(0, 3), 0x6c, &mpllP);
@@ -501,14 +493,17 @@ nouveau_hw_get_clock(struct drm_device *dev, enum pll_types plltype)
 
 		return 400000 / mpllP;
 	} else
-	if (plltype == MPLL && (dev->pci_device & 0xff0) == CHIPSET_NFORCE2) {
+	if (plltype == PLL_MEMORY &&
+	    (dev->pci_device & 0xff0) == CHIPSET_NFORCE2) {
 		uint32_t clock;
 
 		pci_read_config_dword(pci_get_bus_and_slot(0, 5), 0x4c, &clock);
 		return clock;
 	}
 
-	nouveau_hw_get_pllvals(dev, plltype, &pllvals);
+	ret = nouveau_hw_get_pllvals(dev, plltype, &pllvals);
+	if (ret)
+		return ret;
 
 	return nouveau_hw_pllvals_to_clk(&pllvals);
 }
@@ -526,9 +521,9 @@ nouveau_hw_fix_bad_vpll(struct drm_device *dev, int head)
 	struct nouveau_pll_vals pv;
 	uint32_t pllreg = head ? NV_RAMDAC_VPLL2 : NV_PRAMDAC_VPLL_COEFF;
 
-	if (get_pll_limits(dev, head ? VPLL2 : VPLL1, &pll_lim))
+	if (get_pll_limits(dev, pllreg, &pll_lim))
 		return;
-	nouveau_hw_get_pllvals(dev, head ? VPLL2 : VPLL1, &pv);
+	nouveau_hw_get_pllvals(dev, pllreg, &pv);
 
 	if (pv.M1 >= pll_lim.vco1.min_m && pv.M1 <= pll_lim.vco1.max_m &&
 	    pv.N1 >= pll_lim.vco1.min_n && pv.N1 <= pll_lim.vco1.max_n &&
@@ -661,7 +656,7 @@ nv_save_state_ramdac(struct drm_device *dev, int head,
 	if (dev_priv->card_type >= NV_10)
 		regp->nv10_cursync = NVReadRAMDAC(dev, head, NV_RAMDAC_NV10_CURSYNC);
 
-	nouveau_hw_get_pllvals(dev, head ? VPLL2 : VPLL1, &regp->pllvals);
+	nouveau_hw_get_pllvals(dev, head ? PLL_VPLL1 : PLL_VPLL0, &regp->pllvals);
 	state->pllsel = NVReadRAMDAC(dev, 0, NV_PRAMDAC_PLL_COEFF_SELECT);
 	if (nv_two_heads(dev))
 		state->sel_clk = NVReadRAMDAC(dev, 0, NV_PRAMDAC_SEL_CLK);
@@ -866,10 +861,11 @@ nv_save_state_ext(struct drm_device *dev, int head,
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_FFLWM__INDEX);
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_21);
 
-	if (dev_priv->card_type >= NV_30) {
+	if (dev_priv->card_type >= NV_20)
 		rd_cio_state(dev, head, regp, NV_CIO_CRE_47);
+
+	if (dev_priv->card_type >= NV_30)
 		rd_cio_state(dev, head, regp, 0x9f);
-	}
 
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_49);
 	rd_cio_state(dev, head, regp, NV_CIO_CRE_HCUR_ADDR0_INDEX);
@@ -976,10 +972,11 @@ nv_load_state_ext(struct drm_device *dev, int head,
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_FF_INDEX);
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_FFLWM__INDEX);
 
-	if (dev_priv->card_type >= NV_30) {
+	if (dev_priv->card_type >= NV_20)
 		wr_cio_state(dev, head, regp, NV_CIO_CRE_47);
+
+	if (dev_priv->card_type >= NV_30)
 		wr_cio_state(dev, head, regp, 0x9f);
-	}
 
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_49);
 	wr_cio_state(dev, head, regp, NV_CIO_CRE_HCUR_ADDR0_INDEX);
diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.c b/drivers/gpu/drm/nouveau/nouveau_i2c.c
index 84614858728b..fdd7e3de79c8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_i2c.c
+++ b/drivers/gpu/drm/nouveau/nouveau_i2c.c
@@ -299,7 +299,10 @@ nouveau_probe_i2c_addr(struct nouveau_i2c_chan *i2c, int addr)
 
 int
 nouveau_i2c_identify(struct drm_device *dev, const char *what,
-		     struct i2c_board_info *info, int index)
+		     struct i2c_board_info *info,
+		     bool (*match)(struct nouveau_i2c_chan *,
+				   struct i2c_board_info *),
+		     int index)
 {
 	struct nouveau_i2c_chan *i2c = nouveau_i2c_find(dev, index);
 	int i;
@@ -307,7 +310,8 @@ nouveau_i2c_identify(struct drm_device *dev, const char *what,
 	NV_DEBUG(dev, "Probing %ss on I2C bus: %d\n", what, index);
 
 	for (i = 0; info[i].addr; i++) {
-		if (nouveau_probe_i2c_addr(i2c, info[i].addr)) {
+		if (nouveau_probe_i2c_addr(i2c, info[i].addr) &&
+		    (!match || match(i2c, &info[i]))) {
 			NV_INFO(dev, "Detected %s: %s\n", what, info[i].type);
 			return i;
 		}
diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.h b/drivers/gpu/drm/nouveau/nouveau_i2c.h
index f71cb32f7571..c77a6ba66b7c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_i2c.h
+++ b/drivers/gpu/drm/nouveau/nouveau_i2c.h
@@ -44,7 +44,10 @@ void nouveau_i2c_fini(struct drm_device *, struct dcb_i2c_entry *);
 struct nouveau_i2c_chan *nouveau_i2c_find(struct drm_device *, int index);
 bool nouveau_probe_i2c_addr(struct nouveau_i2c_chan *i2c, int addr);
 int nouveau_i2c_identify(struct drm_device *dev, const char *what,
-			 struct i2c_board_info *info, int index);
+			 struct i2c_board_info *info,
+			 bool (*match)(struct nouveau_i2c_chan *,
+				       struct i2c_board_info *),
+			 int index);
 
 extern const struct i2c_algorithm nouveau_dp_i2c_algo;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_irq.c b/drivers/gpu/drm/nouveau/nouveau_irq.c
index 794b0ee30cf6..6fd51a51c608 100644
--- a/drivers/gpu/drm/nouveau/nouveau_irq.c
+++ b/drivers/gpu/drm/nouveau/nouveau_irq.c
@@ -35,6 +35,7 @@
 #include "nouveau_drm.h"
 #include "nouveau_drv.h"
 #include "nouveau_reg.h"
+#include "nouveau_ramht.h"
 #include <linux/ratelimit.h>
 
 /* needed for hotplug irq */
@@ -106,15 +107,16 @@ nouveau_fifo_swmthd(struct nouveau_channel *chan, uint32_t addr, uint32_t data)
 	const int mthd = addr & 0x1ffc;
 
 	if (mthd == 0x0000) {
-		struct nouveau_gpuobj_ref *ref = NULL;
+		struct nouveau_gpuobj *gpuobj;
 
-		if (nouveau_gpuobj_ref_find(chan, data, &ref))
+		gpuobj = nouveau_ramht_find(chan, data);
+		if (!gpuobj)
 			return false;
 
-		if (ref->gpuobj->engine != NVOBJ_ENGINE_SW)
+		if (gpuobj->engine != NVOBJ_ENGINE_SW)
 			return false;
 
-		chan->sw_subchannel[subc] = ref->gpuobj->class;
+		chan->sw_subchannel[subc] = gpuobj->class;
 		nv_wr32(dev, NV04_PFIFO_CACHE1_ENGINE, nv_rd32(dev,
 			NV04_PFIFO_CACHE1_ENGINE) & ~(0xf << subc * 4));
 		return true;
@@ -200,16 +202,45 @@ nouveau_fifo_irq_handler(struct drm_device *dev)
 		}
 
 		if (status & NV_PFIFO_INTR_DMA_PUSHER) {
-			NV_INFO(dev, "PFIFO_DMA_PUSHER - Ch %d\n", chid);
+			u32 get = nv_rd32(dev, 0x003244);
+			u32 put = nv_rd32(dev, 0x003240);
+			u32 push = nv_rd32(dev, 0x003220);
+			u32 state = nv_rd32(dev, 0x003228);
 
+			if (dev_priv->card_type == NV_50) {
+				u32 ho_get = nv_rd32(dev, 0x003328);
+				u32 ho_put = nv_rd32(dev, 0x003320);
+				u32 ib_get = nv_rd32(dev, 0x003334);
+				u32 ib_put = nv_rd32(dev, 0x003330);
+
+				NV_INFO(dev, "PFIFO_DMA_PUSHER - Ch %d Get 0x%02x%08x "
+					     "Put 0x%02x%08x IbGet 0x%08x IbPut 0x%08x "
+					     "State 0x%08x Push 0x%08x\n",
+					chid, ho_get, get, ho_put, put, ib_get, ib_put,
+					state, push);
+
+				/* METHOD_COUNT, in DMA_STATE on earlier chipsets */
+				nv_wr32(dev, 0x003364, 0x00000000);
+				if (get != put || ho_get != ho_put) {
+					nv_wr32(dev, 0x003244, put);
+					nv_wr32(dev, 0x003328, ho_put);
+				} else
+				if (ib_get != ib_put) {
+					nv_wr32(dev, 0x003334, ib_put);
+				}
+			} else {
+				NV_INFO(dev, "PFIFO_DMA_PUSHER - Ch %d Get 0x%08x "
+					     "Put 0x%08x State 0x%08x Push 0x%08x\n",
+					chid, get, put, state, push);
+
+				if (get != put)
+					nv_wr32(dev, 0x003244, put);
+			}
+
+			nv_wr32(dev, 0x003228, 0x00000000);
+			nv_wr32(dev, 0x003220, 0x00000001);
+			nv_wr32(dev, 0x002100, NV_PFIFO_INTR_DMA_PUSHER);
 			status &= ~NV_PFIFO_INTR_DMA_PUSHER;
-			nv_wr32(dev, NV03_PFIFO_INTR_0,
-						NV_PFIFO_INTR_DMA_PUSHER);
-
-			nv_wr32(dev, NV04_PFIFO_CACHE1_DMA_STATE, 0x00000000);
-			if (nv_rd32(dev, NV04_PFIFO_CACHE1_DMA_PUT) != get)
-				nv_wr32(dev, NV04_PFIFO_CACHE1_DMA_GET,
-								get + 4);
 		}
 
 		if (status & NV_PFIFO_INTR_SEMAPHORE) {
@@ -226,6 +257,14 @@ nouveau_fifo_irq_handler(struct drm_device *dev)
 			nv_wr32(dev, NV04_PFIFO_CACHE1_PULL0, 1);
 		}
 
+		if (dev_priv->card_type == NV_50) {
+			if (status & 0x00000010) {
+				nv50_fb_vm_trap(dev, 1, "PFIFO_BAR_FAULT");
+				status &= ~0x00000010;
+				nv_wr32(dev, 0x002100, 0x00000010);
+			}
+		}
+
 		if (status) {
 			NV_INFO(dev, "PFIFO_INTR 0x%08x - Ch %d\n",
 				status, chid);
@@ -357,7 +396,7 @@ nouveau_graph_chid_from_grctx(struct drm_device *dev)
 			if (!chan || !chan->ramin_grctx)
 				continue;
 
-			if (inst == chan->ramin_grctx->instance)
+			if (inst == chan->ramin_grctx->pinst)
 				break;
 		}
 	} else {
@@ -369,7 +408,7 @@ nouveau_graph_chid_from_grctx(struct drm_device *dev)
 			if (!chan || !chan->ramin)
 				continue;
 
-			if (inst == chan->ramin->instance)
+			if (inst == chan->ramin->vinst)
 				break;
 		}
 	}
@@ -605,40 +644,6 @@ nouveau_pgraph_irq_handler(struct drm_device *dev)
 	nv_wr32(dev, NV03_PMC_INTR_0, NV_PMC_INTR_0_PGRAPH_PENDING);
 }
 
-static void
-nv50_pfb_vm_trap(struct drm_device *dev, int display, const char *name)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	uint32_t trap[6];
-	int i, ch;
-	uint32_t idx = nv_rd32(dev, 0x100c90);
-	if (idx & 0x80000000) {
-		idx &= 0xffffff;
-		if (display) {
-			for (i = 0; i < 6; i++) {
-				nv_wr32(dev, 0x100c90, idx | i << 24);
-				trap[i] = nv_rd32(dev, 0x100c94);
-			}
-			for (ch = 0; ch < dev_priv->engine.fifo.channels; ch++) {
-				struct nouveau_channel *chan = dev_priv->fifos[ch];
-
-				if (!chan || !chan->ramin)
-					continue;
-
-				if (trap[1] == chan->ramin->instance >> 12)
-					break;
-			}
-			NV_INFO(dev, "%s - VM: Trapped %s at %02x%04x%04x status %08x %08x channel %d\n",
-					name, (trap[5]&0x100?"read":"write"),
-					trap[5]&0xff, trap[4]&0xffff,
-					trap[3]&0xffff, trap[0], trap[2], ch);
-		}
-		nv_wr32(dev, 0x100c90, idx | 0x80000000);
-	} else if (display) {
-		NV_INFO(dev, "%s - no VM fault?\n", name);
-	}
-}
-
 static struct nouveau_enum_names nv50_mp_exec_error_names[] =
 {
 	{ 3, "STACK_UNDERFLOW" },
@@ -711,7 +716,7 @@ nv50_pgraph_tp_trap(struct drm_device *dev, int type, uint32_t ustatus_old,
 		tps++;
 		switch (type) {
 		case 6: /* texture error... unknown for now */
-			nv50_pfb_vm_trap(dev, display, name);
+			nv50_fb_vm_trap(dev, display, name);
 			if (display) {
 				NV_ERROR(dev, "magic set %d:\n", i);
 				for (r = ustatus_addr + 4; r <= ustatus_addr + 0x10; r += 4)
@@ -734,7 +739,7 @@ nv50_pgraph_tp_trap(struct drm_device *dev, int type, uint32_t ustatus_old,
 			uint32_t e1c = nv_rd32(dev, ustatus_addr + 0x14);
 			uint32_t e20 = nv_rd32(dev, ustatus_addr + 0x18);
 			uint32_t e24 = nv_rd32(dev, ustatus_addr + 0x1c);
-			nv50_pfb_vm_trap(dev, display, name);
+			nv50_fb_vm_trap(dev, display, name);
 			/* 2d engine destination */
 			if (ustatus & 0x00000010) {
 				if (display) {
@@ -817,7 +822,7 @@ nv50_pgraph_trap_handler(struct drm_device *dev)
 
 		/* Known to be triggered by screwed up NOTIFY and COND... */
 		if (ustatus & 0x00000001) {
-			nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_DISPATCH_FAULT");
+			nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_DISPATCH_FAULT");
 			nv_wr32(dev, 0x400500, 0);
 			if (nv_rd32(dev, 0x400808) & 0x80000000) {
 				if (display) {
@@ -842,7 +847,7 @@ nv50_pgraph_trap_handler(struct drm_device *dev)
 			ustatus &= ~0x00000001;
 		}
 		if (ustatus & 0x00000002) {
-			nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_DISPATCH_QUERY");
+			nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_DISPATCH_QUERY");
 			nv_wr32(dev, 0x400500, 0);
 			if (nv_rd32(dev, 0x40084c) & 0x80000000) {
 				if (display) {
@@ -884,15 +889,15 @@ nv50_pgraph_trap_handler(struct drm_device *dev)
 			NV_INFO(dev, "PGRAPH_TRAP_M2MF - no ustatus?\n");
 		}
 		if (ustatus & 0x00000001) {
-			nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_M2MF_NOTIFY");
+			nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_M2MF_NOTIFY");
 			ustatus &= ~0x00000001;
 		}
 		if (ustatus & 0x00000002) {
-			nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_M2MF_IN");
+			nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_M2MF_IN");
 			ustatus &= ~0x00000002;
 		}
 		if (ustatus & 0x00000004) {
-			nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_M2MF_OUT");
+			nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_M2MF_OUT");
 			ustatus &= ~0x00000004;
 		}
 		NV_INFO (dev, "PGRAPH_TRAP_M2MF - %08x %08x %08x %08x\n",
@@ -917,7 +922,7 @@ nv50_pgraph_trap_handler(struct drm_device *dev)
 			NV_INFO(dev, "PGRAPH_TRAP_VFETCH - no ustatus?\n");
 		}
 		if (ustatus & 0x00000001) {
-			nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_VFETCH_FAULT");
+			nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_VFETCH_FAULT");
 			NV_INFO (dev, "PGRAPH_TRAP_VFETCH_FAULT - %08x %08x %08x %08x\n",
 					nv_rd32(dev, 0x400c00),
 					nv_rd32(dev, 0x400c08),
@@ -939,7 +944,7 @@ nv50_pgraph_trap_handler(struct drm_device *dev)
 			NV_INFO(dev, "PGRAPH_TRAP_STRMOUT - no ustatus?\n");
 		}
 		if (ustatus & 0x00000001) {
-			nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_STRMOUT_FAULT");
+			nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_STRMOUT_FAULT");
 			NV_INFO (dev, "PGRAPH_TRAP_STRMOUT_FAULT - %08x %08x %08x %08x\n",
 					nv_rd32(dev, 0x401804),
 					nv_rd32(dev, 0x401808),
@@ -964,7 +969,7 @@ nv50_pgraph_trap_handler(struct drm_device *dev)
 			NV_INFO(dev, "PGRAPH_TRAP_CCACHE - no ustatus?\n");
 		}
 		if (ustatus & 0x00000001) {
-			nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_CCACHE_FAULT");
+			nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_CCACHE_FAULT");
 			NV_INFO (dev, "PGRAPH_TRAP_CCACHE_FAULT - %08x %08x %08x %08x %08x %08x %08x\n",
 					nv_rd32(dev, 0x405800),
 					nv_rd32(dev, 0x405804),
@@ -986,7 +991,7 @@ nv50_pgraph_trap_handler(struct drm_device *dev)
 	 * remaining, so try to handle it anyway. Perhaps related to that
 	 * unknown DMA slot on tesla? */
 	if (status & 0x20) {
-		nv50_pfb_vm_trap(dev, display, "PGRAPH_TRAP_UNKC04");
+		nv50_fb_vm_trap(dev, display, "PGRAPH_TRAP_UNKC04");
 		ustatus = nv_rd32(dev, 0x402000) & 0x7fffffff;
 		if (display)
 			NV_INFO(dev, "PGRAPH_TRAP_UNKC04 - Unhandled ustatus 0x%08x\n", ustatus);
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c
index 9689d4147686..a163c7c612e7 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.c
@@ -35,6 +35,8 @@
 #include "drm_sarea.h"
 #include "nouveau_drv.h"
 
+#define MIN(a,b) a < b ? a : b
+
 /*
  * NV10-NV40 tiling helpers
  */
@@ -47,18 +49,14 @@ nv10_mem_set_region_tiling(struct drm_device *dev, int i, uint32_t addr,
 	struct nouveau_fifo_engine *pfifo = &dev_priv->engine.fifo;
 	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
 	struct nouveau_pgraph_engine *pgraph = &dev_priv->engine.graph;
-	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
+	struct nouveau_tile_reg *tile = &dev_priv->tile[i];
 
 	tile->addr = addr;
 	tile->size = size;
 	tile->used = !!pitch;
 	nouveau_fence_unref((void **)&tile->fence);
 
-	if (!pfifo->cache_flush(dev))
-		return;
-
 	pfifo->reassign(dev, false);
-	pfifo->cache_flush(dev);
 	pfifo->cache_pull(dev, false);
 
 	nouveau_wait_for_idle(dev);
@@ -76,34 +74,36 @@ nv10_mem_set_tiling(struct drm_device *dev, uint32_t addr, uint32_t size,
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
-	struct nouveau_tile_reg *tile = dev_priv->tile.reg, *found = NULL;
-	int i;
+	struct nouveau_tile_reg *found = NULL;
+	unsigned long i, flags;
 
-	spin_lock(&dev_priv->tile.lock);
+	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
 
 	for (i = 0; i < pfb->num_tiles; i++) {
-		if (tile[i].used)
+		struct nouveau_tile_reg *tile = &dev_priv->tile[i];
+
+		if (tile->used)
 			/* Tile region in use. */
 			continue;
 
-		if (tile[i].fence &&
-		    !nouveau_fence_signalled(tile[i].fence, NULL))
+		if (tile->fence &&
+		    !nouveau_fence_signalled(tile->fence, NULL))
 			/* Pending tile region. */
 			continue;
 
-		if (max(tile[i].addr, addr) <
-		    min(tile[i].addr + tile[i].size, addr + size))
+		if (max(tile->addr, addr) <
+		    min(tile->addr + tile->size, addr + size))
 			/* Kill an intersecting tile region. */
 			nv10_mem_set_region_tiling(dev, i, 0, 0, 0);
 
 		if (pitch && !found) {
 			/* Free tile region. */
 			nv10_mem_set_region_tiling(dev, i, addr, size, pitch);
-			found = &tile[i];
+			found = tile;
 		}
 	}
 
-	spin_unlock(&dev_priv->tile.lock);
+	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
 
 	return found;
 }
@@ -169,8 +169,9 @@ nv50_mem_vm_bind_linear(struct drm_device *dev, uint64_t virt, uint32_t size,
 			virt  += (end - pte);
 
 			while (pte < end) {
-				nv_wo32(dev, pgt, pte++, offset_l);
-				nv_wo32(dev, pgt, pte++, offset_h);
+				nv_wo32(pgt, (pte * 4) + 0, offset_l);
+				nv_wo32(pgt, (pte * 4) + 4, offset_h);
+				pte += 2;
 			}
 		}
 	}
@@ -203,8 +204,10 @@ nv50_mem_vm_unbind(struct drm_device *dev, uint64_t virt, uint32_t size)
 		pages -= (end - pte);
 		virt  += (end - pte) << 15;
 
-		while (pte < end)
-			nv_wo32(dev, pgt, pte++, 0);
+		while (pte < end) {
+			nv_wo32(pgt, (pte * 4), 0);
+			pte++;
+		}
 	}
 	dev_priv->engine.instmem.flush(dev);
 
@@ -218,7 +221,7 @@ nv50_mem_vm_unbind(struct drm_device *dev, uint64_t virt, uint32_t size)
  * Cleanup everything
  */
 void
-nouveau_mem_close(struct drm_device *dev)
+nouveau_mem_vram_fini(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 
@@ -229,6 +232,19 @@ nouveau_mem_close(struct drm_device *dev)
 
 	nouveau_ttm_global_release(dev_priv);
 
+	if (dev_priv->fb_mtrr >= 0) {
+		drm_mtrr_del(dev_priv->fb_mtrr,
+			     pci_resource_start(dev->pdev, 1),
+			     pci_resource_len(dev->pdev, 1), DRM_MTRR_WC);
+		dev_priv->fb_mtrr = -1;
+	}
+}
+
+void
+nouveau_mem_gart_fini(struct drm_device *dev)
+{
+	nouveau_sgdma_takedown(dev);
+
 	if (drm_core_has_AGP(dev) && dev->agp) {
 		struct drm_agp_mem *entry, *tempe;
 
@@ -248,13 +264,6 @@ nouveau_mem_close(struct drm_device *dev)
 		dev->agp->acquired = 0;
 		dev->agp->enabled = 0;
 	}
-
-	if (dev_priv->fb_mtrr) {
-		drm_mtrr_del(dev_priv->fb_mtrr,
-			     pci_resource_start(dev->pdev, 1),
-			     pci_resource_len(dev->pdev, 1), DRM_MTRR_WC);
-		dev_priv->fb_mtrr = -1;
-	}
 }
 
 static uint32_t
@@ -305,8 +314,62 @@ nouveau_mem_detect_nforce(struct drm_device *dev)
 	return 0;
 }
 
-/* returns the amount of FB ram in bytes */
-int
+static void
+nv50_vram_preinit(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	int i, parts, colbits, rowbitsa, rowbitsb, banks;
+	u64 rowsize, predicted;
+	u32 r0, r4, rt, ru;
+
+	r0 = nv_rd32(dev, 0x100200);
+	r4 = nv_rd32(dev, 0x100204);
+	rt = nv_rd32(dev, 0x100250);
+	ru = nv_rd32(dev, 0x001540);
+	NV_DEBUG(dev, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt, ru);
+
+	for (i = 0, parts = 0; i < 8; i++) {
+		if (ru & (0x00010000 << i))
+			parts++;
+	}
+
+	colbits  =  (r4 & 0x0000f000) >> 12;
+	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
+	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
+	banks    = ((r4 & 0x01000000) ? 8 : 4);
+
+	rowsize = parts * banks * (1 << colbits) * 8;
+	predicted = rowsize << rowbitsa;
+	if (r0 & 0x00000004)
+		predicted += rowsize << rowbitsb;
+
+	if (predicted != dev_priv->vram_size) {
+		NV_WARN(dev, "memory controller reports %dMiB VRAM\n",
+			(u32)(dev_priv->vram_size >> 20));
+		NV_WARN(dev, "we calculated %dMiB VRAM\n",
+			(u32)(predicted >> 20));
+	}
+
+	dev_priv->vram_rblock_size = rowsize >> 12;
+	if (rt & 1)
+		dev_priv->vram_rblock_size *= 3;
+
+	NV_DEBUG(dev, "rblock %lld bytes\n",
+		 (u64)dev_priv->vram_rblock_size << 12);
+}
+
+static void
+nvaa_vram_preinit(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	/* To our knowledge, there's no large scale reordering of pages
+	 * that occurs on IGP chipsets.
+	 */
+	dev_priv->vram_rblock_size = 1;
+}
+
+static int
 nouveau_mem_detect(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
@@ -325,9 +388,18 @@ nouveau_mem_detect(struct drm_device *dev)
 		dev_priv->vram_size = nv_rd32(dev, NV04_PFB_FIFO_DATA);
 		dev_priv->vram_size |= (dev_priv->vram_size & 0xff) << 32;
 		dev_priv->vram_size &= 0xffffffff00ll;
-		if (dev_priv->chipset == 0xaa || dev_priv->chipset == 0xac) {
+
+		switch (dev_priv->chipset) {
+		case 0xaa:
+		case 0xac:
+		case 0xaf:
 			dev_priv->vram_sys_base = nv_rd32(dev, 0x100e10);
 			dev_priv->vram_sys_base <<= 12;
+			nvaa_vram_preinit(dev);
+			break;
+		default:
+			nv50_vram_preinit(dev);
+			break;
 		}
 	} else {
 		dev_priv->vram_size  = nv_rd32(dev, 0x10f20c) << 20;
@@ -345,6 +417,33 @@ nouveau_mem_detect(struct drm_device *dev)
 	return -ENOMEM;
 }
 
+#if __OS_HAS_AGP
+static unsigned long
+get_agp_mode(struct drm_device *dev, unsigned long mode)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	/*
+	 * FW seems to be broken on nv18, it makes the card lock up
+	 * randomly.
+	 */
+	if (dev_priv->chipset == 0x18)
+		mode &= ~PCI_AGP_COMMAND_FW;
+
+	/*
+	 * AGP mode set in the command line.
+	 */
+	if (nouveau_agpmode > 0) {
+		bool agpv3 = mode & 0x8;
+		int rate = agpv3 ? nouveau_agpmode / 4 : nouveau_agpmode;
+
+		mode = (mode & ~0x7) | (rate & 0x7);
+	}
+
+	return mode;
+}
+#endif
+
 int
 nouveau_mem_reset_agp(struct drm_device *dev)
 {
@@ -355,7 +454,8 @@ nouveau_mem_reset_agp(struct drm_device *dev)
 	/* First of all, disable fast writes, otherwise if it's
 	 * already enabled in the AGP bridge and we disable the card's
 	 * AGP controller we might be locking ourselves out of it. */
-	if (nv_rd32(dev, NV04_PBUS_PCI_NV_19) & PCI_AGP_COMMAND_FW) {
+	if ((nv_rd32(dev, NV04_PBUS_PCI_NV_19) |
+	     dev->agp->mode) & PCI_AGP_COMMAND_FW) {
 		struct drm_agp_info info;
 		struct drm_agp_mode mode;
 
@@ -363,7 +463,7 @@ nouveau_mem_reset_agp(struct drm_device *dev)
 		if (ret)
 			return ret;
 
-		mode.mode = info.mode & ~PCI_AGP_COMMAND_FW;
+		mode.mode = get_agp_mode(dev, info.mode) & ~PCI_AGP_COMMAND_FW;
 		ret = drm_agp_enable(dev, mode);
 		if (ret)
 			return ret;
@@ -418,7 +518,7 @@ nouveau_mem_init_agp(struct drm_device *dev)
 	}
 
 	/* see agp.h for the AGPSTAT_* modes available */
-	mode.mode = info.mode;
+	mode.mode = get_agp_mode(dev, info.mode);
 	ret = drm_agp_enable(dev, mode);
 	if (ret) {
 		NV_ERROR(dev, "Unable to enable AGP: %d\n", ret);
@@ -433,24 +533,27 @@ nouveau_mem_init_agp(struct drm_device *dev)
 }
 
 int
-nouveau_mem_init(struct drm_device *dev)
+nouveau_mem_vram_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct ttm_bo_device *bdev = &dev_priv->ttm.bdev;
-	int ret, dma_bits = 32;
-
-	dev_priv->fb_phys = pci_resource_start(dev->pdev, 1);
-	dev_priv->gart_info.type = NOUVEAU_GART_NONE;
+	int ret, dma_bits;
 
 	if (dev_priv->card_type >= NV_50 &&
 	    pci_dma_supported(dev->pdev, DMA_BIT_MASK(40)))
 		dma_bits = 40;
+	else
+		dma_bits = 32;
 
 	ret = pci_set_dma_mask(dev->pdev, DMA_BIT_MASK(dma_bits));
-	if (ret) {
-		NV_ERROR(dev, "Error setting DMA mask: %d\n", ret);
+	if (ret)
 		return ret;
-	}
+
+	ret = nouveau_mem_detect(dev);
+	if (ret)
+		return ret;
+
+	dev_priv->fb_phys = pci_resource_start(dev->pdev, 1);
 
 	ret = nouveau_ttm_global_init(dev_priv);
 	if (ret)
@@ -465,8 +568,6 @@ nouveau_mem_init(struct drm_device *dev)
 		return ret;
 	}
 
-	spin_lock_init(&dev_priv->tile.lock);
-
 	dev_priv->fb_available_size = dev_priv->vram_size;
 	dev_priv->fb_mappable_pages = dev_priv->fb_available_size;
 	if (dev_priv->fb_mappable_pages > pci_resource_len(dev->pdev, 1))
@@ -474,7 +575,16 @@ nouveau_mem_init(struct drm_device *dev)
 			pci_resource_len(dev->pdev, 1);
 	dev_priv->fb_mappable_pages >>= PAGE_SHIFT;
 
-	/* remove reserved space at end of vram from available amount */
+	/* reserve space at end of VRAM for PRAMIN */
+	if (dev_priv->chipset == 0x40 || dev_priv->chipset == 0x47 ||
+	    dev_priv->chipset == 0x49 || dev_priv->chipset == 0x4b)
+		dev_priv->ramin_rsvd_vram = (2 * 1024 * 1024);
+	else
+	if (dev_priv->card_type >= NV_40)
+		dev_priv->ramin_rsvd_vram = (1 * 1024 * 1024);
+	else
+		dev_priv->ramin_rsvd_vram = (512 * 1024);
+
 	dev_priv->fb_available_size -= dev_priv->ramin_rsvd_vram;
 	dev_priv->fb_aper_free = dev_priv->fb_available_size;
 
@@ -495,9 +605,23 @@ nouveau_mem_init(struct drm_device *dev)
 		nouveau_bo_ref(NULL, &dev_priv->vga_ram);
 	}
 
-	/* GART */
+	dev_priv->fb_mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 1),
+					 pci_resource_len(dev->pdev, 1),
+					 DRM_MTRR_WC);
+	return 0;
+}
+
+int
+nouveau_mem_gart_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct ttm_bo_device *bdev = &dev_priv->ttm.bdev;
+	int ret;
+
+	dev_priv->gart_info.type = NOUVEAU_GART_NONE;
+
 #if !defined(__powerpc__) && !defined(__ia64__)
-	if (drm_device_is_agp(dev) && dev->agp && !nouveau_noagp) {
+	if (drm_device_is_agp(dev) && dev->agp && nouveau_agpmode) {
 		ret = nouveau_mem_init_agp(dev);
 		if (ret)
 			NV_ERROR(dev, "Error initialising AGP: %d\n", ret);
@@ -523,11 +647,150 @@ nouveau_mem_init(struct drm_device *dev)
 		return ret;
 	}
 
-	dev_priv->fb_mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 1),
-					 pci_resource_len(dev->pdev, 1),
-					 DRM_MTRR_WC);
-
 	return 0;
 }
 
+void
+nouveau_mem_timing_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_memtimings *memtimings = &pm->memtimings;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct bit_entry P;
+	u8 tUNK_0, tUNK_1, tUNK_2;
+	u8 tRP;		/* Byte 3 */
+	u8 tRAS;	/* Byte 5 */
+	u8 tRFC;	/* Byte 7 */
+	u8 tRC;		/* Byte 9 */
+	u8 tUNK_10, tUNK_11, tUNK_12, tUNK_13, tUNK_14;
+	u8 tUNK_18, tUNK_19, tUNK_20, tUNK_21;
+	u8 *mem = NULL, *entry;
+	int i, recordlen, entries;
 
+	if (bios->type == NVBIOS_BIT) {
+		if (bit_table(dev, 'P', &P))
+			return;
+
+		if (P.version == 1)
+			mem = ROMPTR(bios, P.data[4]);
+		else
+		if (P.version == 2)
+			mem = ROMPTR(bios, P.data[8]);
+		else {
+			NV_WARN(dev, "unknown mem for BIT P %d\n", P.version);
+		}
+	} else {
+		NV_DEBUG(dev, "BMP version too old for memory\n");
+		return;
+	}
+
+	if (!mem) {
+		NV_DEBUG(dev, "memory timing table pointer invalid\n");
+		return;
+	}
+
+	if (mem[0] != 0x10) {
+		NV_WARN(dev, "memory timing table 0x%02x unknown\n", mem[0]);
+		return;
+	}
+
+	/* validate record length */
+	entries   = mem[2];
+	recordlen = mem[3];
+	if (recordlen < 15) {
+		NV_ERROR(dev, "mem timing table length unknown: %d\n", mem[3]);
+		return;
+	}
+
+	/* parse vbios entries into common format */
+	memtimings->timing =
+		kcalloc(entries, sizeof(*memtimings->timing), GFP_KERNEL);
+	if (!memtimings->timing)
+		return;
+
+	entry = mem + mem[1];
+	for (i = 0; i < entries; i++, entry += recordlen) {
+		struct nouveau_pm_memtiming *timing = &pm->memtimings.timing[i];
+		if (entry[0] == 0)
+			continue;
+
+		tUNK_18 = 1;
+		tUNK_19 = 1;
+		tUNK_20 = 0;
+		tUNK_21 = 0;
+		switch (MIN(recordlen,21)) {
+		case 21:
+			tUNK_21 = entry[21];
+		case 20:
+			tUNK_20 = entry[20];
+		case 19:
+			tUNK_19 = entry[19];
+		case 18:
+			tUNK_18 = entry[18];
+		default:
+			tUNK_0  = entry[0];
+			tUNK_1  = entry[1];
+			tUNK_2  = entry[2];
+			tRP     = entry[3];
+			tRAS    = entry[5];
+			tRFC    = entry[7];
+			tRC     = entry[9];
+			tUNK_10 = entry[10];
+			tUNK_11 = entry[11];
+			tUNK_12 = entry[12];
+			tUNK_13 = entry[13];
+			tUNK_14 = entry[14];
+			break;
+		}
+
+		timing->reg_100220 = (tRC << 24 | tRFC << 16 | tRAS << 8 | tRP);
+
+		/* XXX: I don't trust the -1's and +1's... they must come
+		 *      from somewhere! */
+		timing->reg_100224 = ((tUNK_0 + tUNK_19 + 1) << 24 |
+				      tUNK_18 << 16 |
+				      (tUNK_1 + tUNK_19 + 1) << 8 |
+				      (tUNK_2 - 1));
+
+		timing->reg_100228 = (tUNK_12 << 16 | tUNK_11 << 8 | tUNK_10);
+		if(recordlen > 19) {
+			timing->reg_100228 += (tUNK_19 - 1) << 24;
+		} else {
+			timing->reg_100228 += tUNK_12 << 24;
+		}
+
+		/* XXX: reg_10022c */
+
+		timing->reg_100230 = (tUNK_20 << 24 | tUNK_21 << 16 |
+				      tUNK_13 << 8  | tUNK_13);
+
+		/* XXX: +6? */
+		timing->reg_100234 = (tRAS << 24 | (tUNK_19 + 6) << 8 | tRC);
+		if(tUNK_10 > tUNK_11) {
+			timing->reg_100234 += tUNK_10 << 16;
+		} else {
+			timing->reg_100234 += tUNK_11 << 16;
+		}
+
+		/* XXX; reg_100238, reg_10023c */
+		NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x %08x\n", i,
+			 timing->reg_100220, timing->reg_100224,
+			 timing->reg_100228, timing->reg_10022c);
+		NV_DEBUG(dev, "         230: %08x %08x %08x %08x\n",
+			 timing->reg_100230, timing->reg_100234,
+			 timing->reg_100238, timing->reg_10023c);
+	}
+
+	memtimings->nr_timing  = entries;
+	memtimings->supported = true;
+}
+
+void
+nouveau_mem_timing_fini(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_memtimings *mem = &dev_priv->engine.pm.memtimings;
+
+	kfree(mem->timing);
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_notifier.c b/drivers/gpu/drm/nouveau/nouveau_notifier.c
index 3c9964a8fbad..d670839cb34d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_notifier.c
+++ b/drivers/gpu/drm/nouveau/nouveau_notifier.c
@@ -28,6 +28,7 @@
 #include "drmP.h"
 #include "drm.h"
 #include "nouveau_drv.h"
+#include "nouveau_ramht.h"
 
 int
 nouveau_notifier_init_channel(struct nouveau_channel *chan)
@@ -113,7 +114,7 @@ nouveau_notifier_alloc(struct nouveau_channel *chan, uint32_t handle,
 		return -ENOMEM;
 	}
 
-	offset = chan->notifier_bo->bo.mem.mm_node->start << PAGE_SHIFT;
+	offset = chan->notifier_bo->bo.mem.start << PAGE_SHIFT;
 	if (chan->notifier_bo->bo.mem.mem_type == TTM_PL_VRAM) {
 		target = NV_DMA_TARGET_VIDMEM;
 	} else
@@ -147,11 +148,11 @@ nouveau_notifier_alloc(struct nouveau_channel *chan, uint32_t handle,
 	nobj->dtor = nouveau_notifier_gpuobj_dtor;
 	nobj->priv = mem;
 
-	ret = nouveau_gpuobj_ref_add(dev, chan, handle, nobj, NULL);
+	ret = nouveau_ramht_insert(chan, handle, nobj);
+	nouveau_gpuobj_ref(NULL, &nobj);
 	if (ret) {
-		nouveau_gpuobj_del(dev, &nobj);
 		drm_mm_put_block(mem);
-		NV_ERROR(dev, "Error referencing notifier ctxdma: %d\n", ret);
+		NV_ERROR(dev, "Error adding notifier to ramht: %d\n", ret);
 		return ret;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_object.c b/drivers/gpu/drm/nouveau/nouveau_object.c
index b6bcb254f4ab..896cf8634144 100644
--- a/drivers/gpu/drm/nouveau/nouveau_object.c
+++ b/drivers/gpu/drm/nouveau/nouveau_object.c
@@ -34,6 +34,7 @@
 #include "drm.h"
 #include "nouveau_drv.h"
 #include "nouveau_drm.h"
+#include "nouveau_ramht.h"
 
 /* NVidia uses context objects to drive drawing operations.
 
@@ -65,137 +66,6 @@
    The key into the hash table depends on the object handle and channel id and
    is given as:
 */
-static uint32_t
-nouveau_ramht_hash_handle(struct drm_device *dev, int channel, uint32_t handle)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	uint32_t hash = 0;
-	int i;
-
-	NV_DEBUG(dev, "ch%d handle=0x%08x\n", channel, handle);
-
-	for (i = 32; i > 0; i -= dev_priv->ramht_bits) {
-		hash ^= (handle & ((1 << dev_priv->ramht_bits) - 1));
-		handle >>= dev_priv->ramht_bits;
-	}
-
-	if (dev_priv->card_type < NV_50)
-		hash ^= channel << (dev_priv->ramht_bits - 4);
-	hash <<= 3;
-
-	NV_DEBUG(dev, "hash=0x%08x\n", hash);
-	return hash;
-}
-
-static int
-nouveau_ramht_entry_valid(struct drm_device *dev, struct nouveau_gpuobj *ramht,
-			  uint32_t offset)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	uint32_t ctx = nv_ro32(dev, ramht, (offset + 4)/4);
-
-	if (dev_priv->card_type < NV_40)
-		return ((ctx & NV_RAMHT_CONTEXT_VALID) != 0);
-	return (ctx != 0);
-}
-
-static int
-nouveau_ramht_insert(struct drm_device *dev, struct nouveau_gpuobj_ref *ref)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_instmem_engine *instmem = &dev_priv->engine.instmem;
-	struct nouveau_channel *chan = ref->channel;
-	struct nouveau_gpuobj *ramht = chan->ramht ? chan->ramht->gpuobj : NULL;
-	uint32_t ctx, co, ho;
-
-	if (!ramht) {
-		NV_ERROR(dev, "No hash table!\n");
-		return -EINVAL;
-	}
-
-	if (dev_priv->card_type < NV_40) {
-		ctx = NV_RAMHT_CONTEXT_VALID | (ref->instance >> 4) |
-		      (chan->id << NV_RAMHT_CONTEXT_CHANNEL_SHIFT) |
-		      (ref->gpuobj->engine << NV_RAMHT_CONTEXT_ENGINE_SHIFT);
-	} else
-	if (dev_priv->card_type < NV_50) {
-		ctx = (ref->instance >> 4) |
-		      (chan->id << NV40_RAMHT_CONTEXT_CHANNEL_SHIFT) |
-		      (ref->gpuobj->engine << NV40_RAMHT_CONTEXT_ENGINE_SHIFT);
-	} else {
-		if (ref->gpuobj->engine == NVOBJ_ENGINE_DISPLAY) {
-			ctx = (ref->instance << 10) | 2;
-		} else {
-			ctx = (ref->instance >> 4) |
-			      ((ref->gpuobj->engine <<
-				NV40_RAMHT_CONTEXT_ENGINE_SHIFT));
-		}
-	}
-
-	co = ho = nouveau_ramht_hash_handle(dev, chan->id, ref->handle);
-	do {
-		if (!nouveau_ramht_entry_valid(dev, ramht, co)) {
-			NV_DEBUG(dev,
-				 "insert ch%d 0x%08x: h=0x%08x, c=0x%08x\n",
-				 chan->id, co, ref->handle, ctx);
-			nv_wo32(dev, ramht, (co + 0)/4, ref->handle);
-			nv_wo32(dev, ramht, (co + 4)/4, ctx);
-
-			list_add_tail(&ref->list, &chan->ramht_refs);
-			instmem->flush(dev);
-			return 0;
-		}
-		NV_DEBUG(dev, "collision ch%d 0x%08x: h=0x%08x\n",
-			 chan->id, co, nv_ro32(dev, ramht, co/4));
-
-		co += 8;
-		if (co >= dev_priv->ramht_size)
-			co = 0;
-	} while (co != ho);
-
-	NV_ERROR(dev, "RAMHT space exhausted. ch=%d\n", chan->id);
-	return -ENOMEM;
-}
-
-static void
-nouveau_ramht_remove(struct drm_device *dev, struct nouveau_gpuobj_ref *ref)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_instmem_engine *instmem = &dev_priv->engine.instmem;
-	struct nouveau_channel *chan = ref->channel;
-	struct nouveau_gpuobj *ramht = chan->ramht ? chan->ramht->gpuobj : NULL;
-	uint32_t co, ho;
-
-	if (!ramht) {
-		NV_ERROR(dev, "No hash table!\n");
-		return;
-	}
-
-	co = ho = nouveau_ramht_hash_handle(dev, chan->id, ref->handle);
-	do {
-		if (nouveau_ramht_entry_valid(dev, ramht, co) &&
-		    (ref->handle == nv_ro32(dev, ramht, (co/4)))) {
-			NV_DEBUG(dev,
-				 "remove ch%d 0x%08x: h=0x%08x, c=0x%08x\n",
-				 chan->id, co, ref->handle,
-				 nv_ro32(dev, ramht, (co + 4)));
-			nv_wo32(dev, ramht, (co + 0)/4, 0x00000000);
-			nv_wo32(dev, ramht, (co + 4)/4, 0x00000000);
-
-			list_del(&ref->list);
-			instmem->flush(dev);
-			return;
-		}
-
-		co += 8;
-		if (co >= dev_priv->ramht_size)
-			co = 0;
-	} while (co != ho);
-	list_del(&ref->list);
-
-	NV_ERROR(dev, "RAMHT entry not found. ch=%d, handle=0x%08x\n",
-		 chan->id, ref->handle);
-}
 
 int
 nouveau_gpuobj_new(struct drm_device *dev, struct nouveau_channel *chan,
@@ -205,7 +75,7 @@ nouveau_gpuobj_new(struct drm_device *dev, struct nouveau_channel *chan,
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_engine *engine = &dev_priv->engine;
 	struct nouveau_gpuobj *gpuobj;
-	struct drm_mm *pramin = NULL;
+	struct drm_mm_node *ramin = NULL;
 	int ret;
 
 	NV_DEBUG(dev, "ch%d size=%u align=%d flags=0x%08x\n",
@@ -218,88 +88,115 @@ nouveau_gpuobj_new(struct drm_device *dev, struct nouveau_channel *chan,
 	if (!gpuobj)
 		return -ENOMEM;
 	NV_DEBUG(dev, "gpuobj %p\n", gpuobj);
+	gpuobj->dev = dev;
 	gpuobj->flags = flags;
-	gpuobj->im_channel = chan;
+	kref_init(&gpuobj->refcount);
+	gpuobj->size = size;
 
+	spin_lock(&dev_priv->ramin_lock);
 	list_add_tail(&gpuobj->list, &dev_priv->gpuobj_list);
+	spin_unlock(&dev_priv->ramin_lock);
 
-	/* Choose between global instmem heap, and per-channel private
-	 * instmem heap.  On <NV50 allow requests for private instmem
-	 * to be satisfied from global heap if no per-channel area
-	 * available.
-	 */
 	if (chan) {
 		NV_DEBUG(dev, "channel heap\n");
-		pramin = &chan->ramin_heap;
+
+		ramin = drm_mm_search_free(&chan->ramin_heap, size, align, 0);
+		if (ramin)
+			ramin = drm_mm_get_block(ramin, size, align);
+
+		if (!ramin) {
+			nouveau_gpuobj_ref(NULL, &gpuobj);
+			return -ENOMEM;
+		}
 	} else {
 		NV_DEBUG(dev, "global heap\n");
-		pramin = &dev_priv->ramin_heap;
 
+		/* allocate backing pages, sets vinst */
 		ret = engine->instmem.populate(dev, gpuobj, &size);
 		if (ret) {
-			nouveau_gpuobj_del(dev, &gpuobj);
+			nouveau_gpuobj_ref(NULL, &gpuobj);
 			return ret;
 		}
+
+		/* try and get aperture space */
+		do {
+			if (drm_mm_pre_get(&dev_priv->ramin_heap))
+				return -ENOMEM;
+
+			spin_lock(&dev_priv->ramin_lock);
+			ramin = drm_mm_search_free(&dev_priv->ramin_heap, size,
+						   align, 0);
+			if (ramin == NULL) {
+				spin_unlock(&dev_priv->ramin_lock);
+				nouveau_gpuobj_ref(NULL, &gpuobj);
+				return ret;
+			}
+
+			ramin = drm_mm_get_block_atomic(ramin, size, align);
+			spin_unlock(&dev_priv->ramin_lock);
+		} while (ramin == NULL);
+
+		/* on nv50 it's ok to fail, we have a fallback path */
+		if (!ramin && dev_priv->card_type < NV_50) {
+			nouveau_gpuobj_ref(NULL, &gpuobj);
+			return -ENOMEM;
+		}
 	}
 
-	/* Allocate a chunk of the PRAMIN aperture */
-	gpuobj->im_pramin = drm_mm_search_free(pramin, size, align, 0);
-	if (gpuobj->im_pramin)
-		gpuobj->im_pramin = drm_mm_get_block(gpuobj->im_pramin, size, align);
-
-	if (!gpuobj->im_pramin) {
-		nouveau_gpuobj_del(dev, &gpuobj);
-		return -ENOMEM;
-	}
-
-	if (!chan) {
+	/* if we got a chunk of the aperture, map pages into it */
+	gpuobj->im_pramin = ramin;
+	if (!chan && gpuobj->im_pramin && dev_priv->ramin_available) {
 		ret = engine->instmem.bind(dev, gpuobj);
 		if (ret) {
-			nouveau_gpuobj_del(dev, &gpuobj);
+			nouveau_gpuobj_ref(NULL, &gpuobj);
 			return ret;
 		}
 	}
 
+	/* calculate the various different addresses for the object */
+	if (chan) {
+		gpuobj->pinst = chan->ramin->pinst;
+		if (gpuobj->pinst != ~0)
+			gpuobj->pinst += gpuobj->im_pramin->start;
+
+		if (dev_priv->card_type < NV_50) {
+			gpuobj->cinst = gpuobj->pinst;
+		} else {
+			gpuobj->cinst = gpuobj->im_pramin->start;
+			gpuobj->vinst = gpuobj->im_pramin->start +
+					chan->ramin->vinst;
+		}
+	} else {
+		if (gpuobj->im_pramin)
+			gpuobj->pinst = gpuobj->im_pramin->start;
+		else
+			gpuobj->pinst = ~0;
+		gpuobj->cinst = 0xdeadbeef;
+	}
+
 	if (gpuobj->flags & NVOBJ_FLAG_ZERO_ALLOC) {
 		int i;
 
-		for (i = 0; i < gpuobj->im_pramin->size; i += 4)
-			nv_wo32(dev, gpuobj, i/4, 0);
+		for (i = 0; i < gpuobj->size; i += 4)
+			nv_wo32(gpuobj, i, 0);
 		engine->instmem.flush(dev);
 	}
 
+
 	*gpuobj_ret = gpuobj;
 	return 0;
 }
 
-int
-nouveau_gpuobj_early_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	NV_DEBUG(dev, "\n");
-
-	INIT_LIST_HEAD(&dev_priv->gpuobj_list);
-
-	return 0;
-}
-
 int
 nouveau_gpuobj_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int ret;
 
 	NV_DEBUG(dev, "\n");
 
-	if (dev_priv->card_type < NV_50) {
-		ret = nouveau_gpuobj_new_fake(dev,
-			dev_priv->ramht_offset, ~0, dev_priv->ramht_size,
-			NVOBJ_FLAG_ZERO_ALLOC | NVOBJ_FLAG_ALLOW_NO_REFS,
-						&dev_priv->ramht, NULL);
-		if (ret)
-			return ret;
-	}
+	INIT_LIST_HEAD(&dev_priv->gpuobj_list);
+	spin_lock_init(&dev_priv->ramin_lock);
+	dev_priv->ramin_base = ~0;
 
 	return 0;
 }
@@ -311,297 +208,89 @@ nouveau_gpuobj_takedown(struct drm_device *dev)
 
 	NV_DEBUG(dev, "\n");
 
-	nouveau_gpuobj_del(dev, &dev_priv->ramht);
+	BUG_ON(!list_empty(&dev_priv->gpuobj_list));
 }
 
-void
-nouveau_gpuobj_late_takedown(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *gpuobj = NULL;
-	struct list_head *entry, *tmp;
-
-	NV_DEBUG(dev, "\n");
-
-	list_for_each_safe(entry, tmp, &dev_priv->gpuobj_list) {
-		gpuobj = list_entry(entry, struct nouveau_gpuobj, list);
-
-		NV_ERROR(dev, "gpuobj %p still exists at takedown, refs=%d\n",
-			 gpuobj, gpuobj->refcount);
-		gpuobj->refcount = 0;
-		nouveau_gpuobj_del(dev, &gpuobj);
-	}
-}
-
-int
-nouveau_gpuobj_del(struct drm_device *dev, struct nouveau_gpuobj **pgpuobj)
+
+static void
+nouveau_gpuobj_del(struct kref *ref)
 {
+	struct nouveau_gpuobj *gpuobj =
+		container_of(ref, struct nouveau_gpuobj, refcount);
+	struct drm_device *dev = gpuobj->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_engine *engine = &dev_priv->engine;
-	struct nouveau_gpuobj *gpuobj;
 	int i;
 
-	NV_DEBUG(dev, "gpuobj %p\n", pgpuobj ? *pgpuobj : NULL);
-
-	if (!dev_priv || !pgpuobj || !(*pgpuobj))
-		return -EINVAL;
-	gpuobj = *pgpuobj;
-
-	if (gpuobj->refcount != 0) {
-		NV_ERROR(dev, "gpuobj refcount is %d\n", gpuobj->refcount);
-		return -EINVAL;
-	}
+	NV_DEBUG(dev, "gpuobj %p\n", gpuobj);
 
 	if (gpuobj->im_pramin && (gpuobj->flags & NVOBJ_FLAG_ZERO_FREE)) {
-		for (i = 0; i < gpuobj->im_pramin->size; i += 4)
-			nv_wo32(dev, gpuobj, i/4, 0);
+		for (i = 0; i < gpuobj->size; i += 4)
+			nv_wo32(gpuobj, i, 0);
 		engine->instmem.flush(dev);
 	}
 
 	if (gpuobj->dtor)
 		gpuobj->dtor(dev, gpuobj);
 
-	if (gpuobj->im_backing && !(gpuobj->flags & NVOBJ_FLAG_FAKE))
+	if (gpuobj->im_backing)
 		engine->instmem.clear(dev, gpuobj);
 
-	if (gpuobj->im_pramin) {
-		if (gpuobj->flags & NVOBJ_FLAG_FAKE)
-			kfree(gpuobj->im_pramin);
-		else
-			drm_mm_put_block(gpuobj->im_pramin);
-	}
-
+	spin_lock(&dev_priv->ramin_lock);
+	if (gpuobj->im_pramin)
+		drm_mm_put_block(gpuobj->im_pramin);
 	list_del(&gpuobj->list);
+	spin_unlock(&dev_priv->ramin_lock);
 
-	*pgpuobj = NULL;
 	kfree(gpuobj);
-	return 0;
 }
 
-static int
-nouveau_gpuobj_instance_get(struct drm_device *dev,
-			    struct nouveau_channel *chan,
-			    struct nouveau_gpuobj *gpuobj, uint32_t *inst)
+void
+nouveau_gpuobj_ref(struct nouveau_gpuobj *ref, struct nouveau_gpuobj **ptr)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *cpramin;
+	if (ref)
+		kref_get(&ref->refcount);
 
-	/* <NV50 use PRAMIN address everywhere */
-	if (dev_priv->card_type < NV_50) {
-		*inst = gpuobj->im_pramin->start;
-		return 0;
-	}
+	if (*ptr)
+		kref_put(&(*ptr)->refcount, nouveau_gpuobj_del);
 
-	if (chan && gpuobj->im_channel != chan) {
-		NV_ERROR(dev, "Channel mismatch: obj %d, ref %d\n",
-			 gpuobj->im_channel->id, chan->id);
-		return -EINVAL;
-	}
-
-	/* NV50 channel-local instance */
-	if (chan) {
-		cpramin = chan->ramin->gpuobj;
-		*inst = gpuobj->im_pramin->start - cpramin->im_pramin->start;
-		return 0;
-	}
-
-	/* NV50 global (VRAM) instance */
-	if (!gpuobj->im_channel) {
-		/* ...from global heap */
-		if (!gpuobj->im_backing) {
-			NV_ERROR(dev, "AII, no VRAM backing gpuobj\n");
-			return -EINVAL;
-		}
-		*inst = gpuobj->im_backing_start;
-		return 0;
-	} else {
-		/* ...from local heap */
-		cpramin = gpuobj->im_channel->ramin->gpuobj;
-		*inst = cpramin->im_backing_start +
-			(gpuobj->im_pramin->start - cpramin->im_pramin->start);
-		return 0;
-	}
-
-	return -EINVAL;
+	*ptr = ref;
 }
 
 int
-nouveau_gpuobj_ref_add(struct drm_device *dev, struct nouveau_channel *chan,
-		       uint32_t handle, struct nouveau_gpuobj *gpuobj,
-		       struct nouveau_gpuobj_ref **ref_ret)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj_ref *ref;
-	uint32_t instance;
-	int ret;
-
-	NV_DEBUG(dev, "ch%d h=0x%08x gpuobj=%p\n",
-		 chan ? chan->id : -1, handle, gpuobj);
-
-	if (!dev_priv || !gpuobj || (ref_ret && *ref_ret != NULL))
-		return -EINVAL;
-
-	if (!chan && !ref_ret)
-		return -EINVAL;
-
-	if (gpuobj->engine == NVOBJ_ENGINE_SW && !gpuobj->im_pramin) {
-		/* sw object */
-		instance = 0x40;
-	} else {
-		ret = nouveau_gpuobj_instance_get(dev, chan, gpuobj, &instance);
-		if (ret)
-			return ret;
-	}
-
-	ref = kzalloc(sizeof(*ref), GFP_KERNEL);
-	if (!ref)
-		return -ENOMEM;
-	INIT_LIST_HEAD(&ref->list);
-	ref->gpuobj   = gpuobj;
-	ref->channel  = chan;
-	ref->instance = instance;
-
-	if (!ref_ret) {
-		ref->handle = handle;
-
-		ret = nouveau_ramht_insert(dev, ref);
-		if (ret) {
-			kfree(ref);
-			return ret;
-		}
-	} else {
-		ref->handle = ~0;
-		*ref_ret = ref;
-	}
-
-	ref->gpuobj->refcount++;
-	return 0;
-}
-
-int nouveau_gpuobj_ref_del(struct drm_device *dev, struct nouveau_gpuobj_ref **pref)
-{
-	struct nouveau_gpuobj_ref *ref;
-
-	NV_DEBUG(dev, "ref %p\n", pref ? *pref : NULL);
-
-	if (!dev || !pref || *pref == NULL)
-		return -EINVAL;
-	ref = *pref;
-
-	if (ref->handle != ~0)
-		nouveau_ramht_remove(dev, ref);
-
-	if (ref->gpuobj) {
-		ref->gpuobj->refcount--;
-
-		if (ref->gpuobj->refcount == 0) {
-			if (!(ref->gpuobj->flags & NVOBJ_FLAG_ALLOW_NO_REFS))
-				nouveau_gpuobj_del(dev, &ref->gpuobj);
-		}
-	}
-
-	*pref = NULL;
-	kfree(ref);
-	return 0;
-}
-
-int
-nouveau_gpuobj_new_ref(struct drm_device *dev,
-		       struct nouveau_channel *oc, struct nouveau_channel *rc,
-		       uint32_t handle, uint32_t size, int align,
-		       uint32_t flags, struct nouveau_gpuobj_ref **ref)
-{
-	struct nouveau_gpuobj *gpuobj = NULL;
-	int ret;
-
-	ret = nouveau_gpuobj_new(dev, oc, size, align, flags, &gpuobj);
-	if (ret)
-		return ret;
-
-	ret = nouveau_gpuobj_ref_add(dev, rc, handle, gpuobj, ref);
-	if (ret) {
-		nouveau_gpuobj_del(dev, &gpuobj);
-		return ret;
-	}
-
-	return 0;
-}
-
-int
-nouveau_gpuobj_ref_find(struct nouveau_channel *chan, uint32_t handle,
-			struct nouveau_gpuobj_ref **ref_ret)
-{
-	struct nouveau_gpuobj_ref *ref;
-	struct list_head *entry, *tmp;
-
-	list_for_each_safe(entry, tmp, &chan->ramht_refs) {
-		ref = list_entry(entry, struct nouveau_gpuobj_ref, list);
-
-		if (ref->handle == handle) {
-			if (ref_ret)
-				*ref_ret = ref;
-			return 0;
-		}
-	}
-
-	return -EINVAL;
-}
-
-int
-nouveau_gpuobj_new_fake(struct drm_device *dev, uint32_t p_offset,
-			uint32_t b_offset, uint32_t size,
-			uint32_t flags, struct nouveau_gpuobj **pgpuobj,
-			struct nouveau_gpuobj_ref **pref)
+nouveau_gpuobj_new_fake(struct drm_device *dev, u32 pinst, u64 vinst,
+			u32 size, u32 flags, struct nouveau_gpuobj **pgpuobj)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_gpuobj *gpuobj = NULL;
 	int i;
 
 	NV_DEBUG(dev,
-		 "p_offset=0x%08x b_offset=0x%08x size=0x%08x flags=0x%08x\n",
-		 p_offset, b_offset, size, flags);
+		 "pinst=0x%08x vinst=0x%010llx size=0x%08x flags=0x%08x\n",
+		 pinst, vinst, size, flags);
 
 	gpuobj = kzalloc(sizeof(*gpuobj), GFP_KERNEL);
 	if (!gpuobj)
 		return -ENOMEM;
 	NV_DEBUG(dev, "gpuobj %p\n", gpuobj);
-	gpuobj->im_channel = NULL;
-	gpuobj->flags      = flags | NVOBJ_FLAG_FAKE;
-
-	list_add_tail(&gpuobj->list, &dev_priv->gpuobj_list);
-
-	if (p_offset != ~0) {
-		gpuobj->im_pramin = kzalloc(sizeof(struct drm_mm_node),
-					    GFP_KERNEL);
-		if (!gpuobj->im_pramin) {
-			nouveau_gpuobj_del(dev, &gpuobj);
-			return -ENOMEM;
-		}
-		gpuobj->im_pramin->start = p_offset;
-		gpuobj->im_pramin->size  = size;
-	}
-
-	if (b_offset != ~0) {
-		gpuobj->im_backing = (struct nouveau_bo *)-1;
-		gpuobj->im_backing_start = b_offset;
-	}
+	gpuobj->dev = dev;
+	gpuobj->flags = flags;
+	kref_init(&gpuobj->refcount);
+	gpuobj->size  = size;
+	gpuobj->pinst = pinst;
+	gpuobj->cinst = 0xdeadbeef;
+	gpuobj->vinst = vinst;
 
 	if (gpuobj->flags & NVOBJ_FLAG_ZERO_ALLOC) {
-		for (i = 0; i < gpuobj->im_pramin->size; i += 4)
-			nv_wo32(dev, gpuobj, i/4, 0);
+		for (i = 0; i < gpuobj->size; i += 4)
+			nv_wo32(gpuobj, i, 0);
 		dev_priv->engine.instmem.flush(dev);
 	}
 
-	if (pref) {
-		i = nouveau_gpuobj_ref_add(dev, NULL, 0, gpuobj, pref);
-		if (i) {
-			nouveau_gpuobj_del(dev, &gpuobj);
-			return i;
-		}
-	}
-
-	if (pgpuobj)
-		*pgpuobj = gpuobj;
+	spin_lock(&dev_priv->ramin_lock);
+	list_add_tail(&gpuobj->list, &dev_priv->gpuobj_list);
+	spin_unlock(&dev_priv->ramin_lock);
+	*pgpuobj = gpuobj;
 	return 0;
 }
 
@@ -685,14 +374,12 @@ nouveau_gpuobj_dma_new(struct nouveau_channel *chan, int class,
 		adjust = offset &  0x00000fff;
 		frame  = offset & ~0x00000fff;
 
-		nv_wo32(dev, *gpuobj, 0, ((1<<12) | (1<<13) |
-				(adjust << 20) |
-				 (access << 14) |
-				 (target << 16) |
-				  class));
-		nv_wo32(dev, *gpuobj, 1, size - 1);
-		nv_wo32(dev, *gpuobj, 2, frame | pte_flags);
-		nv_wo32(dev, *gpuobj, 3, frame | pte_flags);
+		nv_wo32(*gpuobj,  0, ((1<<12) | (1<<13) | (adjust << 20) |
+				      (access << 14) | (target << 16) |
+				      class));
+		nv_wo32(*gpuobj,  4, size - 1);
+		nv_wo32(*gpuobj,  8, frame | pte_flags);
+		nv_wo32(*gpuobj, 12, frame | pte_flags);
 	} else {
 		uint64_t limit = offset + size - 1;
 		uint32_t flags0, flags5;
@@ -705,12 +392,12 @@ nouveau_gpuobj_dma_new(struct nouveau_channel *chan, int class,
 			flags5 = 0x00080000;
 		}
 
-		nv_wo32(dev, *gpuobj, 0, flags0 | class);
-		nv_wo32(dev, *gpuobj, 1, lower_32_bits(limit));
-		nv_wo32(dev, *gpuobj, 2, lower_32_bits(offset));
-		nv_wo32(dev, *gpuobj, 3, ((upper_32_bits(limit) & 0xff) << 24) |
-					(upper_32_bits(offset) & 0xff));
-		nv_wo32(dev, *gpuobj, 5, flags5);
+		nv_wo32(*gpuobj,  0, flags0 | class);
+		nv_wo32(*gpuobj,  4, lower_32_bits(limit));
+		nv_wo32(*gpuobj,  8, lower_32_bits(offset));
+		nv_wo32(*gpuobj, 12, ((upper_32_bits(limit) & 0xff) << 24) |
+				      (upper_32_bits(offset) & 0xff));
+		nv_wo32(*gpuobj, 20, flags5);
 	}
 
 	instmem->flush(dev);
@@ -741,7 +428,7 @@ nouveau_gpuobj_gart_dma_new(struct nouveau_channel *chan,
 			*o_ret = 0;
 	} else
 	if (dev_priv->gart_info.type == NOUVEAU_GART_SGDMA) {
-		*gpuobj = dev_priv->gart_info.sg_ctxdma;
+		nouveau_gpuobj_ref(dev_priv->gart_info.sg_ctxdma, gpuobj);
 		if (offset & ~0xffffffffULL) {
 			NV_ERROR(dev, "obj offset exceeds 32-bits\n");
 			return -EINVAL;
@@ -829,25 +516,25 @@ nouveau_gpuobj_gr_new(struct nouveau_channel *chan, int class,
 	}
 
 	if (dev_priv->card_type >= NV_50) {
-		nv_wo32(dev, *gpuobj, 0, class);
-		nv_wo32(dev, *gpuobj, 5, 0x00010000);
+		nv_wo32(*gpuobj,  0, class);
+		nv_wo32(*gpuobj, 20, 0x00010000);
 	} else {
 		switch (class) {
 		case NV_CLASS_NULL:
-			nv_wo32(dev, *gpuobj, 0, 0x00001030);
-			nv_wo32(dev, *gpuobj, 1, 0xFFFFFFFF);
+			nv_wo32(*gpuobj, 0, 0x00001030);
+			nv_wo32(*gpuobj, 4, 0xFFFFFFFF);
 			break;
 		default:
 			if (dev_priv->card_type >= NV_40) {
-				nv_wo32(dev, *gpuobj, 0, class);
+				nv_wo32(*gpuobj, 0, class);
 #ifdef __BIG_ENDIAN
-				nv_wo32(dev, *gpuobj, 2, 0x01000000);
+				nv_wo32(*gpuobj, 8, 0x01000000);
 #endif
 			} else {
 #ifdef __BIG_ENDIAN
-				nv_wo32(dev, *gpuobj, 0, class | 0x00080000);
+				nv_wo32(*gpuobj, 0, class | 0x00080000);
 #else
-				nv_wo32(dev, *gpuobj, 0, class);
+				nv_wo32(*gpuobj, 0, class);
 #endif
 			}
 		}
@@ -873,10 +560,15 @@ nouveau_gpuobj_sw_new(struct nouveau_channel *chan, int class,
 	gpuobj = kzalloc(sizeof(*gpuobj), GFP_KERNEL);
 	if (!gpuobj)
 		return -ENOMEM;
+	gpuobj->dev = chan->dev;
 	gpuobj->engine = NVOBJ_ENGINE_SW;
 	gpuobj->class = class;
+	kref_init(&gpuobj->refcount);
+	gpuobj->cinst = 0x40;
 
+	spin_lock(&dev_priv->ramin_lock);
 	list_add_tail(&gpuobj->list, &dev_priv->gpuobj_list);
+	spin_unlock(&dev_priv->ramin_lock);
 	*gpuobj_ret = gpuobj;
 	return 0;
 }
@@ -886,7 +578,6 @@ nouveau_gpuobj_channel_init_pramin(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *pramin = NULL;
 	uint32_t size;
 	uint32_t base;
 	int ret;
@@ -911,18 +602,16 @@ nouveau_gpuobj_channel_init_pramin(struct nouveau_channel *chan)
 		size += 0x1000;
 	}
 
-	ret = nouveau_gpuobj_new_ref(dev, NULL, NULL, 0, size, 0x1000, 0,
-				     &chan->ramin);
+	ret = nouveau_gpuobj_new(dev, NULL, size, 0x1000, 0, &chan->ramin);
 	if (ret) {
 		NV_ERROR(dev, "Error allocating channel PRAMIN: %d\n", ret);
 		return ret;
 	}
-	pramin = chan->ramin->gpuobj;
 
-	ret = drm_mm_init(&chan->ramin_heap, pramin->im_pramin->start + base, size);
+	ret = drm_mm_init(&chan->ramin_heap, base, size);
 	if (ret) {
 		NV_ERROR(dev, "Error creating PRAMIN heap: %d\n", ret);
-		nouveau_gpuobj_ref_del(dev, &chan->ramin);
+		nouveau_gpuobj_ref(NULL, &chan->ramin);
 		return ret;
 	}
 
@@ -939,8 +628,6 @@ nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 	struct nouveau_gpuobj *vram = NULL, *tt = NULL;
 	int ret, i;
 
-	INIT_LIST_HEAD(&chan->ramht_refs);
-
 	NV_DEBUG(dev, "ch%d vram=0x%08x tt=0x%08x\n", chan->id, vram_h, tt_h);
 
 	/* Allocate a chunk of memory for per-channel object storage */
@@ -956,41 +643,38 @@ nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 	 *    locations determined during init.
 	 */
 	if (dev_priv->card_type >= NV_50) {
-		uint32_t vm_offset, pde;
+		u32 pgd_offs = (dev_priv->chipset == 0x50) ? 0x1400 : 0x0200;
+		u64 vm_vinst = chan->ramin->vinst + pgd_offs;
+		u32 vm_pinst = chan->ramin->pinst;
+		u32 pde;
 
-		vm_offset = (dev_priv->chipset & 0xf0) == 0x50 ? 0x1400 : 0x200;
-		vm_offset += chan->ramin->gpuobj->im_pramin->start;
+		if (vm_pinst != ~0)
+			vm_pinst += pgd_offs;
 
-		ret = nouveau_gpuobj_new_fake(dev, vm_offset, ~0, 0x4000,
-							0, &chan->vm_pd, NULL);
+		ret = nouveau_gpuobj_new_fake(dev, vm_pinst, vm_vinst, 0x4000,
+					      0, &chan->vm_pd);
 		if (ret)
 			return ret;
 		for (i = 0; i < 0x4000; i += 8) {
-			nv_wo32(dev, chan->vm_pd, (i+0)/4, 0x00000000);
-			nv_wo32(dev, chan->vm_pd, (i+4)/4, 0xdeadcafe);
+			nv_wo32(chan->vm_pd, i + 0, 0x00000000);
+			nv_wo32(chan->vm_pd, i + 4, 0xdeadcafe);
 		}
 
-		pde = (dev_priv->vm_gart_base / (512*1024*1024)) * 2;
-		ret = nouveau_gpuobj_ref_add(dev, NULL, 0,
-					     dev_priv->gart_info.sg_ctxdma,
-					     &chan->vm_gart_pt);
-		if (ret)
-			return ret;
-		nv_wo32(dev, chan->vm_pd, pde++,
-			    chan->vm_gart_pt->instance | 0x03);
-		nv_wo32(dev, chan->vm_pd, pde++, 0x00000000);
+		nouveau_gpuobj_ref(dev_priv->gart_info.sg_ctxdma,
+				   &chan->vm_gart_pt);
+		pde = (dev_priv->vm_gart_base / (512*1024*1024)) * 8;
+		nv_wo32(chan->vm_pd, pde + 0, chan->vm_gart_pt->vinst | 3);
+		nv_wo32(chan->vm_pd, pde + 4, 0x00000000);
 
-		pde = (dev_priv->vm_vram_base / (512*1024*1024)) * 2;
+		pde = (dev_priv->vm_vram_base / (512*1024*1024)) * 8;
 		for (i = 0; i < dev_priv->vm_vram_pt_nr; i++) {
-			ret = nouveau_gpuobj_ref_add(dev, NULL, 0,
-						     dev_priv->vm_vram_pt[i],
-						     &chan->vm_vram_pt[i]);
-			if (ret)
-				return ret;
+			nouveau_gpuobj_ref(dev_priv->vm_vram_pt[i],
+					   &chan->vm_vram_pt[i]);
 
-			nv_wo32(dev, chan->vm_pd, pde++,
-				    chan->vm_vram_pt[i]->instance | 0x61);
-			nv_wo32(dev, chan->vm_pd, pde++, 0x00000000);
+			nv_wo32(chan->vm_pd, pde + 0,
+				chan->vm_vram_pt[i]->vinst | 0x61);
+			nv_wo32(chan->vm_pd, pde + 4, 0x00000000);
+			pde += 8;
 		}
 
 		instmem->flush(dev);
@@ -998,15 +682,17 @@ nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 
 	/* RAMHT */
 	if (dev_priv->card_type < NV_50) {
-		ret = nouveau_gpuobj_ref_add(dev, NULL, 0, dev_priv->ramht,
-					     &chan->ramht);
+		nouveau_ramht_ref(dev_priv->ramht, &chan->ramht, NULL);
+	} else {
+		struct nouveau_gpuobj *ramht = NULL;
+
+		ret = nouveau_gpuobj_new(dev, chan, 0x8000, 16,
+					 NVOBJ_FLAG_ZERO_ALLOC, &ramht);
 		if (ret)
 			return ret;
-	} else {
-		ret = nouveau_gpuobj_new_ref(dev, chan, chan, 0,
-					     0x8000, 16,
-					     NVOBJ_FLAG_ZERO_ALLOC,
-					     &chan->ramht);
+
+		ret = nouveau_ramht_new(dev, ramht, &chan->ramht);
+		nouveau_gpuobj_ref(NULL, &ramht);
 		if (ret)
 			return ret;
 	}
@@ -1023,24 +709,32 @@ nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 		}
 	} else {
 		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY,
-						0, dev_priv->fb_available_size,
-						NV_DMA_ACCESS_RW,
-						NV_DMA_TARGET_VIDMEM, &vram);
+					     0, dev_priv->fb_available_size,
+					     NV_DMA_ACCESS_RW,
+					     NV_DMA_TARGET_VIDMEM, &vram);
 		if (ret) {
 			NV_ERROR(dev, "Error creating VRAM ctxdma: %d\n", ret);
 			return ret;
 		}
 	}
 
-	ret = nouveau_gpuobj_ref_add(dev, chan, vram_h, vram, NULL);
+	ret = nouveau_ramht_insert(chan, vram_h, vram);
+	nouveau_gpuobj_ref(NULL, &vram);
 	if (ret) {
-		NV_ERROR(dev, "Error referencing VRAM ctxdma: %d\n", ret);
+		NV_ERROR(dev, "Error adding VRAM ctxdma to RAMHT: %d\n", ret);
 		return ret;
 	}
 
 	/* TT memory ctxdma */
 	if (dev_priv->card_type >= NV_50) {
-		tt = vram;
+		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY,
+					     0, dev_priv->vm_end,
+					     NV_DMA_ACCESS_RW,
+					     NV_DMA_TARGET_AGP, &tt);
+		if (ret) {
+			NV_ERROR(dev, "Error creating VRAM ctxdma: %d\n", ret);
+			return ret;
+		}
 	} else
 	if (dev_priv->gart_info.type != NOUVEAU_GART_NONE) {
 		ret = nouveau_gpuobj_gart_dma_new(chan, 0,
@@ -1056,9 +750,10 @@ nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 		return ret;
 	}
 
-	ret = nouveau_gpuobj_ref_add(dev, chan, tt_h, tt, NULL);
+	ret = nouveau_ramht_insert(chan, tt_h, tt);
+	nouveau_gpuobj_ref(NULL, &tt);
 	if (ret) {
-		NV_ERROR(dev, "Error referencing TT ctxdma: %d\n", ret);
+		NV_ERROR(dev, "Error adding TT ctxdma to RAMHT: %d\n", ret);
 		return ret;
 	}
 
@@ -1070,33 +765,23 @@ nouveau_gpuobj_channel_takedown(struct nouveau_channel *chan)
 {
 	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
 	struct drm_device *dev = chan->dev;
-	struct list_head *entry, *tmp;
-	struct nouveau_gpuobj_ref *ref;
 	int i;
 
 	NV_DEBUG(dev, "ch%d\n", chan->id);
 
-	if (!chan->ramht_refs.next)
+	if (!chan->ramht)
 		return;
 
-	list_for_each_safe(entry, tmp, &chan->ramht_refs) {
-		ref = list_entry(entry, struct nouveau_gpuobj_ref, list);
+	nouveau_ramht_ref(NULL, &chan->ramht, chan);
 
-		nouveau_gpuobj_ref_del(dev, &ref);
-	}
-
-	nouveau_gpuobj_ref_del(dev, &chan->ramht);
-
-	nouveau_gpuobj_del(dev, &chan->vm_pd);
-	nouveau_gpuobj_ref_del(dev, &chan->vm_gart_pt);
+	nouveau_gpuobj_ref(NULL, &chan->vm_pd);
+	nouveau_gpuobj_ref(NULL, &chan->vm_gart_pt);
 	for (i = 0; i < dev_priv->vm_vram_pt_nr; i++)
-		nouveau_gpuobj_ref_del(dev, &chan->vm_vram_pt[i]);
+		nouveau_gpuobj_ref(NULL, &chan->vm_vram_pt[i]);
 
 	if (chan->ramin_heap.free_stack.next)
 		drm_mm_takedown(&chan->ramin_heap);
-	if (chan->ramin)
-		nouveau_gpuobj_ref_del(dev, &chan->ramin);
-
+	nouveau_gpuobj_ref(NULL, &chan->ramin);
 }
 
 int
@@ -1117,17 +802,17 @@ nouveau_gpuobj_suspend(struct drm_device *dev)
 	}
 
 	list_for_each_entry(gpuobj, &dev_priv->gpuobj_list, list) {
-		if (!gpuobj->im_backing || (gpuobj->flags & NVOBJ_FLAG_FAKE))
+		if (!gpuobj->im_backing)
 			continue;
 
-		gpuobj->im_backing_suspend = vmalloc(gpuobj->im_pramin->size);
+		gpuobj->im_backing_suspend = vmalloc(gpuobj->size);
 		if (!gpuobj->im_backing_suspend) {
 			nouveau_gpuobj_resume(dev);
 			return -ENOMEM;
 		}
 
-		for (i = 0; i < gpuobj->im_pramin->size / 4; i++)
-			gpuobj->im_backing_suspend[i] = nv_ro32(dev, gpuobj, i);
+		for (i = 0; i < gpuobj->size; i += 4)
+			gpuobj->im_backing_suspend[i/4] = nv_ro32(gpuobj, i);
 	}
 
 	return 0;
@@ -1172,8 +857,8 @@ nouveau_gpuobj_resume(struct drm_device *dev)
 		if (!gpuobj->im_backing_suspend)
 			continue;
 
-		for (i = 0; i < gpuobj->im_pramin->size / 4; i++)
-			nv_wo32(dev, gpuobj, i, gpuobj->im_backing_suspend[i]);
+		for (i = 0; i < gpuobj->size; i += 4)
+			nv_wo32(gpuobj, i, gpuobj->im_backing_suspend[i/4]);
 		dev_priv->engine.instmem.flush(dev);
 	}
 
@@ -1208,25 +893,24 @@ int nouveau_ioctl_grobj_alloc(struct drm_device *dev, void *data,
 		return -EPERM;
 	}
 
-	if (nouveau_gpuobj_ref_find(chan, init->handle, NULL) == 0)
+	if (nouveau_ramht_find(chan, init->handle))
 		return -EEXIST;
 
 	if (!grc->software)
 		ret = nouveau_gpuobj_gr_new(chan, grc->id, &gr);
 	else
 		ret = nouveau_gpuobj_sw_new(chan, grc->id, &gr);
-
 	if (ret) {
 		NV_ERROR(dev, "Error creating object: %d (%d/0x%08x)\n",
 			 ret, init->channel, init->handle);
 		return ret;
 	}
 
-	ret = nouveau_gpuobj_ref_add(dev, chan, init->handle, gr, NULL);
+	ret = nouveau_ramht_insert(chan, init->handle, gr);
+	nouveau_gpuobj_ref(NULL, &gr);
 	if (ret) {
 		NV_ERROR(dev, "Error referencing object: %d (%d/0x%08x)\n",
 			 ret, init->channel, init->handle);
-		nouveau_gpuobj_del(dev, &gr);
 		return ret;
 	}
 
@@ -1237,16 +921,62 @@ int nouveau_ioctl_gpuobj_free(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv)
 {
 	struct drm_nouveau_gpuobj_free *objfree = data;
-	struct nouveau_gpuobj_ref *ref;
+	struct nouveau_gpuobj *gpuobj;
 	struct nouveau_channel *chan;
-	int ret;
 
 	NOUVEAU_GET_USER_CHANNEL_WITH_RETURN(objfree->channel, file_priv, chan);
 
-	ret = nouveau_gpuobj_ref_find(chan, objfree->handle, &ref);
-	if (ret)
-		return ret;
-	nouveau_gpuobj_ref_del(dev, &ref);
+	gpuobj = nouveau_ramht_find(chan, objfree->handle);
+	if (!gpuobj)
+		return -ENOENT;
 
+	nouveau_ramht_remove(chan, objfree->handle);
 	return 0;
 }
+
+u32
+nv_ro32(struct nouveau_gpuobj *gpuobj, u32 offset)
+{
+	struct drm_nouveau_private *dev_priv = gpuobj->dev->dev_private;
+	struct drm_device *dev = gpuobj->dev;
+
+	if (gpuobj->pinst == ~0 || !dev_priv->ramin_available) {
+		u64  ptr = gpuobj->vinst + offset;
+		u32 base = ptr >> 16;
+		u32  val;
+
+		spin_lock(&dev_priv->ramin_lock);
+		if (dev_priv->ramin_base != base) {
+			dev_priv->ramin_base = base;
+			nv_wr32(dev, 0x001700, dev_priv->ramin_base);
+		}
+		val = nv_rd32(dev, 0x700000 + (ptr & 0xffff));
+		spin_unlock(&dev_priv->ramin_lock);
+		return val;
+	}
+
+	return nv_ri32(dev, gpuobj->pinst + offset);
+}
+
+void
+nv_wo32(struct nouveau_gpuobj *gpuobj, u32 offset, u32 val)
+{
+	struct drm_nouveau_private *dev_priv = gpuobj->dev->dev_private;
+	struct drm_device *dev = gpuobj->dev;
+
+	if (gpuobj->pinst == ~0 || !dev_priv->ramin_available) {
+		u64  ptr = gpuobj->vinst + offset;
+		u32 base = ptr >> 16;
+
+		spin_lock(&dev_priv->ramin_lock);
+		if (dev_priv->ramin_base != base) {
+			dev_priv->ramin_base = base;
+			nv_wr32(dev, 0x001700, dev_priv->ramin_base);
+		}
+		nv_wr32(dev, 0x700000 + (ptr & 0xffff), val);
+		spin_unlock(&dev_priv->ramin_lock);
+		return;
+	}
+
+	nv_wi32(dev, gpuobj->pinst + offset, val);
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_perf.c b/drivers/gpu/drm/nouveau/nouveau_perf.c
new file mode 100644
index 000000000000..ac62a1b8c4fc
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_perf.c
@@ -0,0 +1,205 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "drmP.h"
+
+#include "nouveau_drv.h"
+#include "nouveau_pm.h"
+
+static void
+legacy_perf_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	char *perf, *entry, *bmp = &bios->data[bios->offset];
+	int headerlen, use_straps;
+
+	if (bmp[5] < 0x5 || bmp[6] < 0x14) {
+		NV_DEBUG(dev, "BMP version too old for perf\n");
+		return;
+	}
+
+	perf = ROMPTR(bios, bmp[0x73]);
+	if (!perf) {
+		NV_DEBUG(dev, "No memclock table pointer found.\n");
+		return;
+	}
+
+	switch (perf[0]) {
+	case 0x12:
+	case 0x14:
+	case 0x18:
+		use_straps = 0;
+		headerlen = 1;
+		break;
+	case 0x01:
+		use_straps = perf[1] & 1;
+		headerlen = (use_straps ? 8 : 2);
+		break;
+	default:
+		NV_WARN(dev, "Unknown memclock table version %x.\n", perf[0]);
+		return;
+	}
+
+	entry = perf + headerlen;
+	if (use_straps)
+		entry += (nv_rd32(dev, NV_PEXTDEV_BOOT_0) & 0x3c) >> 1;
+
+	sprintf(pm->perflvl[0].name, "performance_level_0");
+	pm->perflvl[0].memory = ROM16(entry[0]) * 20;
+	pm->nr_perflvl = 1;
+}
+
+void
+nouveau_perf_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct bit_entry P;
+	u8 version, headerlen, recordlen, entries;
+	u8 *perf, *entry;
+	int vid, i;
+
+	if (bios->type == NVBIOS_BIT) {
+		if (bit_table(dev, 'P', &P))
+			return;
+
+		if (P.version != 1 && P.version != 2) {
+			NV_WARN(dev, "unknown perf for BIT P %d\n", P.version);
+			return;
+		}
+
+		perf = ROMPTR(bios, P.data[0]);
+		version   = perf[0];
+		headerlen = perf[1];
+		if (version < 0x40) {
+			recordlen = perf[3] + (perf[4] * perf[5]);
+			entries   = perf[2];
+		} else {
+			recordlen = perf[2] + (perf[3] * perf[4]);
+			entries   = perf[5];
+		}
+	} else {
+		if (bios->data[bios->offset + 6] < 0x25) {
+			legacy_perf_init(dev);
+			return;
+		}
+
+		perf = ROMPTR(bios, bios->data[bios->offset + 0x94]);
+		if (!perf) {
+			NV_DEBUG(dev, "perf table pointer invalid\n");
+			return;
+		}
+
+		version   = perf[1];
+		headerlen = perf[0];
+		recordlen = perf[3];
+		entries   = perf[2];
+	}
+
+	entry = perf + headerlen;
+	for (i = 0; i < entries; i++) {
+		struct nouveau_pm_level *perflvl = &pm->perflvl[pm->nr_perflvl];
+
+		if (entry[0] == 0xff) {
+			entry += recordlen;
+			continue;
+		}
+
+		switch (version) {
+		case 0x12:
+		case 0x13:
+		case 0x15:
+			perflvl->fanspeed = entry[55];
+			perflvl->voltage = entry[56];
+			perflvl->core = ROM32(entry[1]) * 10;
+			perflvl->memory = ROM32(entry[5]) * 20;
+			break;
+		case 0x21:
+		case 0x23:
+		case 0x24:
+			perflvl->fanspeed = entry[4];
+			perflvl->voltage = entry[5];
+			perflvl->core = ROM16(entry[6]) * 1000;
+
+			if (dev_priv->chipset == 0x49 ||
+			    dev_priv->chipset == 0x4b)
+				perflvl->memory = ROM16(entry[11]) * 1000;
+			else
+				perflvl->memory = ROM16(entry[11]) * 2000;
+
+			break;
+		case 0x25:
+			perflvl->fanspeed = entry[4];
+			perflvl->voltage = entry[5];
+			perflvl->core = ROM16(entry[6]) * 1000;
+			perflvl->shader = ROM16(entry[10]) * 1000;
+			perflvl->memory = ROM16(entry[12]) * 1000;
+			break;
+		case 0x30:
+			perflvl->memscript = ROM16(entry[2]);
+		case 0x35:
+			perflvl->fanspeed = entry[6];
+			perflvl->voltage = entry[7];
+			perflvl->core = ROM16(entry[8]) * 1000;
+			perflvl->shader = ROM16(entry[10]) * 1000;
+			perflvl->memory = ROM16(entry[12]) * 1000;
+			/*XXX: confirm on 0x35 */
+			perflvl->unk05 = ROM16(entry[16]) * 1000;
+			break;
+		case 0x40:
+#define subent(n) entry[perf[2] + ((n) * perf[3])]
+			perflvl->fanspeed = 0; /*XXX*/
+			perflvl->voltage = entry[2];
+			perflvl->core = (ROM16(subent(0)) & 0xfff) * 1000;
+			perflvl->shader = (ROM16(subent(1)) & 0xfff) * 1000;
+			perflvl->memory = (ROM16(subent(2)) & 0xfff) * 1000;
+			break;
+		}
+
+		/* make sure vid is valid */
+		if (pm->voltage.supported && perflvl->voltage) {
+			vid = nouveau_volt_vid_lookup(dev, perflvl->voltage);
+			if (vid < 0) {
+				NV_DEBUG(dev, "drop perflvl %d, bad vid\n", i);
+				entry += recordlen;
+				continue;
+			}
+		}
+
+		snprintf(perflvl->name, sizeof(perflvl->name),
+			 "performance_level_%d", i);
+		perflvl->id = i;
+		pm->nr_perflvl++;
+
+		entry += recordlen;
+	}
+}
+
+void
+nouveau_perf_fini(struct drm_device *dev)
+{
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.c b/drivers/gpu/drm/nouveau/nouveau_pm.c
new file mode 100644
index 000000000000..1c99c55d6d46
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.c
@@ -0,0 +1,518 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "drmP.h"
+
+#include "nouveau_drv.h"
+#include "nouveau_pm.h"
+
+#include <linux/hwmon.h>
+#include <linux/hwmon-sysfs.h>
+
+static int
+nouveau_pm_clock_set(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+		     u8 id, u32 khz)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	void *pre_state;
+
+	if (khz == 0)
+		return 0;
+
+	pre_state = pm->clock_pre(dev, perflvl, id, khz);
+	if (IS_ERR(pre_state))
+		return PTR_ERR(pre_state);
+
+	if (pre_state)
+		pm->clock_set(dev, pre_state);
+	return 0;
+}
+
+static int
+nouveau_pm_perflvl_set(struct drm_device *dev, struct nouveau_pm_level *perflvl)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	int ret;
+
+	if (perflvl == pm->cur)
+		return 0;
+
+	if (pm->voltage.supported && pm->voltage_set && perflvl->voltage) {
+		ret = pm->voltage_set(dev, perflvl->voltage);
+		if (ret) {
+			NV_ERROR(dev, "voltage_set %d failed: %d\n",
+				 perflvl->voltage, ret);
+		}
+	}
+
+	nouveau_pm_clock_set(dev, perflvl, PLL_CORE, perflvl->core);
+	nouveau_pm_clock_set(dev, perflvl, PLL_SHADER, perflvl->shader);
+	nouveau_pm_clock_set(dev, perflvl, PLL_MEMORY, perflvl->memory);
+	nouveau_pm_clock_set(dev, perflvl, PLL_UNK05, perflvl->unk05);
+
+	pm->cur = perflvl;
+	return 0;
+}
+
+static int
+nouveau_pm_profile_set(struct drm_device *dev, const char *profile)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_level *perflvl = NULL;
+
+	/* safety precaution, for now */
+	if (nouveau_perflvl_wr != 7777)
+		return -EPERM;
+
+	if (!pm->clock_set)
+		return -EINVAL;
+
+	if (!strncmp(profile, "boot", 4))
+		perflvl = &pm->boot;
+	else {
+		int pl = simple_strtol(profile, NULL, 10);
+		int i;
+
+		for (i = 0; i < pm->nr_perflvl; i++) {
+			if (pm->perflvl[i].id == pl) {
+				perflvl = &pm->perflvl[i];
+				break;
+			}
+		}
+
+		if (!perflvl)
+			return -EINVAL;
+	}
+
+	NV_INFO(dev, "setting performance level: %s\n", profile);
+	return nouveau_pm_perflvl_set(dev, perflvl);
+}
+
+static int
+nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	int ret;
+
+	if (!pm->clock_get)
+		return -EINVAL;
+
+	memset(perflvl, 0, sizeof(*perflvl));
+
+	ret = pm->clock_get(dev, PLL_CORE);
+	if (ret > 0)
+		perflvl->core = ret;
+
+	ret = pm->clock_get(dev, PLL_MEMORY);
+	if (ret > 0)
+		perflvl->memory = ret;
+
+	ret = pm->clock_get(dev, PLL_SHADER);
+	if (ret > 0)
+		perflvl->shader = ret;
+
+	ret = pm->clock_get(dev, PLL_UNK05);
+	if (ret > 0)
+		perflvl->unk05 = ret;
+
+	if (pm->voltage.supported && pm->voltage_get) {
+		ret = pm->voltage_get(dev);
+		if (ret > 0)
+			perflvl->voltage = ret;
+	}
+
+	return 0;
+}
+
+static void
+nouveau_pm_perflvl_info(struct nouveau_pm_level *perflvl, char *ptr, int len)
+{
+	char c[16], s[16], v[16], f[16];
+
+	c[0] = '\0';
+	if (perflvl->core)
+		snprintf(c, sizeof(c), " core %dMHz", perflvl->core / 1000);
+
+	s[0] = '\0';
+	if (perflvl->shader)
+		snprintf(s, sizeof(s), " shader %dMHz", perflvl->shader / 1000);
+
+	v[0] = '\0';
+	if (perflvl->voltage)
+		snprintf(v, sizeof(v), " voltage %dmV", perflvl->voltage * 10);
+
+	f[0] = '\0';
+	if (perflvl->fanspeed)
+		snprintf(f, sizeof(f), " fanspeed %d%%", perflvl->fanspeed);
+
+	snprintf(ptr, len, "memory %dMHz%s%s%s%s\n", perflvl->memory / 1000,
+		 c, s, v, f);
+}
+
+static ssize_t
+nouveau_pm_get_perflvl_info(struct device *d,
+			    struct device_attribute *a, char *buf)
+{
+	struct nouveau_pm_level *perflvl = (struct nouveau_pm_level *)a;
+	char *ptr = buf;
+	int len = PAGE_SIZE;
+
+	snprintf(ptr, len, "%d: ", perflvl->id);
+	ptr += strlen(buf);
+	len -= strlen(buf);
+
+	nouveau_pm_perflvl_info(perflvl, ptr, len);
+	return strlen(buf);
+}
+
+static ssize_t
+nouveau_pm_get_perflvl(struct device *d, struct device_attribute *a, char *buf)
+{
+	struct drm_device *dev = pci_get_drvdata(to_pci_dev(d));
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_level cur;
+	int len = PAGE_SIZE, ret;
+	char *ptr = buf;
+
+	if (!pm->cur)
+		snprintf(ptr, len, "setting: boot\n");
+	else if (pm->cur == &pm->boot)
+		snprintf(ptr, len, "setting: boot\nc: ");
+	else
+		snprintf(ptr, len, "setting: static %d\nc: ", pm->cur->id);
+	ptr += strlen(buf);
+	len -= strlen(buf);
+
+	ret = nouveau_pm_perflvl_get(dev, &cur);
+	if (ret == 0)
+		nouveau_pm_perflvl_info(&cur, ptr, len);
+	return strlen(buf);
+}
+
+static ssize_t
+nouveau_pm_set_perflvl(struct device *d, struct device_attribute *a,
+		       const char *buf, size_t count)
+{
+	struct drm_device *dev = pci_get_drvdata(to_pci_dev(d));
+	int ret;
+
+	ret = nouveau_pm_profile_set(dev, buf);
+	if (ret)
+		return ret;
+	return strlen(buf);
+}
+
+static DEVICE_ATTR(performance_level, S_IRUGO | S_IWUSR,
+		   nouveau_pm_get_perflvl, nouveau_pm_set_perflvl);
+
+static int
+nouveau_sysfs_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct device *d = &dev->pdev->dev;
+	int ret, i;
+
+	ret = device_create_file(d, &dev_attr_performance_level);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < pm->nr_perflvl; i++) {
+		struct nouveau_pm_level *perflvl = &pm->perflvl[i];
+
+		perflvl->dev_attr.attr.name = perflvl->name;
+		perflvl->dev_attr.attr.mode = S_IRUGO;
+		perflvl->dev_attr.show = nouveau_pm_get_perflvl_info;
+		perflvl->dev_attr.store = NULL;
+		sysfs_attr_init(&perflvl->dev_attr.attr);
+
+		ret = device_create_file(d, &perflvl->dev_attr);
+		if (ret) {
+			NV_ERROR(dev, "failed pervlvl %d sysfs: %d\n",
+				 perflvl->id, i);
+			perflvl->dev_attr.attr.name = NULL;
+			nouveau_pm_fini(dev);
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static void
+nouveau_sysfs_fini(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct device *d = &dev->pdev->dev;
+	int i;
+
+	device_remove_file(d, &dev_attr_performance_level);
+	for (i = 0; i < pm->nr_perflvl; i++) {
+		struct nouveau_pm_level *pl = &pm->perflvl[i];
+
+		if (!pl->dev_attr.attr.name)
+			break;
+
+		device_remove_file(d, &pl->dev_attr);
+	}
+}
+
+static ssize_t
+nouveau_hwmon_show_temp(struct device *d, struct device_attribute *a, char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", pm->temp_get(dev)*1000);
+}
+static SENSOR_DEVICE_ATTR(temp1_input, S_IRUGO, nouveau_hwmon_show_temp,
+						  NULL, 0);
+
+static ssize_t
+nouveau_hwmon_max_temp(struct device *d, struct device_attribute *a, char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp;
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", temp->down_clock*1000);
+}
+static ssize_t
+nouveau_hwmon_set_max_temp(struct device *d, struct device_attribute *a,
+						const char *buf, size_t count)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp;
+	long value;
+
+	if (strict_strtol(buf, 10, &value) == -EINVAL)
+		return count;
+
+	temp->down_clock = value/1000;
+
+	nouveau_temp_safety_checks(dev);
+
+	return count;
+}
+static SENSOR_DEVICE_ATTR(temp1_max, S_IRUGO | S_IWUSR, nouveau_hwmon_max_temp,
+						  nouveau_hwmon_set_max_temp,
+						  0);
+
+static ssize_t
+nouveau_hwmon_critical_temp(struct device *d, struct device_attribute *a,
+							char *buf)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp;
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", temp->critical*1000);
+}
+static ssize_t
+nouveau_hwmon_set_critical_temp(struct device *d, struct device_attribute *a,
+							    const char *buf,
+								size_t count)
+{
+	struct drm_device *dev = dev_get_drvdata(d);
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp;
+	long value;
+
+	if (strict_strtol(buf, 10, &value) == -EINVAL)
+		return count;
+
+	temp->critical = value/1000;
+
+	nouveau_temp_safety_checks(dev);
+
+	return count;
+}
+static SENSOR_DEVICE_ATTR(temp1_crit, S_IRUGO | S_IWUSR,
+						nouveau_hwmon_critical_temp,
+						nouveau_hwmon_set_critical_temp,
+						0);
+
+static ssize_t nouveau_hwmon_show_name(struct device *dev,
+				      struct device_attribute *attr,
+				      char *buf)
+{
+	return sprintf(buf, "nouveau\n");
+}
+static SENSOR_DEVICE_ATTR(name, S_IRUGO, nouveau_hwmon_show_name, NULL, 0);
+
+static ssize_t nouveau_hwmon_show_update_rate(struct device *dev,
+				      struct device_attribute *attr,
+				      char *buf)
+{
+	return sprintf(buf, "1000\n");
+}
+static SENSOR_DEVICE_ATTR(update_rate, S_IRUGO,
+						nouveau_hwmon_show_update_rate,
+						NULL, 0);
+
+static struct attribute *hwmon_attributes[] = {
+	&sensor_dev_attr_temp1_input.dev_attr.attr,
+	&sensor_dev_attr_temp1_max.dev_attr.attr,
+	&sensor_dev_attr_temp1_crit.dev_attr.attr,
+	&sensor_dev_attr_name.dev_attr.attr,
+	&sensor_dev_attr_update_rate.dev_attr.attr,
+	NULL
+};
+
+static const struct attribute_group hwmon_attrgroup = {
+	.attrs = hwmon_attributes,
+};
+
+static int
+nouveau_hwmon_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct device *hwmon_dev;
+	int ret;
+
+	if (!pm->temp_get)
+		return -ENODEV;
+
+	hwmon_dev = hwmon_device_register(&dev->pdev->dev);
+	if (IS_ERR(hwmon_dev)) {
+		ret = PTR_ERR(hwmon_dev);
+		NV_ERROR(dev,
+			"Unable to register hwmon device: %d\n", ret);
+		return ret;
+	}
+	dev_set_drvdata(hwmon_dev, dev);
+	ret = sysfs_create_group(&hwmon_dev->kobj,
+					&hwmon_attrgroup);
+	if (ret) {
+		NV_ERROR(dev,
+			"Unable to create hwmon sysfs file: %d\n", ret);
+		hwmon_device_unregister(hwmon_dev);
+		return ret;
+	}
+
+	pm->hwmon = hwmon_dev;
+
+	return 0;
+}
+
+static void
+nouveau_hwmon_fini(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+
+	if (pm->hwmon) {
+		sysfs_remove_group(&pm->hwmon->kobj, &hwmon_attrgroup);
+		hwmon_device_unregister(pm->hwmon);
+	}
+}
+
+int
+nouveau_pm_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	char info[256];
+	int ret, i;
+
+	nouveau_volt_init(dev);
+	nouveau_perf_init(dev);
+	nouveau_temp_init(dev);
+	nouveau_mem_timing_init(dev);
+
+	NV_INFO(dev, "%d available performance level(s)\n", pm->nr_perflvl);
+	for (i = 0; i < pm->nr_perflvl; i++) {
+		nouveau_pm_perflvl_info(&pm->perflvl[i], info, sizeof(info));
+		NV_INFO(dev, "%d: %s", pm->perflvl[i].id, info);
+	}
+
+	/* determine current ("boot") performance level */
+	ret = nouveau_pm_perflvl_get(dev, &pm->boot);
+	if (ret == 0) {
+		pm->cur = &pm->boot;
+
+		nouveau_pm_perflvl_info(&pm->boot, info, sizeof(info));
+		NV_INFO(dev, "c: %s", info);
+	}
+
+	/* switch performance levels now if requested */
+	if (nouveau_perflvl != NULL) {
+		ret = nouveau_pm_profile_set(dev, nouveau_perflvl);
+		if (ret) {
+			NV_ERROR(dev, "error setting perflvl \"%s\": %d\n",
+				 nouveau_perflvl, ret);
+		}
+	}
+
+	nouveau_sysfs_init(dev);
+	nouveau_hwmon_init(dev);
+
+	return 0;
+}
+
+void
+nouveau_pm_fini(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+
+	if (pm->cur != &pm->boot)
+		nouveau_pm_perflvl_set(dev, &pm->boot);
+
+	nouveau_mem_timing_fini(dev);
+	nouveau_temp_fini(dev);
+	nouveau_perf_fini(dev);
+	nouveau_volt_fini(dev);
+
+	nouveau_hwmon_fini(dev);
+	nouveau_sysfs_fini(dev);
+}
+
+void
+nouveau_pm_resume(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_level *perflvl;
+
+	if (pm->cur == &pm->boot)
+		return;
+
+	perflvl = pm->cur;
+	pm->cur = &pm->boot;
+	nouveau_pm_perflvl_set(dev, perflvl);
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.h b/drivers/gpu/drm/nouveau/nouveau_pm.h
new file mode 100644
index 000000000000..4a9838ddacec
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.h
@@ -0,0 +1,74 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#ifndef __NOUVEAU_PM_H__
+#define __NOUVEAU_PM_H__
+
+/* nouveau_pm.c */
+int  nouveau_pm_init(struct drm_device *dev);
+void nouveau_pm_fini(struct drm_device *dev);
+void nouveau_pm_resume(struct drm_device *dev);
+
+/* nouveau_volt.c */
+void nouveau_volt_init(struct drm_device *);
+void nouveau_volt_fini(struct drm_device *);
+int  nouveau_volt_vid_lookup(struct drm_device *, int voltage);
+int  nouveau_volt_lvl_lookup(struct drm_device *, int vid);
+int  nouveau_voltage_gpio_get(struct drm_device *);
+int  nouveau_voltage_gpio_set(struct drm_device *, int voltage);
+
+/* nouveau_perf.c */
+void nouveau_perf_init(struct drm_device *);
+void nouveau_perf_fini(struct drm_device *);
+
+/* nouveau_mem.c */
+void nouveau_mem_timing_init(struct drm_device *);
+void nouveau_mem_timing_fini(struct drm_device *);
+
+/* nv04_pm.c */
+int nv04_pm_clock_get(struct drm_device *, u32 id);
+void *nv04_pm_clock_pre(struct drm_device *, struct nouveau_pm_level *,
+			u32 id, int khz);
+void nv04_pm_clock_set(struct drm_device *, void *);
+
+/* nv50_pm.c */
+int nv50_pm_clock_get(struct drm_device *, u32 id);
+void *nv50_pm_clock_pre(struct drm_device *, struct nouveau_pm_level *,
+			u32 id, int khz);
+void nv50_pm_clock_set(struct drm_device *, void *);
+
+/* nva3_pm.c */
+int nva3_pm_clock_get(struct drm_device *, u32 id);
+void *nva3_pm_clock_pre(struct drm_device *, struct nouveau_pm_level *,
+			u32 id, int khz);
+void nva3_pm_clock_set(struct drm_device *, void *);
+
+/* nouveau_temp.c */
+void nouveau_temp_init(struct drm_device *dev);
+void nouveau_temp_fini(struct drm_device *dev);
+void nouveau_temp_safety_checks(struct drm_device *dev);
+int nv40_temp_get(struct drm_device *dev);
+int nv84_temp_get(struct drm_device *dev);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_ramht.c b/drivers/gpu/drm/nouveau/nouveau_ramht.c
new file mode 100644
index 000000000000..7f16697cc96c
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_ramht.c
@@ -0,0 +1,289 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "drmP.h"
+
+#include "nouveau_drv.h"
+#include "nouveau_ramht.h"
+
+static u32
+nouveau_ramht_hash_handle(struct nouveau_channel *chan, u32 handle)
+{
+	struct drm_device *dev = chan->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_ramht *ramht = chan->ramht;
+	u32 hash = 0;
+	int i;
+
+	NV_DEBUG(dev, "ch%d handle=0x%08x\n", chan->id, handle);
+
+	for (i = 32; i > 0; i -= ramht->bits) {
+		hash ^= (handle & ((1 << ramht->bits) - 1));
+		handle >>= ramht->bits;
+	}
+
+	if (dev_priv->card_type < NV_50)
+		hash ^= chan->id << (ramht->bits - 4);
+	hash <<= 3;
+
+	NV_DEBUG(dev, "hash=0x%08x\n", hash);
+	return hash;
+}
+
+static int
+nouveau_ramht_entry_valid(struct drm_device *dev, struct nouveau_gpuobj *ramht,
+			  u32 offset)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 ctx = nv_ro32(ramht, offset + 4);
+
+	if (dev_priv->card_type < NV_40)
+		return ((ctx & NV_RAMHT_CONTEXT_VALID) != 0);
+	return (ctx != 0);
+}
+
+static int
+nouveau_ramht_entry_same_channel(struct nouveau_channel *chan,
+				 struct nouveau_gpuobj *ramht, u32 offset)
+{
+	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
+	u32 ctx = nv_ro32(ramht, offset + 4);
+
+	if (dev_priv->card_type >= NV_50)
+		return true;
+	else if (dev_priv->card_type >= NV_40)
+		return chan->id ==
+			((ctx >> NV40_RAMHT_CONTEXT_CHANNEL_SHIFT) & 0x1f);
+	else
+		return chan->id ==
+			((ctx >> NV_RAMHT_CONTEXT_CHANNEL_SHIFT) & 0x1f);
+}
+
+int
+nouveau_ramht_insert(struct nouveau_channel *chan, u32 handle,
+		     struct nouveau_gpuobj *gpuobj)
+{
+	struct drm_device *dev = chan->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_instmem_engine *instmem = &dev_priv->engine.instmem;
+	struct nouveau_ramht_entry *entry;
+	struct nouveau_gpuobj *ramht = chan->ramht->gpuobj;
+	unsigned long flags;
+	u32 ctx, co, ho;
+
+	if (nouveau_ramht_find(chan, handle))
+		return -EEXIST;
+
+	entry = kmalloc(sizeof(*entry), GFP_KERNEL);
+	if (!entry)
+		return -ENOMEM;
+	entry->channel = chan;
+	entry->gpuobj = NULL;
+	entry->handle = handle;
+	nouveau_gpuobj_ref(gpuobj, &entry->gpuobj);
+
+	if (dev_priv->card_type < NV_40) {
+		ctx = NV_RAMHT_CONTEXT_VALID | (gpuobj->cinst >> 4) |
+		      (chan->id << NV_RAMHT_CONTEXT_CHANNEL_SHIFT) |
+		      (gpuobj->engine << NV_RAMHT_CONTEXT_ENGINE_SHIFT);
+	} else
+	if (dev_priv->card_type < NV_50) {
+		ctx = (gpuobj->cinst >> 4) |
+		      (chan->id << NV40_RAMHT_CONTEXT_CHANNEL_SHIFT) |
+		      (gpuobj->engine << NV40_RAMHT_CONTEXT_ENGINE_SHIFT);
+	} else {
+		if (gpuobj->engine == NVOBJ_ENGINE_DISPLAY) {
+			ctx = (gpuobj->cinst << 10) | 2;
+		} else {
+			ctx = (gpuobj->cinst >> 4) |
+			      ((gpuobj->engine <<
+				NV40_RAMHT_CONTEXT_ENGINE_SHIFT));
+		}
+	}
+
+	spin_lock_irqsave(&chan->ramht->lock, flags);
+	list_add(&entry->head, &chan->ramht->entries);
+
+	co = ho = nouveau_ramht_hash_handle(chan, handle);
+	do {
+		if (!nouveau_ramht_entry_valid(dev, ramht, co)) {
+			NV_DEBUG(dev,
+				 "insert ch%d 0x%08x: h=0x%08x, c=0x%08x\n",
+				 chan->id, co, handle, ctx);
+			nv_wo32(ramht, co + 0, handle);
+			nv_wo32(ramht, co + 4, ctx);
+
+			spin_unlock_irqrestore(&chan->ramht->lock, flags);
+			instmem->flush(dev);
+			return 0;
+		}
+		NV_DEBUG(dev, "collision ch%d 0x%08x: h=0x%08x\n",
+			 chan->id, co, nv_ro32(ramht, co));
+
+		co += 8;
+		if (co >= ramht->size)
+			co = 0;
+	} while (co != ho);
+
+	NV_ERROR(dev, "RAMHT space exhausted. ch=%d\n", chan->id);
+	list_del(&entry->head);
+	spin_unlock_irqrestore(&chan->ramht->lock, flags);
+	kfree(entry);
+	return -ENOMEM;
+}
+
+static void
+nouveau_ramht_remove_locked(struct nouveau_channel *chan, u32 handle)
+{
+	struct drm_device *dev = chan->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_instmem_engine *instmem = &dev_priv->engine.instmem;
+	struct nouveau_gpuobj *ramht = chan->ramht->gpuobj;
+	struct nouveau_ramht_entry *entry, *tmp;
+	u32 co, ho;
+
+	list_for_each_entry_safe(entry, tmp, &chan->ramht->entries, head) {
+		if (entry->channel != chan || entry->handle != handle)
+			continue;
+
+		nouveau_gpuobj_ref(NULL, &entry->gpuobj);
+		list_del(&entry->head);
+		kfree(entry);
+		break;
+	}
+
+	co = ho = nouveau_ramht_hash_handle(chan, handle);
+	do {
+		if (nouveau_ramht_entry_valid(dev, ramht, co) &&
+		    nouveau_ramht_entry_same_channel(chan, ramht, co) &&
+		    (handle == nv_ro32(ramht, co))) {
+			NV_DEBUG(dev,
+				 "remove ch%d 0x%08x: h=0x%08x, c=0x%08x\n",
+				 chan->id, co, handle, nv_ro32(ramht, co + 4));
+			nv_wo32(ramht, co + 0, 0x00000000);
+			nv_wo32(ramht, co + 4, 0x00000000);
+			instmem->flush(dev);
+			return;
+		}
+
+		co += 8;
+		if (co >= ramht->size)
+			co = 0;
+	} while (co != ho);
+
+	NV_ERROR(dev, "RAMHT entry not found. ch=%d, handle=0x%08x\n",
+		 chan->id, handle);
+}
+
+void
+nouveau_ramht_remove(struct nouveau_channel *chan, u32 handle)
+{
+	struct nouveau_ramht *ramht = chan->ramht;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ramht->lock, flags);
+	nouveau_ramht_remove_locked(chan, handle);
+	spin_unlock_irqrestore(&ramht->lock, flags);
+}
+
+struct nouveau_gpuobj *
+nouveau_ramht_find(struct nouveau_channel *chan, u32 handle)
+{
+	struct nouveau_ramht *ramht = chan->ramht;
+	struct nouveau_ramht_entry *entry;
+	struct nouveau_gpuobj *gpuobj = NULL;
+	unsigned long flags;
+
+	if (unlikely(!chan->ramht))
+		return NULL;
+
+	spin_lock_irqsave(&ramht->lock, flags);
+	list_for_each_entry(entry, &chan->ramht->entries, head) {
+		if (entry->channel == chan && entry->handle == handle) {
+			gpuobj = entry->gpuobj;
+			break;
+		}
+	}
+	spin_unlock_irqrestore(&ramht->lock, flags);
+
+	return gpuobj;
+}
+
+int
+nouveau_ramht_new(struct drm_device *dev, struct nouveau_gpuobj *gpuobj,
+		  struct nouveau_ramht **pramht)
+{
+	struct nouveau_ramht *ramht;
+
+	ramht = kzalloc(sizeof(*ramht), GFP_KERNEL);
+	if (!ramht)
+		return -ENOMEM;
+
+	ramht->dev = dev;
+	kref_init(&ramht->refcount);
+	ramht->bits = drm_order(gpuobj->size / 8);
+	INIT_LIST_HEAD(&ramht->entries);
+	spin_lock_init(&ramht->lock);
+	nouveau_gpuobj_ref(gpuobj, &ramht->gpuobj);
+
+	*pramht = ramht;
+	return 0;
+}
+
+static void
+nouveau_ramht_del(struct kref *ref)
+{
+	struct nouveau_ramht *ramht =
+		container_of(ref, struct nouveau_ramht, refcount);
+
+	nouveau_gpuobj_ref(NULL, &ramht->gpuobj);
+	kfree(ramht);
+}
+
+void
+nouveau_ramht_ref(struct nouveau_ramht *ref, struct nouveau_ramht **ptr,
+		  struct nouveau_channel *chan)
+{
+	struct nouveau_ramht_entry *entry, *tmp;
+	struct nouveau_ramht *ramht;
+	unsigned long flags;
+
+	if (ref)
+		kref_get(&ref->refcount);
+
+	ramht = *ptr;
+	if (ramht) {
+		spin_lock_irqsave(&ramht->lock, flags);
+		list_for_each_entry_safe(entry, tmp, &ramht->entries, head) {
+			if (entry->channel != chan)
+				continue;
+
+			nouveau_ramht_remove_locked(chan, entry->handle);
+		}
+		spin_unlock_irqrestore(&ramht->lock, flags);
+
+		kref_put(&ramht->refcount, nouveau_ramht_del);
+	}
+	*ptr = ref;
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_ramht.h b/drivers/gpu/drm/nouveau/nouveau_ramht.h
new file mode 100644
index 000000000000..b79cb5e1a8f1
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_ramht.h
@@ -0,0 +1,55 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#ifndef __NOUVEAU_RAMHT_H__
+#define __NOUVEAU_RAMHT_H__
+
+struct nouveau_ramht_entry {
+	struct list_head head;
+	struct nouveau_channel *channel;
+	struct nouveau_gpuobj *gpuobj;
+	u32 handle;
+};
+
+struct nouveau_ramht {
+	struct drm_device *dev;
+	struct kref refcount;
+	spinlock_t lock;
+	struct nouveau_gpuobj *gpuobj;
+	struct list_head entries;
+	int bits;
+};
+
+extern int  nouveau_ramht_new(struct drm_device *, struct nouveau_gpuobj *,
+			      struct nouveau_ramht **);
+extern void nouveau_ramht_ref(struct nouveau_ramht *, struct nouveau_ramht **,
+			      struct nouveau_channel *unref_channel);
+
+extern int  nouveau_ramht_insert(struct nouveau_channel *, u32 handle,
+				 struct nouveau_gpuobj *);
+extern void nouveau_ramht_remove(struct nouveau_channel *, u32 handle);
+extern struct nouveau_gpuobj *
+nouveau_ramht_find(struct nouveau_channel *chan, u32 handle);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_reg.h b/drivers/gpu/drm/nouveau/nouveau_reg.h
index 21a6e453b975..1b42541ca9e5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_reg.h
+++ b/drivers/gpu/drm/nouveau/nouveau_reg.h
@@ -551,6 +551,8 @@
 #define NV10_PFIFO_CACHE1_DMA_SUBROUTINE                   0x0000324C
 #define NV03_PFIFO_CACHE1_PULL0                            0x00003240
 #define NV04_PFIFO_CACHE1_PULL0                            0x00003250
+#    define NV04_PFIFO_CACHE1_PULL0_HASH_FAILED            0x00000010
+#    define NV04_PFIFO_CACHE1_PULL0_HASH_BUSY              0x00001000
 #define NV03_PFIFO_CACHE1_PULL1                            0x00003250
 #define NV04_PFIFO_CACHE1_PULL1                            0x00003254
 #define NV04_PFIFO_CACHE1_HASH                             0x00003258
@@ -785,15 +787,12 @@
 #define NV50_PDISPLAY_DAC_MODE_CTRL_C(i)                (0x00610b5c + (i) * 0x8)
 #define NV50_PDISPLAY_SOR_MODE_CTRL_P(i)                (0x00610b70 + (i) * 0x8)
 #define NV50_PDISPLAY_SOR_MODE_CTRL_C(i)                (0x00610b74 + (i) * 0x8)
+#define NV50_PDISPLAY_EXT_MODE_CTRL_P(i)                (0x00610b80 + (i) * 0x8)
+#define NV50_PDISPLAY_EXT_MODE_CTRL_C(i)                (0x00610b84 + (i) * 0x8)
 #define NV50_PDISPLAY_DAC_MODE_CTRL2_P(i)               (0x00610bdc + (i) * 0x8)
 #define NV50_PDISPLAY_DAC_MODE_CTRL2_C(i)               (0x00610be0 + (i) * 0x8)
-
 #define NV90_PDISPLAY_SOR_MODE_CTRL_P(i)                (0x00610794 + (i) * 0x8)
 #define NV90_PDISPLAY_SOR_MODE_CTRL_C(i)                (0x00610798 + (i) * 0x8)
-#define NV90_PDISPLAY_DAC_MODE_CTRL_P(i)                (0x00610b58 + (i) * 0x8)
-#define NV90_PDISPLAY_DAC_MODE_CTRL_C(i)                (0x00610b5c + (i) * 0x8)
-#define NV90_PDISPLAY_DAC_MODE_CTRL2_P(i)               (0x00610b80 + (i) * 0x8)
-#define NV90_PDISPLAY_DAC_MODE_CTRL2_C(i)               (0x00610b84 + (i) * 0x8)
 
 #define NV50_PDISPLAY_CRTC_CLK                                       0x00614000
 #define NV50_PDISPLAY_CRTC_CLK_CTRL1(i)                 ((i) * 0x800 + 0x614100)
diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
index 6b9187d7f67d..288bacac7e5a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
@@ -95,9 +95,9 @@ nouveau_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem)
 	struct nouveau_gpuobj *gpuobj = dev_priv->gart_info.sg_ctxdma;
 	unsigned i, j, pte;
 
-	NV_DEBUG(dev, "pg=0x%lx\n", mem->mm_node->start);
+	NV_DEBUG(dev, "pg=0x%lx\n", mem->start);
 
-	pte = nouveau_sgdma_pte(nvbe->dev, mem->mm_node->start << PAGE_SHIFT);
+	pte = nouveau_sgdma_pte(nvbe->dev, mem->start << PAGE_SHIFT);
 	nvbe->pte_start = pte;
 	for (i = 0; i < nvbe->nr_pages; i++) {
 		dma_addr_t dma_offset = nvbe->pages[i];
@@ -105,11 +105,13 @@ nouveau_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem)
 		uint32_t offset_h = upper_32_bits(dma_offset);
 
 		for (j = 0; j < PAGE_SIZE / NV_CTXDMA_PAGE_SIZE; j++) {
-			if (dev_priv->card_type < NV_50)
-				nv_wo32(dev, gpuobj, pte++, offset_l | 3);
-			else {
-				nv_wo32(dev, gpuobj, pte++, offset_l | 0x21);
-				nv_wo32(dev, gpuobj, pte++, offset_h & 0xff);
+			if (dev_priv->card_type < NV_50) {
+				nv_wo32(gpuobj, (pte * 4) + 0, offset_l | 3);
+				pte += 1;
+			} else {
+				nv_wo32(gpuobj, (pte * 4) + 0, offset_l | 0x21);
+				nv_wo32(gpuobj, (pte * 4) + 4, offset_h & 0xff);
+				pte += 2;
 			}
 
 			dma_offset += NV_CTXDMA_PAGE_SIZE;
@@ -145,11 +147,13 @@ nouveau_sgdma_unbind(struct ttm_backend *be)
 		dma_addr_t dma_offset = dev_priv->gart_info.sg_dummy_bus;
 
 		for (j = 0; j < PAGE_SIZE / NV_CTXDMA_PAGE_SIZE; j++) {
-			if (dev_priv->card_type < NV_50)
-				nv_wo32(dev, gpuobj, pte++, dma_offset | 3);
-			else {
-				nv_wo32(dev, gpuobj, pte++, dma_offset | 0x21);
-				nv_wo32(dev, gpuobj, pte++, 0x00000000);
+			if (dev_priv->card_type < NV_50) {
+				nv_wo32(gpuobj, (pte * 4) + 0, dma_offset | 3);
+				pte += 1;
+			} else {
+				nv_wo32(gpuobj, (pte * 4) + 0, 0x00000000);
+				nv_wo32(gpuobj, (pte * 4) + 4, 0x00000000);
+				pte += 2;
 			}
 
 			dma_offset += NV_CTXDMA_PAGE_SIZE;
@@ -230,7 +234,6 @@ nouveau_sgdma_init(struct drm_device *dev)
 	}
 
 	ret = nouveau_gpuobj_new(dev, NULL, obj_size, 16,
-				      NVOBJ_FLAG_ALLOW_NO_REFS |
 				      NVOBJ_FLAG_ZERO_ALLOC |
 				      NVOBJ_FLAG_ZERO_FREE, &gpuobj);
 	if (ret) {
@@ -239,9 +242,9 @@ nouveau_sgdma_init(struct drm_device *dev)
 	}
 
 	dev_priv->gart_info.sg_dummy_page =
-		alloc_page(GFP_KERNEL|__GFP_DMA32);
+		alloc_page(GFP_KERNEL|__GFP_DMA32|__GFP_ZERO);
 	if (!dev_priv->gart_info.sg_dummy_page) {
-		nouveau_gpuobj_del(dev, &gpuobj);
+		nouveau_gpuobj_ref(NULL, &gpuobj);
 		return -ENOMEM;
 	}
 
@@ -250,29 +253,34 @@ nouveau_sgdma_init(struct drm_device *dev)
 		pci_map_page(pdev, dev_priv->gart_info.sg_dummy_page, 0,
 			     PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
 	if (pci_dma_mapping_error(pdev, dev_priv->gart_info.sg_dummy_bus)) {
-		nouveau_gpuobj_del(dev, &gpuobj);
+		nouveau_gpuobj_ref(NULL, &gpuobj);
 		return -EFAULT;
 	}
 
 	if (dev_priv->card_type < NV_50) {
+		/* special case, allocated from global instmem heap so
+		 * cinst is invalid, we use it on all channels though so
+		 * cinst needs to be valid, set it the same as pinst
+		 */
+		gpuobj->cinst = gpuobj->pinst;
+
 		/* Maybe use NV_DMA_TARGET_AGP for PCIE? NVIDIA do this, and
 		 * confirmed to work on c51.  Perhaps means NV_DMA_TARGET_PCIE
 		 * on those cards? */
-		nv_wo32(dev, gpuobj, 0, NV_CLASS_DMA_IN_MEMORY |
-				       (1 << 12) /* PT present */ |
-				       (0 << 13) /* PT *not* linear */ |
-				       (NV_DMA_ACCESS_RW  << 14) |
-				       (NV_DMA_TARGET_PCI << 16));
-		nv_wo32(dev, gpuobj, 1, aper_size - 1);
+		nv_wo32(gpuobj, 0, NV_CLASS_DMA_IN_MEMORY |
+				   (1 << 12) /* PT present */ |
+				   (0 << 13) /* PT *not* linear */ |
+				   (NV_DMA_ACCESS_RW  << 14) |
+				   (NV_DMA_TARGET_PCI << 16));
+		nv_wo32(gpuobj, 4, aper_size - 1);
 		for (i = 2; i < 2 + (aper_size >> 12); i++) {
-			nv_wo32(dev, gpuobj, i,
-				    dev_priv->gart_info.sg_dummy_bus | 3);
+			nv_wo32(gpuobj, i * 4,
+				dev_priv->gart_info.sg_dummy_bus | 3);
 		}
 	} else {
 		for (i = 0; i < obj_size; i += 8) {
-			nv_wo32(dev, gpuobj, (i+0)/4,
-				    dev_priv->gart_info.sg_dummy_bus | 0x21);
-			nv_wo32(dev, gpuobj, (i+4)/4, 0);
+			nv_wo32(gpuobj, i + 0, 0x00000000);
+			nv_wo32(gpuobj, i + 4, 0x00000000);
 		}
 	}
 	dev_priv->engine.instmem.flush(dev);
@@ -298,7 +306,7 @@ nouveau_sgdma_takedown(struct drm_device *dev)
 		dev_priv->gart_info.sg_dummy_bus = 0;
 	}
 
-	nouveau_gpuobj_del(dev, &dev_priv->gart_info.sg_ctxdma);
+	nouveau_gpuobj_ref(NULL, &dev_priv->gart_info.sg_ctxdma);
 }
 
 int
@@ -308,9 +316,9 @@ nouveau_sgdma_get_page(struct drm_device *dev, uint32_t offset, uint32_t *page)
 	struct nouveau_gpuobj *gpuobj = dev_priv->gart_info.sg_ctxdma;
 	int pte;
 
-	pte = (offset >> NV_CTXDMA_PAGE_SHIFT);
+	pte = (offset >> NV_CTXDMA_PAGE_SHIFT) << 2;
 	if (dev_priv->card_type < NV_50) {
-		*page = nv_ro32(dev, gpuobj, (pte + 2)) & ~NV_CTXDMA_PAGE_MASK;
+		*page = nv_ro32(gpuobj, (pte + 8)) & ~NV_CTXDMA_PAGE_MASK;
 		return 0;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_state.c b/drivers/gpu/drm/nouveau/nouveau_state.c
index 989322be3728..ed7757f14083 100644
--- a/drivers/gpu/drm/nouveau/nouveau_state.c
+++ b/drivers/gpu/drm/nouveau/nouveau_state.c
@@ -35,6 +35,8 @@
 #include "nouveau_drv.h"
 #include "nouveau_drm.h"
 #include "nouveau_fbcon.h"
+#include "nouveau_ramht.h"
+#include "nouveau_pm.h"
 #include "nv50_display.h"
 
 static void nouveau_stub_takedown(struct drm_device *dev) {}
@@ -78,7 +80,6 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->fifo.disable		= nv04_fifo_disable;
 		engine->fifo.enable		= nv04_fifo_enable;
 		engine->fifo.reassign		= nv04_fifo_reassign;
-		engine->fifo.cache_flush	= nv04_fifo_cache_flush;
 		engine->fifo.cache_pull		= nv04_fifo_cache_pull;
 		engine->fifo.channel_id		= nv04_fifo_channel_id;
 		engine->fifo.create_context	= nv04_fifo_create_context;
@@ -95,6 +96,9 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->gpio.get		= NULL;
 		engine->gpio.set		= NULL;
 		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
 		break;
 	case 0x10:
 		engine->instmem.init		= nv04_instmem_init;
@@ -130,7 +134,6 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->fifo.disable		= nv04_fifo_disable;
 		engine->fifo.enable		= nv04_fifo_enable;
 		engine->fifo.reassign		= nv04_fifo_reassign;
-		engine->fifo.cache_flush	= nv04_fifo_cache_flush;
 		engine->fifo.cache_pull		= nv04_fifo_cache_pull;
 		engine->fifo.channel_id		= nv10_fifo_channel_id;
 		engine->fifo.create_context	= nv10_fifo_create_context;
@@ -147,6 +150,9 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->gpio.get		= nv10_gpio_get;
 		engine->gpio.set		= nv10_gpio_set;
 		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
 		break;
 	case 0x20:
 		engine->instmem.init		= nv04_instmem_init;
@@ -182,7 +188,6 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->fifo.disable		= nv04_fifo_disable;
 		engine->fifo.enable		= nv04_fifo_enable;
 		engine->fifo.reassign		= nv04_fifo_reassign;
-		engine->fifo.cache_flush	= nv04_fifo_cache_flush;
 		engine->fifo.cache_pull		= nv04_fifo_cache_pull;
 		engine->fifo.channel_id		= nv10_fifo_channel_id;
 		engine->fifo.create_context	= nv10_fifo_create_context;
@@ -199,6 +204,9 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->gpio.get		= nv10_gpio_get;
 		engine->gpio.set		= nv10_gpio_set;
 		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
 		break;
 	case 0x30:
 		engine->instmem.init		= nv04_instmem_init;
@@ -234,7 +242,6 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->fifo.disable		= nv04_fifo_disable;
 		engine->fifo.enable		= nv04_fifo_enable;
 		engine->fifo.reassign		= nv04_fifo_reassign;
-		engine->fifo.cache_flush	= nv04_fifo_cache_flush;
 		engine->fifo.cache_pull		= nv04_fifo_cache_pull;
 		engine->fifo.channel_id		= nv10_fifo_channel_id;
 		engine->fifo.create_context	= nv10_fifo_create_context;
@@ -251,6 +258,11 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->gpio.get		= nv10_gpio_get;
 		engine->gpio.set		= nv10_gpio_set;
 		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
+		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
+		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
 		break;
 	case 0x40:
 	case 0x60:
@@ -287,7 +299,6 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->fifo.disable		= nv04_fifo_disable;
 		engine->fifo.enable		= nv04_fifo_enable;
 		engine->fifo.reassign		= nv04_fifo_reassign;
-		engine->fifo.cache_flush	= nv04_fifo_cache_flush;
 		engine->fifo.cache_pull		= nv04_fifo_cache_pull;
 		engine->fifo.channel_id		= nv10_fifo_channel_id;
 		engine->fifo.create_context	= nv40_fifo_create_context;
@@ -304,6 +315,12 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->gpio.get		= nv10_gpio_get;
 		engine->gpio.set		= nv10_gpio_set;
 		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
+		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
+		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
+		engine->pm.temp_get		= nv40_temp_get;
 		break;
 	case 0x50:
 	case 0x80: /* gotta love NVIDIA's consistency.. */
@@ -358,6 +375,27 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->gpio.get		= nv50_gpio_get;
 		engine->gpio.set		= nv50_gpio_set;
 		engine->gpio.irq_enable		= nv50_gpio_irq_enable;
+		switch (dev_priv->chipset) {
+		case 0xa3:
+		case 0xa5:
+		case 0xa8:
+		case 0xaf:
+			engine->pm.clock_get	= nva3_pm_clock_get;
+			engine->pm.clock_pre	= nva3_pm_clock_pre;
+			engine->pm.clock_set	= nva3_pm_clock_set;
+			break;
+		default:
+			engine->pm.clock_get	= nv50_pm_clock_get;
+			engine->pm.clock_pre	= nv50_pm_clock_pre;
+			engine->pm.clock_set	= nv50_pm_clock_set;
+			break;
+		}
+		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
+		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
+		if (dev_priv->chipset >= 0x84)
+			engine->pm.temp_get	= nv84_temp_get;
+		else
+			engine->pm.temp_get	= nv40_temp_get;
 		break;
 	case 0xC0:
 		engine->instmem.init		= nvc0_instmem_init;
@@ -437,16 +475,14 @@ static int
 nouveau_card_init_channel(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *gpuobj;
+	struct nouveau_gpuobj *gpuobj = NULL;
 	int ret;
 
 	ret = nouveau_channel_alloc(dev, &dev_priv->channel,
-				    (struct drm_file *)-2,
-				    NvDmaFB, NvDmaTT);
+				    (struct drm_file *)-2, NvDmaFB, NvDmaTT);
 	if (ret)
 		return ret;
 
-	gpuobj = NULL;
 	ret = nouveau_gpuobj_dma_new(dev_priv->channel, NV_CLASS_DMA_IN_MEMORY,
 				     0, dev_priv->vram_size,
 				     NV_DMA_ACCESS_RW, NV_DMA_TARGET_VIDMEM,
@@ -454,26 +490,25 @@ nouveau_card_init_channel(struct drm_device *dev)
 	if (ret)
 		goto out_err;
 
-	ret = nouveau_gpuobj_ref_add(dev, dev_priv->channel, NvDmaVRAM,
-				     gpuobj, NULL);
+	ret = nouveau_ramht_insert(dev_priv->channel, NvDmaVRAM, gpuobj);
+	nouveau_gpuobj_ref(NULL, &gpuobj);
 	if (ret)
 		goto out_err;
 
-	gpuobj = NULL;
 	ret = nouveau_gpuobj_gart_dma_new(dev_priv->channel, 0,
 					  dev_priv->gart_info.aper_size,
 					  NV_DMA_ACCESS_RW, &gpuobj, NULL);
 	if (ret)
 		goto out_err;
 
-	ret = nouveau_gpuobj_ref_add(dev, dev_priv->channel, NvDmaGART,
-				     gpuobj, NULL);
+	ret = nouveau_ramht_insert(dev_priv->channel, NvDmaGART, gpuobj);
+	nouveau_gpuobj_ref(NULL, &gpuobj);
 	if (ret)
 		goto out_err;
 
 	return 0;
+
 out_err:
-	nouveau_gpuobj_del(dev, &gpuobj);
 	nouveau_channel_free(dev_priv->channel);
 	dev_priv->channel = NULL;
 	return ret;
@@ -534,35 +569,28 @@ nouveau_card_init(struct drm_device *dev)
 	if (ret)
 		goto out_display_early;
 
-	ret = nouveau_mem_detect(dev);
+	nouveau_pm_init(dev);
+
+	ret = nouveau_mem_vram_init(dev);
 	if (ret)
 		goto out_bios;
 
-	ret = nouveau_gpuobj_early_init(dev);
-	if (ret)
-		goto out_bios;
-
-	/* Initialise instance memory, must happen before mem_init so we
-	 * know exactly how much VRAM we're able to use for "normal"
-	 * purposes.
-	 */
-	ret = engine->instmem.init(dev);
-	if (ret)
-		goto out_gpuobj_early;
-
-	/* Setup the memory manager */
-	ret = nouveau_mem_init(dev);
-	if (ret)
-		goto out_instmem;
-
 	ret = nouveau_gpuobj_init(dev);
 	if (ret)
-		goto out_mem;
+		goto out_vram;
+
+	ret = engine->instmem.init(dev);
+	if (ret)
+		goto out_gpuobj;
+
+	ret = nouveau_mem_gart_init(dev);
+	if (ret)
+		goto out_instmem;
 
 	/* PMC */
 	ret = engine->mc.init(dev);
 	if (ret)
-		goto out_gpuobj;
+		goto out_gart;
 
 	/* PGPIO */
 	ret = engine->gpio.init(dev);
@@ -611,9 +639,13 @@ nouveau_card_init(struct drm_device *dev)
 	/* what about PVIDEO/PCRTC/PRAMDAC etc? */
 
 	if (!engine->graph.accel_blocked) {
-		ret = nouveau_card_init_channel(dev);
+		ret = nouveau_fence_init(dev);
 		if (ret)
 			goto out_irq;
+
+		ret = nouveau_card_init_channel(dev);
+		if (ret)
+			goto out_fence;
 	}
 
 	ret = nouveau_backlight_init(dev);
@@ -624,6 +656,8 @@ nouveau_card_init(struct drm_device *dev)
 	drm_kms_helper_poll_init(dev);
 	return 0;
 
+out_fence:
+	nouveau_fence_fini(dev);
 out_irq:
 	drm_irq_uninstall(dev);
 out_display:
@@ -642,16 +676,16 @@ out_gpio:
 	engine->gpio.takedown(dev);
 out_mc:
 	engine->mc.takedown(dev);
-out_gpuobj:
-	nouveau_gpuobj_takedown(dev);
-out_mem:
-	nouveau_sgdma_takedown(dev);
-	nouveau_mem_close(dev);
+out_gart:
+	nouveau_mem_gart_fini(dev);
 out_instmem:
 	engine->instmem.takedown(dev);
-out_gpuobj_early:
-	nouveau_gpuobj_late_takedown(dev);
+out_gpuobj:
+	nouveau_gpuobj_takedown(dev);
+out_vram:
+	nouveau_mem_vram_fini(dev);
 out_bios:
+	nouveau_pm_fini(dev);
 	nouveau_bios_takedown(dev);
 out_display_early:
 	engine->display.late_takedown(dev);
@@ -667,7 +701,8 @@ static void nouveau_card_takedown(struct drm_device *dev)
 
 	nouveau_backlight_exit(dev);
 
-	if (dev_priv->channel) {
+	if (!engine->graph.accel_blocked) {
+		nouveau_fence_fini(dev);
 		nouveau_channel_free(dev_priv->channel);
 		dev_priv->channel = NULL;
 	}
@@ -686,15 +721,15 @@ static void nouveau_card_takedown(struct drm_device *dev)
 	ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_VRAM);
 	ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_TT);
 	mutex_unlock(&dev->struct_mutex);
-	nouveau_sgdma_takedown(dev);
+	nouveau_mem_gart_fini(dev);
 
-	nouveau_gpuobj_takedown(dev);
-	nouveau_mem_close(dev);
 	engine->instmem.takedown(dev);
+	nouveau_gpuobj_takedown(dev);
+	nouveau_mem_vram_fini(dev);
 
 	drm_irq_uninstall(dev);
 
-	nouveau_gpuobj_late_takedown(dev);
+	nouveau_pm_fini(dev);
 	nouveau_bios_takedown(dev);
 
 	vga_client_register(dev->pdev, NULL, NULL, NULL);
@@ -1057,7 +1092,7 @@ bool nouveau_wait_until(struct drm_device *dev, uint64_t timeout,
 /* Waits for PGRAPH to go completely idle */
 bool nouveau_wait_for_idle(struct drm_device *dev)
 {
-	if (!nv_wait(NV04_PGRAPH_STATUS, 0xffffffff, 0x00000000)) {
+	if (!nv_wait(dev, NV04_PGRAPH_STATUS, 0xffffffff, 0x00000000)) {
 		NV_ERROR(dev, "PGRAPH idle timed out with status 0x%08x\n",
 			 nv_rd32(dev, NV04_PGRAPH_STATUS));
 		return false;
diff --git a/drivers/gpu/drm/nouveau/nouveau_temp.c b/drivers/gpu/drm/nouveau/nouveau_temp.c
new file mode 100644
index 000000000000..16bbbf1eff63
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_temp.c
@@ -0,0 +1,309 @@
+/*
+ * Copyright 2010 PathScale inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Martin Peres
+ */
+
+#include "drmP.h"
+
+#include "nouveau_drv.h"
+#include "nouveau_pm.h"
+
+static void
+nouveau_temp_vbios_parse(struct drm_device *dev, u8 *temp)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_temp_sensor_constants *sensor = &pm->sensor_constants;
+	struct nouveau_pm_threshold_temp *temps = &pm->threshold_temp;
+	int i, headerlen, recordlen, entries;
+
+	if (!temp) {
+		NV_DEBUG(dev, "temperature table pointer invalid\n");
+		return;
+	}
+
+	/* Set the default sensor's contants */
+	sensor->offset_constant = 0;
+	sensor->offset_mult = 1;
+	sensor->offset_div = 1;
+	sensor->slope_mult = 1;
+	sensor->slope_div = 1;
+
+	/* Set the default temperature thresholds */
+	temps->critical = 110;
+	temps->down_clock = 100;
+	temps->fan_boost = 90;
+
+	/* Set the known default values to setup the temperature sensor */
+	if (dev_priv->card_type >= NV_40) {
+		switch (dev_priv->chipset) {
+		case 0x43:
+			sensor->offset_mult = 32060;
+			sensor->offset_div = 1000;
+			sensor->slope_mult = 792;
+			sensor->slope_div = 1000;
+			break;
+
+		case 0x44:
+		case 0x47:
+		case 0x4a:
+			sensor->offset_mult = 27839;
+			sensor->offset_div = 1000;
+			sensor->slope_mult = 780;
+			sensor->slope_div = 1000;
+			break;
+
+		case 0x46:
+			sensor->offset_mult = -24775;
+			sensor->offset_div = 100;
+			sensor->slope_mult = 467;
+			sensor->slope_div = 10000;
+			break;
+
+		case 0x49:
+			sensor->offset_mult = -25051;
+			sensor->offset_div = 100;
+			sensor->slope_mult = 458;
+			sensor->slope_div = 10000;
+			break;
+
+		case 0x4b:
+			sensor->offset_mult = -24088;
+			sensor->offset_div = 100;
+			sensor->slope_mult = 442;
+			sensor->slope_div = 10000;
+			break;
+
+		case 0x50:
+			sensor->offset_mult = -22749;
+			sensor->offset_div = 100;
+			sensor->slope_mult = 431;
+			sensor->slope_div = 10000;
+			break;
+		}
+	}
+
+	headerlen = temp[1];
+	recordlen = temp[2];
+	entries = temp[3];
+	temp = temp + headerlen;
+
+	/* Read the entries from the table */
+	for (i = 0; i < entries; i++) {
+		u16 value = ROM16(temp[1]);
+
+		switch (temp[0]) {
+		case 0x01:
+			if ((value & 0x8f) == 0)
+				sensor->offset_constant = (value >> 9) & 0x7f;
+			break;
+
+		case 0x04:
+			if ((value & 0xf00f) == 0xa000) /* core */
+				temps->critical = (value&0x0ff0) >> 4;
+			break;
+
+		case 0x07:
+			if ((value & 0xf00f) == 0xa000) /* core */
+				temps->down_clock = (value&0x0ff0) >> 4;
+			break;
+
+		case 0x08:
+			if ((value & 0xf00f) == 0xa000) /* core */
+				temps->fan_boost = (value&0x0ff0) >> 4;
+			break;
+
+		case 0x10:
+			sensor->offset_mult = value;
+			break;
+
+		case 0x11:
+			sensor->offset_div = value;
+			break;
+
+		case 0x12:
+			sensor->slope_mult = value;
+			break;
+
+		case 0x13:
+			sensor->slope_div = value;
+			break;
+		}
+		temp += recordlen;
+	}
+
+	nouveau_temp_safety_checks(dev);
+}
+
+static int
+nv40_sensor_setup(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_temp_sensor_constants *sensor = &pm->sensor_constants;
+	u32 offset = sensor->offset_mult / sensor->offset_div;
+	u32 sensor_calibration;
+
+	/* set up the sensors */
+	sensor_calibration = 120 - offset - sensor->offset_constant;
+	sensor_calibration = sensor_calibration * sensor->slope_div /
+				sensor->slope_mult;
+
+	if (dev_priv->chipset >= 0x46)
+		sensor_calibration |= 0x80000000;
+	else
+		sensor_calibration |= 0x10000000;
+
+	nv_wr32(dev, 0x0015b0, sensor_calibration);
+
+	/* Wait for the sensor to update */
+	msleep(5);
+
+	/* read */
+	return nv_rd32(dev, 0x0015b4) & 0x1fff;
+}
+
+int
+nv40_temp_get(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_temp_sensor_constants *sensor = &pm->sensor_constants;
+	int offset = sensor->offset_mult / sensor->offset_div;
+	int core_temp;
+
+	if (dev_priv->chipset >= 0x50) {
+		core_temp = nv_rd32(dev, 0x20008);
+	} else {
+		core_temp = nv_rd32(dev, 0x0015b4) & 0x1fff;
+		/* Setup the sensor if the temperature is 0 */
+		if (core_temp == 0)
+			core_temp = nv40_sensor_setup(dev);
+	}
+
+	core_temp = core_temp * sensor->slope_mult / sensor->slope_div;
+	core_temp = core_temp + offset + sensor->offset_constant;
+
+	return core_temp;
+}
+
+int
+nv84_temp_get(struct drm_device *dev)
+{
+	return nv_rd32(dev, 0x20400);
+}
+
+void
+nouveau_temp_safety_checks(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_threshold_temp *temps = &pm->threshold_temp;
+
+	if (temps->critical > 120)
+		temps->critical = 120;
+	else if (temps->critical < 80)
+		temps->critical = 80;
+
+	if (temps->down_clock > 110)
+		temps->down_clock = 110;
+	else if (temps->down_clock < 60)
+		temps->down_clock = 60;
+
+	if (temps->fan_boost > 100)
+		temps->fan_boost = 100;
+	else if (temps->fan_boost < 40)
+		temps->fan_boost = 40;
+}
+
+static bool
+probe_monitoring_device(struct nouveau_i2c_chan *i2c,
+			struct i2c_board_info *info)
+{
+	char modalias[16] = "i2c:";
+	struct i2c_client *client;
+
+	strlcat(modalias, info->type, sizeof(modalias));
+	request_module(modalias);
+
+	client = i2c_new_device(&i2c->adapter, info);
+	if (!client)
+		return false;
+
+	if (!client->driver || client->driver->detect(client, info)) {
+		i2c_unregister_device(client);
+		return false;
+	}
+
+	return true;
+}
+
+static void
+nouveau_temp_probe_i2c(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct dcb_table *dcb = &dev_priv->vbios.dcb;
+	struct i2c_board_info info[] = {
+		{ I2C_BOARD_INFO("w83l785ts", 0x2d) },
+		{ I2C_BOARD_INFO("w83781d", 0x2d) },
+		{ I2C_BOARD_INFO("f75375", 0x2e) },
+		{ I2C_BOARD_INFO("adt7473", 0x2e) },
+		{ I2C_BOARD_INFO("lm99", 0x4c) },
+		{ }
+	};
+	int idx = (dcb->version >= 0x40 ?
+		   dcb->i2c_default_indices & 0xf : 2);
+
+	nouveau_i2c_identify(dev, "monitoring device", info,
+			     probe_monitoring_device, idx);
+}
+
+void
+nouveau_temp_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct bit_entry P;
+	u8 *temp = NULL;
+
+	if (bios->type == NVBIOS_BIT) {
+		if (bit_table(dev, 'P', &P))
+			return;
+
+		if (P.version == 1)
+			temp = ROMPTR(bios, P.data[12]);
+		else if (P.version == 2)
+			temp = ROMPTR(bios, P.data[16]);
+		else
+			NV_WARN(dev, "unknown temp for BIT P %d\n", P.version);
+
+		nouveau_temp_vbios_parse(dev, temp);
+	}
+
+	nouveau_temp_probe_i2c(dev);
+}
+
+void
+nouveau_temp_fini(struct drm_device *dev)
+{
+
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_volt.c b/drivers/gpu/drm/nouveau/nouveau_volt.c
new file mode 100644
index 000000000000..04fdc00a67d5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nouveau_volt.c
@@ -0,0 +1,212 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "drmP.h"
+
+#include "nouveau_drv.h"
+#include "nouveau_pm.h"
+
+static const enum dcb_gpio_tag vidtag[] = { 0x04, 0x05, 0x06, 0x1a };
+static int nr_vidtag = sizeof(vidtag) / sizeof(vidtag[0]);
+
+int
+nouveau_voltage_gpio_get(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio;
+	struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage;
+	u8 vid = 0;
+	int i;
+
+	for (i = 0; i < nr_vidtag; i++) {
+		if (!(volt->vid_mask & (1 << i)))
+			continue;
+
+		vid |= gpio->get(dev, vidtag[i]) << i;
+	}
+
+	return nouveau_volt_lvl_lookup(dev, vid);
+}
+
+int
+nouveau_voltage_gpio_set(struct drm_device *dev, int voltage)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio;
+	struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage;
+	int vid, i;
+
+	vid = nouveau_volt_vid_lookup(dev, voltage);
+	if (vid < 0)
+		return vid;
+
+	for (i = 0; i < nr_vidtag; i++) {
+		if (!(volt->vid_mask & (1 << i)))
+			continue;
+
+		gpio->set(dev, vidtag[i], !!(vid & (1 << i)));
+	}
+
+	return 0;
+}
+
+int
+nouveau_volt_vid_lookup(struct drm_device *dev, int voltage)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage;
+	int i;
+
+	for (i = 0; i < volt->nr_level; i++) {
+		if (volt->level[i].voltage == voltage)
+			return volt->level[i].vid;
+	}
+
+	return -ENOENT;
+}
+
+int
+nouveau_volt_lvl_lookup(struct drm_device *dev, int vid)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage;
+	int i;
+
+	for (i = 0; i < volt->nr_level; i++) {
+		if (volt->level[i].vid == vid)
+			return volt->level[i].voltage;
+	}
+
+	return -ENOENT;
+}
+
+void
+nouveau_volt_init(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_voltage *voltage = &pm->voltage;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct bit_entry P;
+	u8 *volt = NULL, *entry;
+	int i, headerlen, recordlen, entries, vidmask, vidshift;
+
+	if (bios->type == NVBIOS_BIT) {
+		if (bit_table(dev, 'P', &P))
+			return;
+
+		if (P.version == 1)
+			volt = ROMPTR(bios, P.data[16]);
+		else
+		if (P.version == 2)
+			volt = ROMPTR(bios, P.data[12]);
+		else {
+			NV_WARN(dev, "unknown volt for BIT P %d\n", P.version);
+		}
+	} else {
+		if (bios->data[bios->offset + 6] < 0x27) {
+			NV_DEBUG(dev, "BMP version too old for voltage\n");
+			return;
+		}
+
+		volt = ROMPTR(bios, bios->data[bios->offset + 0x98]);
+	}
+
+	if (!volt) {
+		NV_DEBUG(dev, "voltage table pointer invalid\n");
+		return;
+	}
+
+	switch (volt[0]) {
+	case 0x10:
+	case 0x11:
+	case 0x12:
+		headerlen = 5;
+		recordlen = volt[1];
+		entries   = volt[2];
+		vidshift  = 0;
+		vidmask   = volt[4];
+		break;
+	case 0x20:
+		headerlen = volt[1];
+		recordlen = volt[3];
+		entries   = volt[2];
+		vidshift  = 0; /* could be vidshift like 0x30? */
+		vidmask   = volt[5];
+		break;
+	case 0x30:
+		headerlen = volt[1];
+		recordlen = volt[2];
+		entries   = volt[3];
+		vidshift  = hweight8(volt[5]);
+		vidmask   = volt[4];
+		break;
+	default:
+		NV_WARN(dev, "voltage table 0x%02x unknown\n", volt[0]);
+		return;
+	}
+
+	/* validate vid mask */
+	voltage->vid_mask = vidmask;
+	if (!voltage->vid_mask)
+		return;
+
+	i = 0;
+	while (vidmask) {
+		if (i > nr_vidtag) {
+			NV_DEBUG(dev, "vid bit %d unknown\n", i);
+			return;
+		}
+
+		if (!nouveau_bios_gpio_entry(dev, vidtag[i])) {
+			NV_DEBUG(dev, "vid bit %d has no gpio tag\n", i);
+			return;
+		}
+
+		vidmask >>= 1;
+		i++;
+	}
+
+	/* parse vbios entries into common format */
+	voltage->level = kcalloc(entries, sizeof(*voltage->level), GFP_KERNEL);
+	if (!voltage->level)
+		return;
+
+	entry = volt + headerlen;
+	for (i = 0; i < entries; i++, entry += recordlen) {
+		voltage->level[i].voltage = entry[0];
+		voltage->level[i].vid     = entry[1] >> vidshift;
+	}
+	voltage->nr_level  = entries;
+	voltage->supported = true;
+}
+
+void
+nouveau_volt_fini(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage;
+
+	kfree(volt->level);
+}
diff --git a/drivers/gpu/drm/nouveau/nv04_crtc.c b/drivers/gpu/drm/nouveau/nv04_crtc.c
index 497df8765f28..ef480281afec 100644
--- a/drivers/gpu/drm/nouveau/nv04_crtc.c
+++ b/drivers/gpu/drm/nouveau/nv04_crtc.c
@@ -109,7 +109,7 @@ static void nv_crtc_calc_state_ext(struct drm_crtc *crtc, struct drm_display_mod
 	struct nouveau_pll_vals *pv = &regp->pllvals;
 	struct pll_lims pll_lim;
 
-	if (get_pll_limits(dev, nv_crtc->index ? VPLL2 : VPLL1, &pll_lim))
+	if (get_pll_limits(dev, nv_crtc->index ? PLL_VPLL1 : PLL_VPLL0, &pll_lim))
 		return;
 
 	/* NM2 == 0 is used to determine single stage mode on two stage plls */
@@ -718,6 +718,7 @@ static void nv_crtc_destroy(struct drm_crtc *crtc)
 
 	drm_crtc_cleanup(crtc);
 
+	nouveau_bo_unmap(nv_crtc->cursor.nvbo);
 	nouveau_bo_ref(NULL, &nv_crtc->cursor.nvbo);
 	kfree(nv_crtc);
 }
@@ -826,7 +827,7 @@ nv04_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_FF_INDEX);
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_FFLWM__INDEX);
 
-	if (dev_priv->card_type >= NV_30) {
+	if (dev_priv->card_type >= NV_20) {
 		regp->CRTC[NV_CIO_CRE_47] = arb_lwm >> 8;
 		crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_47);
 	}
diff --git a/drivers/gpu/drm/nouveau/nv04_dac.c b/drivers/gpu/drm/nouveau/nv04_dac.c
index ea3627041ecf..ba6423f2ffcc 100644
--- a/drivers/gpu/drm/nouveau/nv04_dac.c
+++ b/drivers/gpu/drm/nouveau/nv04_dac.c
@@ -291,6 +291,8 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder)
 	msleep(5);
 
 	sample = NVReadRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + regoffset);
+	/* do it again just in case it's a residual current */
+	sample &= NVReadRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + regoffset);
 
 	temp = NVReadRAMDAC(dev, head, NV_PRAMDAC_TEST_CONTROL);
 	NVWriteRAMDAC(dev, head, NV_PRAMDAC_TEST_CONTROL,
@@ -343,22 +345,13 @@ static void nv04_dac_prepare(struct drm_encoder *encoder)
 {
 	struct drm_encoder_helper_funcs *helper = encoder->helper_private;
 	struct drm_device *dev = encoder->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	int head = nouveau_crtc(encoder->crtc)->index;
-	struct nv04_crtc_reg *crtcstate = dev_priv->mode_reg.crtc_reg;
 
 	helper->dpms(encoder, DRM_MODE_DPMS_OFF);
 
 	nv04_dfp_disable(dev, head);
-
-	/* Some NV4x have unknown values (0x3f, 0x50, 0x54, 0x6b, 0x79, 0x7f)
-	 * at LCD__INDEX which we don't alter
-	 */
-	if (!(crtcstate[head].CRTC[NV_CIO_CRE_LCD__INDEX] & 0x44))
-		crtcstate[head].CRTC[NV_CIO_CRE_LCD__INDEX] = 0;
 }
 
-
 static void nv04_dac_mode_set(struct drm_encoder *encoder,
 			      struct drm_display_mode *mode,
 			      struct drm_display_mode *adjusted_mode)
diff --git a/drivers/gpu/drm/nouveau/nv04_dfp.c b/drivers/gpu/drm/nouveau/nv04_dfp.c
index 0d3206a7046c..c936403b26e2 100644
--- a/drivers/gpu/drm/nouveau/nv04_dfp.c
+++ b/drivers/gpu/drm/nouveau/nv04_dfp.c
@@ -104,6 +104,8 @@ void nv04_dfp_disable(struct drm_device *dev, int head)
 	}
 	/* don't inadvertently turn it on when state written later */
 	crtcstate[head].fp_control = FP_TG_CONTROL_OFF;
+	crtcstate[head].CRTC[NV_CIO_CRE_LCD__INDEX] &=
+		~NV_CIO_CRE_LCD_ROUTE_MASK;
 }
 
 void nv04_dfp_update_fp_control(struct drm_encoder *encoder, int mode)
@@ -253,26 +255,21 @@ static void nv04_dfp_prepare(struct drm_encoder *encoder)
 
 	nv04_dfp_prepare_sel_clk(dev, nv_encoder, head);
 
-	/* Some NV4x have unknown values (0x3f, 0x50, 0x54, 0x6b, 0x79, 0x7f)
-	 * at LCD__INDEX which we don't alter
-	 */
-	if (!(*cr_lcd & 0x44)) {
-		*cr_lcd = 0x3;
+	*cr_lcd = (*cr_lcd & ~NV_CIO_CRE_LCD_ROUTE_MASK) | 0x3;
 
-		if (nv_two_heads(dev)) {
-			if (nv_encoder->dcb->location == DCB_LOC_ON_CHIP)
-				*cr_lcd |= head ? 0x0 : 0x8;
-			else {
-				*cr_lcd |= (nv_encoder->dcb->or << 4) & 0x30;
-				if (nv_encoder->dcb->type == OUTPUT_LVDS)
-					*cr_lcd |= 0x30;
-				if ((*cr_lcd & 0x30) == (*cr_lcd_oth & 0x30)) {
-					/* avoid being connected to both crtcs */
-					*cr_lcd_oth &= ~0x30;
-					NVWriteVgaCrtc(dev, head ^ 1,
-						       NV_CIO_CRE_LCD__INDEX,
-						       *cr_lcd_oth);
-				}
+	if (nv_two_heads(dev)) {
+		if (nv_encoder->dcb->location == DCB_LOC_ON_CHIP)
+			*cr_lcd |= head ? 0x0 : 0x8;
+		else {
+			*cr_lcd |= (nv_encoder->dcb->or << 4) & 0x30;
+			if (nv_encoder->dcb->type == OUTPUT_LVDS)
+				*cr_lcd |= 0x30;
+			if ((*cr_lcd & 0x30) == (*cr_lcd_oth & 0x30)) {
+				/* avoid being connected to both crtcs */
+				*cr_lcd_oth &= ~0x30;
+				NVWriteVgaCrtc(dev, head ^ 1,
+					       NV_CIO_CRE_LCD__INDEX,
+					       *cr_lcd_oth);
 			}
 		}
 	}
@@ -640,7 +637,7 @@ static void nv04_tmds_slave_init(struct drm_encoder *encoder)
 	    get_tmds_slave(encoder))
 		return;
 
-	type = nouveau_i2c_identify(dev, "TMDS transmitter", info, 2);
+	type = nouveau_i2c_identify(dev, "TMDS transmitter", info, NULL, 2);
 	if (type < 0)
 		return;
 
diff --git a/drivers/gpu/drm/nouveau/nv04_fbcon.c b/drivers/gpu/drm/nouveau/nv04_fbcon.c
index 1eeac4fae73d..33e4c9388bc1 100644
--- a/drivers/gpu/drm/nouveau/nv04_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nv04_fbcon.c
@@ -25,6 +25,7 @@
 #include "drmP.h"
 #include "nouveau_drv.h"
 #include "nouveau_dma.h"
+#include "nouveau_ramht.h"
 #include "nouveau_fbcon.h"
 
 void
@@ -169,11 +170,9 @@ nv04_fbcon_grobj_new(struct drm_device *dev, int class, uint32_t handle)
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_ref_add(dev, dev_priv->channel, handle, obj, NULL);
-	if (ret)
-		return ret;
-
-	return 0;
+	ret = nouveau_ramht_insert(dev_priv->channel, handle, obj);
+	nouveau_gpuobj_ref(NULL, &obj);
+	return ret;
 }
 
 int
diff --git a/drivers/gpu/drm/nouveau/nv04_fifo.c b/drivers/gpu/drm/nouveau/nv04_fifo.c
index 06cedd99c26a..708293b7ddcd 100644
--- a/drivers/gpu/drm/nouveau/nv04_fifo.c
+++ b/drivers/gpu/drm/nouveau/nv04_fifo.c
@@ -27,8 +27,9 @@
 #include "drmP.h"
 #include "drm.h"
 #include "nouveau_drv.h"
+#include "nouveau_ramht.h"
 
-#define NV04_RAMFC(c) (dev_priv->ramfc_offset + ((c) * NV04_RAMFC__SIZE))
+#define NV04_RAMFC(c) (dev_priv->ramfc->pinst + ((c) * NV04_RAMFC__SIZE))
 #define NV04_RAMFC__SIZE 32
 #define NV04_RAMFC_DMA_PUT                                       0x00
 #define NV04_RAMFC_DMA_GET                                       0x04
@@ -38,10 +39,8 @@
 #define NV04_RAMFC_ENGINE                                        0x14
 #define NV04_RAMFC_PULL1_ENGINE                                  0x18
 
-#define RAMFC_WR(offset, val) nv_wo32(dev, chan->ramfc->gpuobj, \
-					 NV04_RAMFC_##offset/4, (val))
-#define RAMFC_RD(offset)      nv_ro32(dev, chan->ramfc->gpuobj, \
-					 NV04_RAMFC_##offset/4)
+#define RAMFC_WR(offset, val) nv_wo32(chan->ramfc, NV04_RAMFC_##offset, (val))
+#define RAMFC_RD(offset)      nv_ro32(chan->ramfc, NV04_RAMFC_##offset)
 
 void
 nv04_fifo_disable(struct drm_device *dev)
@@ -71,38 +70,33 @@ nv04_fifo_reassign(struct drm_device *dev, bool enable)
 	return (reassign == 1);
 }
 
-bool
-nv04_fifo_cache_flush(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer;
-	uint64_t start = ptimer->read(dev);
-
-	do {
-		if (nv_rd32(dev, NV03_PFIFO_CACHE1_GET) ==
-		    nv_rd32(dev, NV03_PFIFO_CACHE1_PUT))
-			return true;
-
-	} while (ptimer->read(dev) - start < 100000000);
-
-	NV_ERROR(dev, "Timeout flushing the PFIFO cache.\n");
-
-	return false;
-}
-
 bool
 nv04_fifo_cache_pull(struct drm_device *dev, bool enable)
 {
-	uint32_t pull = nv_rd32(dev, NV04_PFIFO_CACHE1_PULL0);
+	int pull = nv_mask(dev, NV04_PFIFO_CACHE1_PULL0, 1, enable);
+
+	if (!enable) {
+		/* In some cases the PFIFO puller may be left in an
+		 * inconsistent state if you try to stop it when it's
+		 * busy translating handles. Sometimes you get a
+		 * PFIFO_CACHE_ERROR, sometimes it just fails silently
+		 * sending incorrect instance offsets to PGRAPH after
+		 * it's started up again. To avoid the latter we
+		 * invalidate the most recently calculated instance.
+		 */
+		if (!nv_wait(dev, NV04_PFIFO_CACHE1_PULL0,
+			     NV04_PFIFO_CACHE1_PULL0_HASH_BUSY, 0))
+			NV_ERROR(dev, "Timeout idling the PFIFO puller.\n");
+
+		if (nv_rd32(dev, NV04_PFIFO_CACHE1_PULL0) &
+		    NV04_PFIFO_CACHE1_PULL0_HASH_FAILED)
+			nv_wr32(dev, NV03_PFIFO_INTR_0,
+				NV_PFIFO_INTR_CACHE_ERROR);
 
-	if (enable) {
-		nv_wr32(dev, NV04_PFIFO_CACHE1_PULL0, pull | 1);
-	} else {
-		nv_wr32(dev, NV04_PFIFO_CACHE1_PULL0, pull & ~1);
 		nv_wr32(dev, NV04_PFIFO_CACHE1_HASH, 0);
 	}
 
-	return !!(pull & 1);
+	return pull & 1;
 }
 
 int
@@ -130,7 +124,7 @@ nv04_fifo_create_context(struct nouveau_channel *chan)
 						NV04_RAMFC__SIZE,
 						NVOBJ_FLAG_ZERO_ALLOC |
 						NVOBJ_FLAG_ZERO_FREE,
-						NULL, &chan->ramfc);
+						&chan->ramfc);
 	if (ret)
 		return ret;
 
@@ -139,7 +133,7 @@ nv04_fifo_create_context(struct nouveau_channel *chan)
 	/* Setup initial state */
 	RAMFC_WR(DMA_PUT, chan->pushbuf_base);
 	RAMFC_WR(DMA_GET, chan->pushbuf_base);
-	RAMFC_WR(DMA_INSTANCE, chan->pushbuf->instance >> 4);
+	RAMFC_WR(DMA_INSTANCE, chan->pushbuf->pinst >> 4);
 	RAMFC_WR(DMA_FETCH, (NV_PFIFO_CACHE1_DMA_FETCH_TRIG_128_BYTES |
 			     NV_PFIFO_CACHE1_DMA_FETCH_SIZE_128_BYTES |
 			     NV_PFIFO_CACHE1_DMA_FETCH_MAX_REQS_8 |
@@ -161,7 +155,7 @@ nv04_fifo_destroy_context(struct nouveau_channel *chan)
 	nv_wr32(dev, NV04_PFIFO_MODE,
 		nv_rd32(dev, NV04_PFIFO_MODE) & ~(1 << chan->id));
 
-	nouveau_gpuobj_ref_del(dev, &chan->ramfc);
+	nouveau_gpuobj_ref(NULL, &chan->ramfc);
 }
 
 static void
@@ -264,10 +258,10 @@ nv04_fifo_init_ramxx(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 
 	nv_wr32(dev, NV03_PFIFO_RAMHT, (0x03 << 24) /* search 128 */ |
-				       ((dev_priv->ramht_bits - 9) << 16) |
-				       (dev_priv->ramht_offset >> 8));
-	nv_wr32(dev, NV03_PFIFO_RAMRO, dev_priv->ramro_offset>>8);
-	nv_wr32(dev, NV03_PFIFO_RAMFC, dev_priv->ramfc_offset >> 8);
+				       ((dev_priv->ramht->bits - 9) << 16) |
+				       (dev_priv->ramht->gpuobj->pinst >> 8));
+	nv_wr32(dev, NV03_PFIFO_RAMRO, dev_priv->ramro->pinst >> 8);
+	nv_wr32(dev, NV03_PFIFO_RAMFC, dev_priv->ramfc->pinst >> 8);
 }
 
 static void
diff --git a/drivers/gpu/drm/nouveau/nv04_instmem.c b/drivers/gpu/drm/nouveau/nv04_instmem.c
index 4408232d33f1..0b5ae297abde 100644
--- a/drivers/gpu/drm/nouveau/nv04_instmem.c
+++ b/drivers/gpu/drm/nouveau/nv04_instmem.c
@@ -1,6 +1,7 @@
 #include "drmP.h"
 #include "drm.h"
 #include "nouveau_drv.h"
+#include "nouveau_ramht.h"
 
 /* returns the size of fifo context */
 static int
@@ -17,102 +18,51 @@ nouveau_fifo_ctx_size(struct drm_device *dev)
 	return 32;
 }
 
-static void
-nv04_instmem_determine_amount(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int i;
-
-	/* Figure out how much instance memory we need */
-	if (dev_priv->card_type >= NV_40) {
-		/* We'll want more instance memory than this on some NV4x cards.
-		 * There's a 16MB aperture to play with that maps onto the end
-		 * of vram.  For now, only reserve a small piece until we know
-		 * more about what each chipset requires.
-		 */
-		switch (dev_priv->chipset) {
-		case 0x40:
-		case 0x47:
-		case 0x49:
-		case 0x4b:
-			dev_priv->ramin_rsvd_vram = (2 * 1024 * 1024);
-			break;
-		default:
-			dev_priv->ramin_rsvd_vram = (1 * 1024 * 1024);
-			break;
-		}
-	} else {
-		/*XXX: what *are* the limits on <NV40 cards?
-		 */
-		dev_priv->ramin_rsvd_vram = (512 * 1024);
-	}
-	NV_DEBUG(dev, "RAMIN size: %dKiB\n", dev_priv->ramin_rsvd_vram >> 10);
-
-	/* Clear all of it, except the BIOS image that's in the first 64KiB */
-	for (i = 64 * 1024; i < dev_priv->ramin_rsvd_vram; i += 4)
-		nv_wi32(dev, i, 0x00000000);
-}
-
-static void
-nv04_instmem_configure_fixed_tables(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_engine *engine = &dev_priv->engine;
-
-	/* FIFO hash table (RAMHT)
-	 *   use 4k hash table at RAMIN+0x10000
-	 *   TODO: extend the hash table
-	 */
-	dev_priv->ramht_offset = 0x10000;
-	dev_priv->ramht_bits   = 9;
-	dev_priv->ramht_size   = (1 << dev_priv->ramht_bits); /* nr entries */
-	dev_priv->ramht_size  *= 8; /* 2 32-bit values per entry in RAMHT */
-	NV_DEBUG(dev, "RAMHT offset=0x%x, size=%d\n", dev_priv->ramht_offset,
-						      dev_priv->ramht_size);
-
-	/* FIFO runout table (RAMRO) - 512k at 0x11200 */
-	dev_priv->ramro_offset = 0x11200;
-	dev_priv->ramro_size   = 512;
-	NV_DEBUG(dev, "RAMRO offset=0x%x, size=%d\n", dev_priv->ramro_offset,
-						      dev_priv->ramro_size);
-
-	/* FIFO context table (RAMFC)
-	 *   NV40  : Not sure exactly how to position RAMFC on some cards,
-	 *           0x30002 seems to position it at RAMIN+0x20000 on these
-	 *           cards.  RAMFC is 4kb (32 fifos, 128byte entries).
-	 *   Others: Position RAMFC at RAMIN+0x11400
-	 */
-	dev_priv->ramfc_size = engine->fifo.channels *
-						nouveau_fifo_ctx_size(dev);
-	switch (dev_priv->card_type) {
-	case NV_40:
-		dev_priv->ramfc_offset = 0x20000;
-		break;
-	case NV_30:
-	case NV_20:
-	case NV_10:
-	case NV_04:
-	default:
-		dev_priv->ramfc_offset = 0x11400;
-		break;
-	}
-	NV_DEBUG(dev, "RAMFC offset=0x%x, size=%d\n", dev_priv->ramfc_offset,
-						      dev_priv->ramfc_size);
-}
-
 int nv04_instmem_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	uint32_t offset;
+	struct nouveau_gpuobj *ramht = NULL;
+	u32 offset, length;
 	int ret;
 
-	nv04_instmem_determine_amount(dev);
-	nv04_instmem_configure_fixed_tables(dev);
+	/* RAMIN always available */
+	dev_priv->ramin_available = true;
 
-	/* Create a heap to manage RAMIN allocations, we don't allocate
-	 * the space that was reserved for RAMHT/FC/RO.
-	 */
-	offset = dev_priv->ramfc_offset + dev_priv->ramfc_size;
+	/* Setup shared RAMHT */
+	ret = nouveau_gpuobj_new_fake(dev, 0x10000, ~0, 4096,
+				      NVOBJ_FLAG_ZERO_ALLOC, &ramht);
+	if (ret)
+		return ret;
+
+	ret = nouveau_ramht_new(dev, ramht, &dev_priv->ramht);
+	nouveau_gpuobj_ref(NULL, &ramht);
+	if (ret)
+		return ret;
+
+	/* And RAMRO */
+	ret = nouveau_gpuobj_new_fake(dev, 0x11200, ~0, 512,
+				      NVOBJ_FLAG_ZERO_ALLOC, &dev_priv->ramro);
+	if (ret)
+		return ret;
+
+	/* And RAMFC */
+	length = dev_priv->engine.fifo.channels * nouveau_fifo_ctx_size(dev);
+	switch (dev_priv->card_type) {
+	case NV_40:
+		offset = 0x20000;
+		break;
+	default:
+		offset = 0x11400;
+		break;
+	}
+
+	ret = nouveau_gpuobj_new_fake(dev, offset, ~0, length,
+				      NVOBJ_FLAG_ZERO_ALLOC, &dev_priv->ramfc);
+	if (ret)
+		return ret;
+
+	/* Only allow space after RAMFC to be used for object allocation */
+	offset += length;
 
 	/* It appears RAMRO (or something?) is controlled by 0x2220/0x2230
 	 * on certain NV4x chipsets as well as RAMFC.  When 0x2230 == 0
@@ -140,46 +90,34 @@ int nv04_instmem_init(struct drm_device *dev)
 void
 nv04_instmem_takedown(struct drm_device *dev)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	nouveau_ramht_ref(NULL, &dev_priv->ramht, NULL);
+	nouveau_gpuobj_ref(NULL, &dev_priv->ramro);
+	nouveau_gpuobj_ref(NULL, &dev_priv->ramfc);
 }
 
 int
-nv04_instmem_populate(struct drm_device *dev, struct nouveau_gpuobj *gpuobj, uint32_t *sz)
+nv04_instmem_populate(struct drm_device *dev, struct nouveau_gpuobj *gpuobj,
+		      uint32_t *sz)
 {
-	if (gpuobj->im_backing)
-		return -EINVAL;
-
 	return 0;
 }
 
 void
 nv04_instmem_clear(struct drm_device *dev, struct nouveau_gpuobj *gpuobj)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	if (gpuobj && gpuobj->im_backing) {
-		if (gpuobj->im_bound)
-			dev_priv->engine.instmem.unbind(dev, gpuobj);
-		gpuobj->im_backing = NULL;
-	}
 }
 
 int
 nv04_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj)
 {
-	if (!gpuobj->im_pramin || gpuobj->im_bound)
-		return -EINVAL;
-
-	gpuobj->im_bound = 1;
 	return 0;
 }
 
 int
 nv04_instmem_unbind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj)
 {
-	if (gpuobj->im_bound == 0)
-		return -EINVAL;
-
-	gpuobj->im_bound = 0;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nv04_pm.c b/drivers/gpu/drm/nouveau/nv04_pm.c
new file mode 100644
index 000000000000..6a6eb697d38e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nv04_pm.c
@@ -0,0 +1,81 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "drmP.h"
+#include "nouveau_drv.h"
+#include "nouveau_hw.h"
+#include "nouveau_pm.h"
+
+struct nv04_pm_state {
+	struct pll_lims pll;
+	struct nouveau_pll_vals calc;
+};
+
+int
+nv04_pm_clock_get(struct drm_device *dev, u32 id)
+{
+	return nouveau_hw_get_clock(dev, id);
+}
+
+void *
+nv04_pm_clock_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+		  u32 id, int khz)
+{
+	struct nv04_pm_state *state;
+	int ret;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return ERR_PTR(-ENOMEM);
+
+	ret = get_pll_limits(dev, id, &state->pll);
+	if (ret) {
+		kfree(state);
+		return (ret == -ENOENT) ? NULL : ERR_PTR(ret);
+	}
+
+	ret = nouveau_calc_pll_mnp(dev, &state->pll, khz, &state->calc);
+	if (!ret) {
+		kfree(state);
+		return ERR_PTR(-EINVAL);
+	}
+
+	return state;
+}
+
+void
+nv04_pm_clock_set(struct drm_device *dev, void *pre_state)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nv04_pm_state *state = pre_state;
+	u32 reg = state->pll.reg;
+
+	/* thank the insane nouveau_hw_setpll() interface for this */
+	if (dev_priv->card_type >= NV_40)
+		reg += 4;
+
+	nouveau_hw_setpll(dev, reg, &state->calc);
+	kfree(state);
+}
+
diff --git a/drivers/gpu/drm/nouveau/nv04_tv.c b/drivers/gpu/drm/nouveau/nv04_tv.c
index 0b5d012d7c28..3eb605ddfd03 100644
--- a/drivers/gpu/drm/nouveau/nv04_tv.c
+++ b/drivers/gpu/drm/nouveau/nv04_tv.c
@@ -49,8 +49,8 @@ static struct i2c_board_info nv04_tv_encoder_info[] = {
 
 int nv04_tv_identify(struct drm_device *dev, int i2c_index)
 {
-	return nouveau_i2c_identify(dev, "TV encoder",
-				    nv04_tv_encoder_info, i2c_index);
+	return nouveau_i2c_identify(dev, "TV encoder", nv04_tv_encoder_info,
+				    NULL, i2c_index);
 }
 
 
@@ -99,12 +99,10 @@ static void nv04_tv_bind(struct drm_device *dev, int head, bool bind)
 
 	state->tv_setup = 0;
 
-	if (bind) {
-		state->CRTC[NV_CIO_CRE_LCD__INDEX] = 0;
+	if (bind)
 		state->CRTC[NV_CIO_CRE_49] |= 0x10;
-	} else {
+	else
 		state->CRTC[NV_CIO_CRE_49] &= ~0x10;
-	}
 
 	NVWriteVgaCrtc(dev, head, NV_CIO_CRE_LCD__INDEX,
 		       state->CRTC[NV_CIO_CRE_LCD__INDEX]);
diff --git a/drivers/gpu/drm/nouveau/nv10_fifo.c b/drivers/gpu/drm/nouveau/nv10_fifo.c
index 7a4069cf5d0b..f1b03ad58fd5 100644
--- a/drivers/gpu/drm/nouveau/nv10_fifo.c
+++ b/drivers/gpu/drm/nouveau/nv10_fifo.c
@@ -27,8 +27,9 @@
 #include "drmP.h"
 #include "drm.h"
 #include "nouveau_drv.h"
+#include "nouveau_ramht.h"
 
-#define NV10_RAMFC(c) (dev_priv->ramfc_offset + ((c) * NV10_RAMFC__SIZE))
+#define NV10_RAMFC(c) (dev_priv->ramfc->pinst + ((c) * NV10_RAMFC__SIZE))
 #define NV10_RAMFC__SIZE ((dev_priv->chipset) >= 0x17 ? 64 : 32)
 
 int
@@ -48,7 +49,7 @@ nv10_fifo_create_context(struct nouveau_channel *chan)
 
 	ret = nouveau_gpuobj_new_fake(dev, NV10_RAMFC(chan->id), ~0,
 				      NV10_RAMFC__SIZE, NVOBJ_FLAG_ZERO_ALLOC |
-				      NVOBJ_FLAG_ZERO_FREE, NULL, &chan->ramfc);
+				      NVOBJ_FLAG_ZERO_FREE, &chan->ramfc);
 	if (ret)
 		return ret;
 
@@ -57,7 +58,7 @@ nv10_fifo_create_context(struct nouveau_channel *chan)
 	 */
 	nv_wi32(dev, fc +  0, chan->pushbuf_base);
 	nv_wi32(dev, fc +  4, chan->pushbuf_base);
-	nv_wi32(dev, fc + 12, chan->pushbuf->instance >> 4);
+	nv_wi32(dev, fc + 12, chan->pushbuf->pinst >> 4);
 	nv_wi32(dev, fc + 20, NV_PFIFO_CACHE1_DMA_FETCH_TRIG_128_BYTES |
 			      NV_PFIFO_CACHE1_DMA_FETCH_SIZE_128_BYTES |
 			      NV_PFIFO_CACHE1_DMA_FETCH_MAX_REQS_8 |
@@ -80,7 +81,7 @@ nv10_fifo_destroy_context(struct nouveau_channel *chan)
 	nv_wr32(dev, NV04_PFIFO_MODE,
 			nv_rd32(dev, NV04_PFIFO_MODE) & ~(1 << chan->id));
 
-	nouveau_gpuobj_ref_del(dev, &chan->ramfc);
+	nouveau_gpuobj_ref(NULL, &chan->ramfc);
 }
 
 static void
@@ -202,14 +203,14 @@ nv10_fifo_init_ramxx(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 
 	nv_wr32(dev, NV03_PFIFO_RAMHT, (0x03 << 24) /* search 128 */ |
-				       ((dev_priv->ramht_bits - 9) << 16) |
-				       (dev_priv->ramht_offset >> 8));
-	nv_wr32(dev, NV03_PFIFO_RAMRO, dev_priv->ramro_offset>>8);
+				       ((dev_priv->ramht->bits - 9) << 16) |
+				       (dev_priv->ramht->gpuobj->pinst >> 8));
+	nv_wr32(dev, NV03_PFIFO_RAMRO, dev_priv->ramro->pinst >> 8);
 
 	if (dev_priv->chipset < 0x17) {
-		nv_wr32(dev, NV03_PFIFO_RAMFC, dev_priv->ramfc_offset >> 8);
+		nv_wr32(dev, NV03_PFIFO_RAMFC, dev_priv->ramfc->pinst >> 8);
 	} else {
-		nv_wr32(dev, NV03_PFIFO_RAMFC, (dev_priv->ramfc_offset >> 8) |
+		nv_wr32(dev, NV03_PFIFO_RAMFC, (dev_priv->ramfc->pinst >> 8) |
 					       (1 << 16) /* 64 Bytes entry*/);
 		/* XXX nvidia blob set bit 18, 21,23 for nv20 & nv30 */
 	}
diff --git a/drivers/gpu/drm/nouveau/nv10_graph.c b/drivers/gpu/drm/nouveau/nv10_graph.c
index b2f6a57c0cc5..8e68c9731159 100644
--- a/drivers/gpu/drm/nouveau/nv10_graph.c
+++ b/drivers/gpu/drm/nouveau/nv10_graph.c
@@ -803,7 +803,7 @@ nv10_graph_context_switch(struct drm_device *dev)
 	/* Load context for next channel */
 	chid = (nv_rd32(dev, NV04_PGRAPH_TRAPPED_ADDR) >> 20) & 0x1f;
 	chan = dev_priv->fifos[chid];
-	if (chan)
+	if (chan && chan->pgraph_ctx)
 		nv10_graph_load_context(chan);
 
 	pgraph->fifo_access(dev, true);
diff --git a/drivers/gpu/drm/nouveau/nv17_tv.c b/drivers/gpu/drm/nouveau/nv17_tv.c
index 13cdc05b7c2d..28119fd19d03 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv.c
+++ b/drivers/gpu/drm/nouveau/nv17_tv.c
@@ -193,55 +193,56 @@ nv17_tv_detect(struct drm_encoder *encoder, struct drm_connector *connector)
 	}
 }
 
-static const struct {
-	int hdisplay;
-	int vdisplay;
-} modes[] = {
-	{ 640, 400 },
-	{ 640, 480 },
-	{ 720, 480 },
-	{ 720, 576 },
-	{ 800, 600 },
-	{ 1024, 768 },
-	{ 1280, 720 },
-	{ 1280, 1024 },
-	{ 1920, 1080 }
-};
-
-static int nv17_tv_get_modes(struct drm_encoder *encoder,
-			     struct drm_connector *connector)
+static int nv17_tv_get_ld_modes(struct drm_encoder *encoder,
+				struct drm_connector *connector)
 {
 	struct nv17_tv_norm_params *tv_norm = get_tv_norm(encoder);
-	struct drm_display_mode *mode;
-	struct drm_display_mode *output_mode;
+	struct drm_display_mode *mode, *tv_mode;
 	int n = 0;
-	int i;
 
-	if (tv_norm->kind != CTV_ENC_MODE) {
-		struct drm_display_mode *tv_mode;
+	for (tv_mode = nv17_tv_modes; tv_mode->hdisplay; tv_mode++) {
+		mode = drm_mode_duplicate(encoder->dev, tv_mode);
 
-		for (tv_mode = nv17_tv_modes; tv_mode->hdisplay; tv_mode++) {
-			mode = drm_mode_duplicate(encoder->dev, tv_mode);
+		mode->clock = tv_norm->tv_enc_mode.vrefresh *
+			mode->htotal / 1000 *
+			mode->vtotal / 1000;
 
-			mode->clock = tv_norm->tv_enc_mode.vrefresh *
-						mode->htotal / 1000 *
-						mode->vtotal / 1000;
+		if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+			mode->clock *= 2;
 
-			if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
-				mode->clock *= 2;
+		if (mode->hdisplay == tv_norm->tv_enc_mode.hdisplay &&
+		    mode->vdisplay == tv_norm->tv_enc_mode.vdisplay)
+			mode->type |= DRM_MODE_TYPE_PREFERRED;
 
-			if (mode->hdisplay == tv_norm->tv_enc_mode.hdisplay &&
-			    mode->vdisplay == tv_norm->tv_enc_mode.vdisplay)
-				mode->type |= DRM_MODE_TYPE_PREFERRED;
-
-			drm_mode_probed_add(connector, mode);
-			n++;
-		}
-		return n;
+		drm_mode_probed_add(connector, mode);
+		n++;
 	}
 
-	/* tv_norm->kind == CTV_ENC_MODE */
-	output_mode = &tv_norm->ctv_enc_mode.mode;
+	return n;
+}
+
+static int nv17_tv_get_hd_modes(struct drm_encoder *encoder,
+				struct drm_connector *connector)
+{
+	struct nv17_tv_norm_params *tv_norm = get_tv_norm(encoder);
+	struct drm_display_mode *output_mode = &tv_norm->ctv_enc_mode.mode;
+	struct drm_display_mode *mode;
+	const struct {
+		int hdisplay;
+		int vdisplay;
+	} modes[] = {
+		{ 640, 400 },
+		{ 640, 480 },
+		{ 720, 480 },
+		{ 720, 576 },
+		{ 800, 600 },
+		{ 1024, 768 },
+		{ 1280, 720 },
+		{ 1280, 1024 },
+		{ 1920, 1080 }
+	};
+	int i, n = 0;
+
 	for (i = 0; i < ARRAY_SIZE(modes); i++) {
 		if (modes[i].hdisplay > output_mode->hdisplay ||
 		    modes[i].vdisplay > output_mode->vdisplay)
@@ -251,11 +252,12 @@ static int nv17_tv_get_modes(struct drm_encoder *encoder,
 		    modes[i].vdisplay == output_mode->vdisplay) {
 			mode = drm_mode_duplicate(encoder->dev, output_mode);
 			mode->type |= DRM_MODE_TYPE_PREFERRED;
+
 		} else {
 			mode = drm_cvt_mode(encoder->dev, modes[i].hdisplay,
-				modes[i].vdisplay, 60, false,
-				output_mode->flags & DRM_MODE_FLAG_INTERLACE,
-				false);
+					    modes[i].vdisplay, 60, false,
+					    (output_mode->flags &
+					     DRM_MODE_FLAG_INTERLACE), false);
 		}
 
 		/* CVT modes are sometimes unsuitable... */
@@ -266,6 +268,7 @@ static int nv17_tv_get_modes(struct drm_encoder *encoder,
 					     - mode->hdisplay) * 9 / 10) & ~7;
 			mode->hsync_end = mode->hsync_start + 8;
 		}
+
 		if (output_mode->vdisplay >= 1024) {
 			mode->vtotal = output_mode->vtotal;
 			mode->vsync_start = output_mode->vsync_start;
@@ -276,9 +279,21 @@ static int nv17_tv_get_modes(struct drm_encoder *encoder,
 		drm_mode_probed_add(connector, mode);
 		n++;
 	}
+
 	return n;
 }
 
+static int nv17_tv_get_modes(struct drm_encoder *encoder,
+			     struct drm_connector *connector)
+{
+	struct nv17_tv_norm_params *tv_norm = get_tv_norm(encoder);
+
+	if (tv_norm->kind == CTV_ENC_MODE)
+		return nv17_tv_get_hd_modes(encoder, connector);
+	else
+		return nv17_tv_get_ld_modes(encoder, connector);
+}
+
 static int nv17_tv_mode_valid(struct drm_encoder *encoder,
 			      struct drm_display_mode *mode)
 {
@@ -408,15 +423,8 @@ static void nv17_tv_prepare(struct drm_encoder *encoder)
 
 	}
 
-	/* Some NV4x have unknown values (0x3f, 0x50, 0x54, 0x6b, 0x79, 0x7f)
-	 * at LCD__INDEX which we don't alter
-	 */
-	if (!(*cr_lcd & 0x44)) {
-		if (tv_norm->kind == CTV_ENC_MODE)
-			*cr_lcd = 0x1 | (head ? 0x0 : 0x8);
-		else
-			*cr_lcd = 0;
-	}
+	if (tv_norm->kind == CTV_ENC_MODE)
+		*cr_lcd |= 0x1 | (head ? 0x0 : 0x8);
 
 	/* Set the DACCLK register */
 	dacclk = (NVReadRAMDAC(dev, 0, dacclk_off) & ~0x30) | 0x1;
diff --git a/drivers/gpu/drm/nouveau/nv17_tv.h b/drivers/gpu/drm/nouveau/nv17_tv.h
index c00977cedabd..6bf03840f9eb 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv.h
+++ b/drivers/gpu/drm/nouveau/nv17_tv.h
@@ -127,7 +127,8 @@ void nv17_ctv_update_rescaler(struct drm_encoder *encoder);
 
 /* TV hardware access functions */
 
-static inline void nv_write_ptv(struct drm_device *dev, uint32_t reg, uint32_t val)
+static inline void nv_write_ptv(struct drm_device *dev, uint32_t reg,
+				uint32_t val)
 {
 	nv_wr32(dev, reg, val);
 }
@@ -137,7 +138,8 @@ static inline uint32_t nv_read_ptv(struct drm_device *dev, uint32_t reg)
 	return nv_rd32(dev, reg);
 }
 
-static inline void nv_write_tv_enc(struct drm_device *dev, uint8_t reg, uint8_t val)
+static inline void nv_write_tv_enc(struct drm_device *dev, uint8_t reg,
+				   uint8_t val)
 {
 	nv_write_ptv(dev, NV_PTV_TV_INDEX, reg);
 	nv_write_ptv(dev, NV_PTV_TV_DATA, val);
@@ -149,8 +151,11 @@ static inline uint8_t nv_read_tv_enc(struct drm_device *dev, uint8_t reg)
 	return nv_read_ptv(dev, NV_PTV_TV_DATA);
 }
 
-#define nv_load_ptv(dev, state, reg) nv_write_ptv(dev, NV_PTV_OFFSET + 0x##reg, state->ptv_##reg)
-#define nv_save_ptv(dev, state, reg) state->ptv_##reg = nv_read_ptv(dev, NV_PTV_OFFSET + 0x##reg)
-#define nv_load_tv_enc(dev, state, reg) nv_write_tv_enc(dev, 0x##reg, state->tv_enc[0x##reg])
+#define nv_load_ptv(dev, state, reg) \
+	nv_write_ptv(dev, NV_PTV_OFFSET + 0x##reg, state->ptv_##reg)
+#define nv_save_ptv(dev, state, reg) \
+	state->ptv_##reg = nv_read_ptv(dev, NV_PTV_OFFSET + 0x##reg)
+#define nv_load_tv_enc(dev, state, reg) \
+	nv_write_tv_enc(dev, 0x##reg, state->tv_enc[0x##reg])
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/nv17_tv_modes.c b/drivers/gpu/drm/nouveau/nv17_tv_modes.c
index d64683d97e0d..9d3893c50a41 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv_modes.c
+++ b/drivers/gpu/drm/nouveau/nv17_tv_modes.c
@@ -336,12 +336,17 @@ static void tv_setup_filter(struct drm_encoder *encoder)
 			struct filter_params *p = &fparams[k][j];
 
 			for (i = 0; i < 7; i++) {
-				int64_t c = (p->k1 + p->ki*i + p->ki2*i*i + p->ki3*i*i*i)
-					+ (p->kr + p->kir*i + p->ki2r*i*i + p->ki3r*i*i*i)*rs[k]
-					+ (p->kf + p->kif*i + p->ki2f*i*i + p->ki3f*i*i*i)*flicker
-					+ (p->krf + p->kirf*i + p->ki2rf*i*i + p->ki3rf*i*i*i)*flicker*rs[k];
+				int64_t c = (p->k1 + p->ki*i + p->ki2*i*i +
+					     p->ki3*i*i*i)
+					+ (p->kr + p->kir*i + p->ki2r*i*i +
+					   p->ki3r*i*i*i) * rs[k]
+					+ (p->kf + p->kif*i + p->ki2f*i*i +
+					   p->ki3f*i*i*i) * flicker
+					+ (p->krf + p->kirf*i + p->ki2rf*i*i +
+					   p->ki3rf*i*i*i) * flicker * rs[k];
 
-				(*filters[k])[j][i] = (c + id5/2) >> 39 & (0x1 << 31 | 0x7f << 9);
+				(*filters[k])[j][i] = (c + id5/2) >> 39
+					& (0x1 << 31 | 0x7f << 9);
 			}
 		}
 	}
@@ -349,7 +354,8 @@ static void tv_setup_filter(struct drm_encoder *encoder)
 
 /* Hardware state saving/restoring */
 
-static void tv_save_filter(struct drm_device *dev, uint32_t base, uint32_t regs[4][7])
+static void tv_save_filter(struct drm_device *dev, uint32_t base,
+			   uint32_t regs[4][7])
 {
 	int i, j;
 	uint32_t offsets[] = { base, base + 0x1c, base + 0x40, base + 0x5c };
@@ -360,7 +366,8 @@ static void tv_save_filter(struct drm_device *dev, uint32_t base, uint32_t regs[
 	}
 }
 
-static void tv_load_filter(struct drm_device *dev, uint32_t base, uint32_t regs[4][7])
+static void tv_load_filter(struct drm_device *dev, uint32_t base,
+			   uint32_t regs[4][7])
 {
 	int i, j;
 	uint32_t offsets[] = { base, base + 0x1c, base + 0x40, base + 0x5c };
@@ -504,10 +511,10 @@ void nv17_tv_update_properties(struct drm_encoder *encoder)
 		break;
 	}
 
-	regs->tv_enc[0x20] = interpolate(0, tv_norm->tv_enc_mode.tv_enc[0x20], 255,
-					 tv_enc->saturation);
-	regs->tv_enc[0x22] = interpolate(0, tv_norm->tv_enc_mode.tv_enc[0x22], 255,
-					 tv_enc->saturation);
+	regs->tv_enc[0x20] = interpolate(0, tv_norm->tv_enc_mode.tv_enc[0x20],
+					 255, tv_enc->saturation);
+	regs->tv_enc[0x22] = interpolate(0, tv_norm->tv_enc_mode.tv_enc[0x22],
+					 255, tv_enc->saturation);
 	regs->tv_enc[0x25] = tv_enc->hue * 255 / 100;
 
 	nv_load_ptv(dev, regs, 204);
@@ -541,7 +548,8 @@ void nv17_ctv_update_rescaler(struct drm_encoder *encoder)
 	int head = nouveau_crtc(encoder->crtc)->index;
 	struct nv04_crtc_reg *regs = &dev_priv->mode_reg.crtc_reg[head];
 	struct drm_display_mode *crtc_mode = &encoder->crtc->mode;
-	struct drm_display_mode *output_mode = &get_tv_norm(encoder)->ctv_enc_mode.mode;
+	struct drm_display_mode *output_mode =
+		&get_tv_norm(encoder)->ctv_enc_mode.mode;
 	int overscan, hmargin, vmargin, hratio, vratio;
 
 	/* The rescaler doesn't do the right thing for interlaced modes. */
@@ -553,13 +561,15 @@ void nv17_ctv_update_rescaler(struct drm_encoder *encoder)
 	hmargin = (output_mode->hdisplay - crtc_mode->hdisplay) / 2;
 	vmargin = (output_mode->vdisplay - crtc_mode->vdisplay) / 2;
 
-	hmargin = interpolate(0, min(hmargin, output_mode->hdisplay/20), hmargin,
-			      overscan);
-	vmargin = interpolate(0, min(vmargin, output_mode->vdisplay/20), vmargin,
-			      overscan);
+	hmargin = interpolate(0, min(hmargin, output_mode->hdisplay/20),
+			      hmargin, overscan);
+	vmargin = interpolate(0, min(vmargin, output_mode->vdisplay/20),
+			      vmargin, overscan);
 
-	hratio = crtc_mode->hdisplay * 0x800 / (output_mode->hdisplay - 2*hmargin);
-	vratio = crtc_mode->vdisplay * 0x800 / (output_mode->vdisplay - 2*vmargin) & ~3;
+	hratio = crtc_mode->hdisplay * 0x800 /
+		(output_mode->hdisplay - 2*hmargin);
+	vratio = crtc_mode->vdisplay * 0x800 /
+		(output_mode->vdisplay - 2*vmargin) & ~3;
 
 	regs->fp_horiz_regs[FP_VALID_START] = hmargin;
 	regs->fp_horiz_regs[FP_VALID_END] = output_mode->hdisplay - hmargin - 1;
diff --git a/drivers/gpu/drm/nouveau/nv20_graph.c b/drivers/gpu/drm/nouveau/nv20_graph.c
index 17f309b36c91..12ab9cd56eca 100644
--- a/drivers/gpu/drm/nouveau/nv20_graph.c
+++ b/drivers/gpu/drm/nouveau/nv20_graph.c
@@ -37,49 +37,49 @@ nv20_graph_context_init(struct drm_device *dev, struct nouveau_gpuobj *ctx)
 {
 	int i;
 
-	nv_wo32(dev, ctx, 0x033c/4, 0xffff0000);
-	nv_wo32(dev, ctx, 0x03a0/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x03a4/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x047c/4, 0x00000101);
-	nv_wo32(dev, ctx, 0x0490/4, 0x00000111);
-	nv_wo32(dev, ctx, 0x04a8/4, 0x44400000);
+	nv_wo32(ctx, 0x033c, 0xffff0000);
+	nv_wo32(ctx, 0x03a0, 0x0fff0000);
+	nv_wo32(ctx, 0x03a4, 0x0fff0000);
+	nv_wo32(ctx, 0x047c, 0x00000101);
+	nv_wo32(ctx, 0x0490, 0x00000111);
+	nv_wo32(ctx, 0x04a8, 0x44400000);
 	for (i = 0x04d4; i <= 0x04e0; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00030303);
+		nv_wo32(ctx, i, 0x00030303);
 	for (i = 0x04f4; i <= 0x0500; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080000);
+		nv_wo32(ctx, i, 0x00080000);
 	for (i = 0x050c; i <= 0x0518; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x01012000);
+		nv_wo32(ctx, i, 0x01012000);
 	for (i = 0x051c; i <= 0x0528; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x000105b8);
+		nv_wo32(ctx, i, 0x000105b8);
 	for (i = 0x052c; i <= 0x0538; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080008);
+		nv_wo32(ctx, i, 0x00080008);
 	for (i = 0x055c; i <= 0x0598; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x07ff0000);
-	nv_wo32(dev, ctx, 0x05a4/4, 0x4b7fffff);
-	nv_wo32(dev, ctx, 0x05fc/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x0604/4, 0x00004000);
-	nv_wo32(dev, ctx, 0x0610/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x0618/4, 0x00040000);
-	nv_wo32(dev, ctx, 0x061c/4, 0x00010000);
+		nv_wo32(ctx, i, 0x07ff0000);
+	nv_wo32(ctx, 0x05a4, 0x4b7fffff);
+	nv_wo32(ctx, 0x05fc, 0x00000001);
+	nv_wo32(ctx, 0x0604, 0x00004000);
+	nv_wo32(ctx, 0x0610, 0x00000001);
+	nv_wo32(ctx, 0x0618, 0x00040000);
+	nv_wo32(ctx, 0x061c, 0x00010000);
 	for (i = 0x1c1c; i <= 0x248c; i += 16) {
-		nv_wo32(dev, ctx, (i + 0)/4, 0x10700ff9);
-		nv_wo32(dev, ctx, (i + 4)/4, 0x0436086c);
-		nv_wo32(dev, ctx, (i + 8)/4, 0x000c001b);
+		nv_wo32(ctx, (i + 0), 0x10700ff9);
+		nv_wo32(ctx, (i + 4), 0x0436086c);
+		nv_wo32(ctx, (i + 8), 0x000c001b);
 	}
-	nv_wo32(dev, ctx, 0x281c/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2830/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x285c/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x2860/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2864/4, 0x3f000000);
-	nv_wo32(dev, ctx, 0x286c/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x2870/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2878/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x2880/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x34a4/4, 0x000fe000);
-	nv_wo32(dev, ctx, 0x3530/4, 0x000003f8);
-	nv_wo32(dev, ctx, 0x3540/4, 0x002fe000);
+	nv_wo32(ctx, 0x281c, 0x3f800000);
+	nv_wo32(ctx, 0x2830, 0x3f800000);
+	nv_wo32(ctx, 0x285c, 0x40000000);
+	nv_wo32(ctx, 0x2860, 0x3f800000);
+	nv_wo32(ctx, 0x2864, 0x3f000000);
+	nv_wo32(ctx, 0x286c, 0x40000000);
+	nv_wo32(ctx, 0x2870, 0x3f800000);
+	nv_wo32(ctx, 0x2878, 0xbf800000);
+	nv_wo32(ctx, 0x2880, 0xbf800000);
+	nv_wo32(ctx, 0x34a4, 0x000fe000);
+	nv_wo32(ctx, 0x3530, 0x000003f8);
+	nv_wo32(ctx, 0x3540, 0x002fe000);
 	for (i = 0x355c; i <= 0x3578; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x001c527c);
+		nv_wo32(ctx, i, 0x001c527c);
 }
 
 static void
@@ -87,58 +87,58 @@ nv25_graph_context_init(struct drm_device *dev, struct nouveau_gpuobj *ctx)
 {
 	int i;
 
-	nv_wo32(dev, ctx, 0x035c/4, 0xffff0000);
-	nv_wo32(dev, ctx, 0x03c0/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x03c4/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x049c/4, 0x00000101);
-	nv_wo32(dev, ctx, 0x04b0/4, 0x00000111);
-	nv_wo32(dev, ctx, 0x04c8/4, 0x00000080);
-	nv_wo32(dev, ctx, 0x04cc/4, 0xffff0000);
-	nv_wo32(dev, ctx, 0x04d0/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x04e4/4, 0x44400000);
-	nv_wo32(dev, ctx, 0x04fc/4, 0x4b800000);
+	nv_wo32(ctx, 0x035c, 0xffff0000);
+	nv_wo32(ctx, 0x03c0, 0x0fff0000);
+	nv_wo32(ctx, 0x03c4, 0x0fff0000);
+	nv_wo32(ctx, 0x049c, 0x00000101);
+	nv_wo32(ctx, 0x04b0, 0x00000111);
+	nv_wo32(ctx, 0x04c8, 0x00000080);
+	nv_wo32(ctx, 0x04cc, 0xffff0000);
+	nv_wo32(ctx, 0x04d0, 0x00000001);
+	nv_wo32(ctx, 0x04e4, 0x44400000);
+	nv_wo32(ctx, 0x04fc, 0x4b800000);
 	for (i = 0x0510; i <= 0x051c; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00030303);
+		nv_wo32(ctx, i, 0x00030303);
 	for (i = 0x0530; i <= 0x053c; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080000);
+		nv_wo32(ctx, i, 0x00080000);
 	for (i = 0x0548; i <= 0x0554; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x01012000);
+		nv_wo32(ctx, i, 0x01012000);
 	for (i = 0x0558; i <= 0x0564; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x000105b8);
+		nv_wo32(ctx, i, 0x000105b8);
 	for (i = 0x0568; i <= 0x0574; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080008);
+		nv_wo32(ctx, i, 0x00080008);
 	for (i = 0x0598; i <= 0x05d4; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x07ff0000);
-	nv_wo32(dev, ctx, 0x05e0/4, 0x4b7fffff);
-	nv_wo32(dev, ctx, 0x0620/4, 0x00000080);
-	nv_wo32(dev, ctx, 0x0624/4, 0x30201000);
-	nv_wo32(dev, ctx, 0x0628/4, 0x70605040);
-	nv_wo32(dev, ctx, 0x062c/4, 0xb0a09080);
-	nv_wo32(dev, ctx, 0x0630/4, 0xf0e0d0c0);
-	nv_wo32(dev, ctx, 0x0664/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x066c/4, 0x00004000);
-	nv_wo32(dev, ctx, 0x0678/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x0680/4, 0x00040000);
-	nv_wo32(dev, ctx, 0x0684/4, 0x00010000);
+		nv_wo32(ctx, i, 0x07ff0000);
+	nv_wo32(ctx, 0x05e0, 0x4b7fffff);
+	nv_wo32(ctx, 0x0620, 0x00000080);
+	nv_wo32(ctx, 0x0624, 0x30201000);
+	nv_wo32(ctx, 0x0628, 0x70605040);
+	nv_wo32(ctx, 0x062c, 0xb0a09080);
+	nv_wo32(ctx, 0x0630, 0xf0e0d0c0);
+	nv_wo32(ctx, 0x0664, 0x00000001);
+	nv_wo32(ctx, 0x066c, 0x00004000);
+	nv_wo32(ctx, 0x0678, 0x00000001);
+	nv_wo32(ctx, 0x0680, 0x00040000);
+	nv_wo32(ctx, 0x0684, 0x00010000);
 	for (i = 0x1b04; i <= 0x2374; i += 16) {
-		nv_wo32(dev, ctx, (i + 0)/4, 0x10700ff9);
-		nv_wo32(dev, ctx, (i + 4)/4, 0x0436086c);
-		nv_wo32(dev, ctx, (i + 8)/4, 0x000c001b);
+		nv_wo32(ctx, (i + 0), 0x10700ff9);
+		nv_wo32(ctx, (i + 4), 0x0436086c);
+		nv_wo32(ctx, (i + 8), 0x000c001b);
 	}
-	nv_wo32(dev, ctx, 0x2704/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2718/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2744/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x2748/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x274c/4, 0x3f000000);
-	nv_wo32(dev, ctx, 0x2754/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x2758/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2760/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x2768/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x308c/4, 0x000fe000);
-	nv_wo32(dev, ctx, 0x3108/4, 0x000003f8);
-	nv_wo32(dev, ctx, 0x3468/4, 0x002fe000);
+	nv_wo32(ctx, 0x2704, 0x3f800000);
+	nv_wo32(ctx, 0x2718, 0x3f800000);
+	nv_wo32(ctx, 0x2744, 0x40000000);
+	nv_wo32(ctx, 0x2748, 0x3f800000);
+	nv_wo32(ctx, 0x274c, 0x3f000000);
+	nv_wo32(ctx, 0x2754, 0x40000000);
+	nv_wo32(ctx, 0x2758, 0x3f800000);
+	nv_wo32(ctx, 0x2760, 0xbf800000);
+	nv_wo32(ctx, 0x2768, 0xbf800000);
+	nv_wo32(ctx, 0x308c, 0x000fe000);
+	nv_wo32(ctx, 0x3108, 0x000003f8);
+	nv_wo32(ctx, 0x3468, 0x002fe000);
 	for (i = 0x3484; i <= 0x34a0; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x001c527c);
+		nv_wo32(ctx, i, 0x001c527c);
 }
 
 static void
@@ -146,49 +146,49 @@ nv2a_graph_context_init(struct drm_device *dev, struct nouveau_gpuobj *ctx)
 {
 	int i;
 
-	nv_wo32(dev, ctx, 0x033c/4, 0xffff0000);
-	nv_wo32(dev, ctx, 0x03a0/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x03a4/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x047c/4, 0x00000101);
-	nv_wo32(dev, ctx, 0x0490/4, 0x00000111);
-	nv_wo32(dev, ctx, 0x04a8/4, 0x44400000);
+	nv_wo32(ctx, 0x033c, 0xffff0000);
+	nv_wo32(ctx, 0x03a0, 0x0fff0000);
+	nv_wo32(ctx, 0x03a4, 0x0fff0000);
+	nv_wo32(ctx, 0x047c, 0x00000101);
+	nv_wo32(ctx, 0x0490, 0x00000111);
+	nv_wo32(ctx, 0x04a8, 0x44400000);
 	for (i = 0x04d4; i <= 0x04e0; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00030303);
+		nv_wo32(ctx, i, 0x00030303);
 	for (i = 0x04f4; i <= 0x0500; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080000);
+		nv_wo32(ctx, i, 0x00080000);
 	for (i = 0x050c; i <= 0x0518; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x01012000);
+		nv_wo32(ctx, i, 0x01012000);
 	for (i = 0x051c; i <= 0x0528; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x000105b8);
+		nv_wo32(ctx, i, 0x000105b8);
 	for (i = 0x052c; i <= 0x0538; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080008);
+		nv_wo32(ctx, i, 0x00080008);
 	for (i = 0x055c; i <= 0x0598; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x07ff0000);
-	nv_wo32(dev, ctx, 0x05a4/4, 0x4b7fffff);
-	nv_wo32(dev, ctx, 0x05fc/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x0604/4, 0x00004000);
-	nv_wo32(dev, ctx, 0x0610/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x0618/4, 0x00040000);
-	nv_wo32(dev, ctx, 0x061c/4, 0x00010000);
+		nv_wo32(ctx, i, 0x07ff0000);
+	nv_wo32(ctx, 0x05a4, 0x4b7fffff);
+	nv_wo32(ctx, 0x05fc, 0x00000001);
+	nv_wo32(ctx, 0x0604, 0x00004000);
+	nv_wo32(ctx, 0x0610, 0x00000001);
+	nv_wo32(ctx, 0x0618, 0x00040000);
+	nv_wo32(ctx, 0x061c, 0x00010000);
 	for (i = 0x1a9c; i <= 0x22fc; i += 16) { /*XXX: check!! */
-		nv_wo32(dev, ctx, (i + 0)/4, 0x10700ff9);
-		nv_wo32(dev, ctx, (i + 4)/4, 0x0436086c);
-		nv_wo32(dev, ctx, (i + 8)/4, 0x000c001b);
+		nv_wo32(ctx, (i + 0), 0x10700ff9);
+		nv_wo32(ctx, (i + 4), 0x0436086c);
+		nv_wo32(ctx, (i + 8), 0x000c001b);
 	}
-	nv_wo32(dev, ctx, 0x269c/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x26b0/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x26dc/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x26e0/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x26e4/4, 0x3f000000);
-	nv_wo32(dev, ctx, 0x26ec/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x26f0/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x26f8/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x2700/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x3024/4, 0x000fe000);
-	nv_wo32(dev, ctx, 0x30a0/4, 0x000003f8);
-	nv_wo32(dev, ctx, 0x33fc/4, 0x002fe000);
+	nv_wo32(ctx, 0x269c, 0x3f800000);
+	nv_wo32(ctx, 0x26b0, 0x3f800000);
+	nv_wo32(ctx, 0x26dc, 0x40000000);
+	nv_wo32(ctx, 0x26e0, 0x3f800000);
+	nv_wo32(ctx, 0x26e4, 0x3f000000);
+	nv_wo32(ctx, 0x26ec, 0x40000000);
+	nv_wo32(ctx, 0x26f0, 0x3f800000);
+	nv_wo32(ctx, 0x26f8, 0xbf800000);
+	nv_wo32(ctx, 0x2700, 0xbf800000);
+	nv_wo32(ctx, 0x3024, 0x000fe000);
+	nv_wo32(ctx, 0x30a0, 0x000003f8);
+	nv_wo32(ctx, 0x33fc, 0x002fe000);
 	for (i = 0x341c; i <= 0x3438; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x001c527c);
+		nv_wo32(ctx, i, 0x001c527c);
 }
 
 static void
@@ -196,57 +196,57 @@ nv30_31_graph_context_init(struct drm_device *dev, struct nouveau_gpuobj *ctx)
 {
 	int i;
 
-	nv_wo32(dev, ctx, 0x0410/4, 0x00000101);
-	nv_wo32(dev, ctx, 0x0424/4, 0x00000111);
-	nv_wo32(dev, ctx, 0x0428/4, 0x00000060);
-	nv_wo32(dev, ctx, 0x0444/4, 0x00000080);
-	nv_wo32(dev, ctx, 0x0448/4, 0xffff0000);
-	nv_wo32(dev, ctx, 0x044c/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x0460/4, 0x44400000);
-	nv_wo32(dev, ctx, 0x048c/4, 0xffff0000);
+	nv_wo32(ctx, 0x0410, 0x00000101);
+	nv_wo32(ctx, 0x0424, 0x00000111);
+	nv_wo32(ctx, 0x0428, 0x00000060);
+	nv_wo32(ctx, 0x0444, 0x00000080);
+	nv_wo32(ctx, 0x0448, 0xffff0000);
+	nv_wo32(ctx, 0x044c, 0x00000001);
+	nv_wo32(ctx, 0x0460, 0x44400000);
+	nv_wo32(ctx, 0x048c, 0xffff0000);
 	for (i = 0x04e0; i < 0x04e8; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x04ec/4, 0x00011100);
+		nv_wo32(ctx, i, 0x0fff0000);
+	nv_wo32(ctx, 0x04ec, 0x00011100);
 	for (i = 0x0508; i < 0x0548; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x07ff0000);
-	nv_wo32(dev, ctx, 0x0550/4, 0x4b7fffff);
-	nv_wo32(dev, ctx, 0x058c/4, 0x00000080);
-	nv_wo32(dev, ctx, 0x0590/4, 0x30201000);
-	nv_wo32(dev, ctx, 0x0594/4, 0x70605040);
-	nv_wo32(dev, ctx, 0x0598/4, 0xb8a89888);
-	nv_wo32(dev, ctx, 0x059c/4, 0xf8e8d8c8);
-	nv_wo32(dev, ctx, 0x05b0/4, 0xb0000000);
+		nv_wo32(ctx, i, 0x07ff0000);
+	nv_wo32(ctx, 0x0550, 0x4b7fffff);
+	nv_wo32(ctx, 0x058c, 0x00000080);
+	nv_wo32(ctx, 0x0590, 0x30201000);
+	nv_wo32(ctx, 0x0594, 0x70605040);
+	nv_wo32(ctx, 0x0598, 0xb8a89888);
+	nv_wo32(ctx, 0x059c, 0xf8e8d8c8);
+	nv_wo32(ctx, 0x05b0, 0xb0000000);
 	for (i = 0x0600; i < 0x0640; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00010588);
+		nv_wo32(ctx, i, 0x00010588);
 	for (i = 0x0640; i < 0x0680; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00030303);
+		nv_wo32(ctx, i, 0x00030303);
 	for (i = 0x06c0; i < 0x0700; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0008aae4);
+		nv_wo32(ctx, i, 0x0008aae4);
 	for (i = 0x0700; i < 0x0740; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x01012000);
+		nv_wo32(ctx, i, 0x01012000);
 	for (i = 0x0740; i < 0x0780; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080008);
-	nv_wo32(dev, ctx, 0x085c/4, 0x00040000);
-	nv_wo32(dev, ctx, 0x0860/4, 0x00010000);
+		nv_wo32(ctx, i, 0x00080008);
+	nv_wo32(ctx, 0x085c, 0x00040000);
+	nv_wo32(ctx, 0x0860, 0x00010000);
 	for (i = 0x0864; i < 0x0874; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00040004);
+		nv_wo32(ctx, i, 0x00040004);
 	for (i = 0x1f18; i <= 0x3088 ; i += 16) {
-		nv_wo32(dev, ctx, i/4 + 0, 0x10700ff9);
-		nv_wo32(dev, ctx, i/4 + 1, 0x0436086c);
-		nv_wo32(dev, ctx, i/4 + 2, 0x000c001b);
+		nv_wo32(ctx, i + 0, 0x10700ff9);
+		nv_wo32(ctx, i + 1, 0x0436086c);
+		nv_wo32(ctx, i + 2, 0x000c001b);
 	}
 	for (i = 0x30b8; i < 0x30c8; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0000ffff);
-	nv_wo32(dev, ctx, 0x344c/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x3808/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x381c/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x3848/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x384c/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x3850/4, 0x3f000000);
-	nv_wo32(dev, ctx, 0x3858/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x385c/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x3864/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x386c/4, 0xbf800000);
+		nv_wo32(ctx, i, 0x0000ffff);
+	nv_wo32(ctx, 0x344c, 0x3f800000);
+	nv_wo32(ctx, 0x3808, 0x3f800000);
+	nv_wo32(ctx, 0x381c, 0x3f800000);
+	nv_wo32(ctx, 0x3848, 0x40000000);
+	nv_wo32(ctx, 0x384c, 0x3f800000);
+	nv_wo32(ctx, 0x3850, 0x3f000000);
+	nv_wo32(ctx, 0x3858, 0x40000000);
+	nv_wo32(ctx, 0x385c, 0x3f800000);
+	nv_wo32(ctx, 0x3864, 0xbf800000);
+	nv_wo32(ctx, 0x386c, 0xbf800000);
 }
 
 static void
@@ -254,57 +254,57 @@ nv34_graph_context_init(struct drm_device *dev, struct nouveau_gpuobj *ctx)
 {
 	int i;
 
-	nv_wo32(dev, ctx, 0x040c/4, 0x01000101);
-	nv_wo32(dev, ctx, 0x0420/4, 0x00000111);
-	nv_wo32(dev, ctx, 0x0424/4, 0x00000060);
-	nv_wo32(dev, ctx, 0x0440/4, 0x00000080);
-	nv_wo32(dev, ctx, 0x0444/4, 0xffff0000);
-	nv_wo32(dev, ctx, 0x0448/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x045c/4, 0x44400000);
-	nv_wo32(dev, ctx, 0x0480/4, 0xffff0000);
+	nv_wo32(ctx, 0x040c, 0x01000101);
+	nv_wo32(ctx, 0x0420, 0x00000111);
+	nv_wo32(ctx, 0x0424, 0x00000060);
+	nv_wo32(ctx, 0x0440, 0x00000080);
+	nv_wo32(ctx, 0x0444, 0xffff0000);
+	nv_wo32(ctx, 0x0448, 0x00000001);
+	nv_wo32(ctx, 0x045c, 0x44400000);
+	nv_wo32(ctx, 0x0480, 0xffff0000);
 	for (i = 0x04d4; i < 0x04dc; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x04e0/4, 0x00011100);
+		nv_wo32(ctx, i, 0x0fff0000);
+	nv_wo32(ctx, 0x04e0, 0x00011100);
 	for (i = 0x04fc; i < 0x053c; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x07ff0000);
-	nv_wo32(dev, ctx, 0x0544/4, 0x4b7fffff);
-	nv_wo32(dev, ctx, 0x057c/4, 0x00000080);
-	nv_wo32(dev, ctx, 0x0580/4, 0x30201000);
-	nv_wo32(dev, ctx, 0x0584/4, 0x70605040);
-	nv_wo32(dev, ctx, 0x0588/4, 0xb8a89888);
-	nv_wo32(dev, ctx, 0x058c/4, 0xf8e8d8c8);
-	nv_wo32(dev, ctx, 0x05a0/4, 0xb0000000);
+		nv_wo32(ctx, i, 0x07ff0000);
+	nv_wo32(ctx, 0x0544, 0x4b7fffff);
+	nv_wo32(ctx, 0x057c, 0x00000080);
+	nv_wo32(ctx, 0x0580, 0x30201000);
+	nv_wo32(ctx, 0x0584, 0x70605040);
+	nv_wo32(ctx, 0x0588, 0xb8a89888);
+	nv_wo32(ctx, 0x058c, 0xf8e8d8c8);
+	nv_wo32(ctx, 0x05a0, 0xb0000000);
 	for (i = 0x05f0; i < 0x0630; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00010588);
+		nv_wo32(ctx, i, 0x00010588);
 	for (i = 0x0630; i < 0x0670; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00030303);
+		nv_wo32(ctx, i, 0x00030303);
 	for (i = 0x06b0; i < 0x06f0; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0008aae4);
+		nv_wo32(ctx, i, 0x0008aae4);
 	for (i = 0x06f0; i < 0x0730; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x01012000);
+		nv_wo32(ctx, i, 0x01012000);
 	for (i = 0x0730; i < 0x0770; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080008);
-	nv_wo32(dev, ctx, 0x0850/4, 0x00040000);
-	nv_wo32(dev, ctx, 0x0854/4, 0x00010000);
+		nv_wo32(ctx, i, 0x00080008);
+	nv_wo32(ctx, 0x0850, 0x00040000);
+	nv_wo32(ctx, 0x0854, 0x00010000);
 	for (i = 0x0858; i < 0x0868; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00040004);
+		nv_wo32(ctx, i, 0x00040004);
 	for (i = 0x15ac; i <= 0x271c ; i += 16) {
-		nv_wo32(dev, ctx, i/4 + 0, 0x10700ff9);
-		nv_wo32(dev, ctx, i/4 + 1, 0x0436086c);
-		nv_wo32(dev, ctx, i/4 + 2, 0x000c001b);
+		nv_wo32(ctx, i + 0, 0x10700ff9);
+		nv_wo32(ctx, i + 1, 0x0436086c);
+		nv_wo32(ctx, i + 2, 0x000c001b);
 	}
 	for (i = 0x274c; i < 0x275c; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0000ffff);
-	nv_wo32(dev, ctx, 0x2ae0/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2e9c/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2eb0/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2edc/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x2ee0/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2ee4/4, 0x3f000000);
-	nv_wo32(dev, ctx, 0x2eec/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x2ef0/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x2ef8/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x2f00/4, 0xbf800000);
+		nv_wo32(ctx, i, 0x0000ffff);
+	nv_wo32(ctx, 0x2ae0, 0x3f800000);
+	nv_wo32(ctx, 0x2e9c, 0x3f800000);
+	nv_wo32(ctx, 0x2eb0, 0x3f800000);
+	nv_wo32(ctx, 0x2edc, 0x40000000);
+	nv_wo32(ctx, 0x2ee0, 0x3f800000);
+	nv_wo32(ctx, 0x2ee4, 0x3f000000);
+	nv_wo32(ctx, 0x2eec, 0x40000000);
+	nv_wo32(ctx, 0x2ef0, 0x3f800000);
+	nv_wo32(ctx, 0x2ef8, 0xbf800000);
+	nv_wo32(ctx, 0x2f00, 0xbf800000);
 }
 
 static void
@@ -312,57 +312,57 @@ nv35_36_graph_context_init(struct drm_device *dev, struct nouveau_gpuobj *ctx)
 {
 	int i;
 
-	nv_wo32(dev, ctx, 0x040c/4, 0x00000101);
-	nv_wo32(dev, ctx, 0x0420/4, 0x00000111);
-	nv_wo32(dev, ctx, 0x0424/4, 0x00000060);
-	nv_wo32(dev, ctx, 0x0440/4, 0x00000080);
-	nv_wo32(dev, ctx, 0x0444/4, 0xffff0000);
-	nv_wo32(dev, ctx, 0x0448/4, 0x00000001);
-	nv_wo32(dev, ctx, 0x045c/4, 0x44400000);
-	nv_wo32(dev, ctx, 0x0488/4, 0xffff0000);
+	nv_wo32(ctx, 0x040c, 0x00000101);
+	nv_wo32(ctx, 0x0420, 0x00000111);
+	nv_wo32(ctx, 0x0424, 0x00000060);
+	nv_wo32(ctx, 0x0440, 0x00000080);
+	nv_wo32(ctx, 0x0444, 0xffff0000);
+	nv_wo32(ctx, 0x0448, 0x00000001);
+	nv_wo32(ctx, 0x045c, 0x44400000);
+	nv_wo32(ctx, 0x0488, 0xffff0000);
 	for (i = 0x04dc; i < 0x04e4; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0fff0000);
-	nv_wo32(dev, ctx, 0x04e8/4, 0x00011100);
+		nv_wo32(ctx, i, 0x0fff0000);
+	nv_wo32(ctx, 0x04e8, 0x00011100);
 	for (i = 0x0504; i < 0x0544; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x07ff0000);
-	nv_wo32(dev, ctx, 0x054c/4, 0x4b7fffff);
-	nv_wo32(dev, ctx, 0x0588/4, 0x00000080);
-	nv_wo32(dev, ctx, 0x058c/4, 0x30201000);
-	nv_wo32(dev, ctx, 0x0590/4, 0x70605040);
-	nv_wo32(dev, ctx, 0x0594/4, 0xb8a89888);
-	nv_wo32(dev, ctx, 0x0598/4, 0xf8e8d8c8);
-	nv_wo32(dev, ctx, 0x05ac/4, 0xb0000000);
+		nv_wo32(ctx, i, 0x07ff0000);
+	nv_wo32(ctx, 0x054c, 0x4b7fffff);
+	nv_wo32(ctx, 0x0588, 0x00000080);
+	nv_wo32(ctx, 0x058c, 0x30201000);
+	nv_wo32(ctx, 0x0590, 0x70605040);
+	nv_wo32(ctx, 0x0594, 0xb8a89888);
+	nv_wo32(ctx, 0x0598, 0xf8e8d8c8);
+	nv_wo32(ctx, 0x05ac, 0xb0000000);
 	for (i = 0x0604; i < 0x0644; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00010588);
+		nv_wo32(ctx, i, 0x00010588);
 	for (i = 0x0644; i < 0x0684; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00030303);
+		nv_wo32(ctx, i, 0x00030303);
 	for (i = 0x06c4; i < 0x0704; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0008aae4);
+		nv_wo32(ctx, i, 0x0008aae4);
 	for (i = 0x0704; i < 0x0744; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x01012000);
+		nv_wo32(ctx, i, 0x01012000);
 	for (i = 0x0744; i < 0x0784; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00080008);
-	nv_wo32(dev, ctx, 0x0860/4, 0x00040000);
-	nv_wo32(dev, ctx, 0x0864/4, 0x00010000);
+		nv_wo32(ctx, i, 0x00080008);
+	nv_wo32(ctx, 0x0860, 0x00040000);
+	nv_wo32(ctx, 0x0864, 0x00010000);
 	for (i = 0x0868; i < 0x0878; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x00040004);
+		nv_wo32(ctx, i, 0x00040004);
 	for (i = 0x1f1c; i <= 0x308c ; i += 16) {
-		nv_wo32(dev, ctx, i/4 + 0, 0x10700ff9);
-		nv_wo32(dev, ctx, i/4 + 1, 0x0436086c);
-		nv_wo32(dev, ctx, i/4 + 2, 0x000c001b);
+		nv_wo32(ctx, i + 0, 0x10700ff9);
+		nv_wo32(ctx, i + 4, 0x0436086c);
+		nv_wo32(ctx, i + 8, 0x000c001b);
 	}
 	for (i = 0x30bc; i < 0x30cc; i += 4)
-		nv_wo32(dev, ctx, i/4, 0x0000ffff);
-	nv_wo32(dev, ctx, 0x3450/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x380c/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x3820/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x384c/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x3850/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x3854/4, 0x3f000000);
-	nv_wo32(dev, ctx, 0x385c/4, 0x40000000);
-	nv_wo32(dev, ctx, 0x3860/4, 0x3f800000);
-	nv_wo32(dev, ctx, 0x3868/4, 0xbf800000);
-	nv_wo32(dev, ctx, 0x3870/4, 0xbf800000);
+		nv_wo32(ctx, i, 0x0000ffff);
+	nv_wo32(ctx, 0x3450, 0x3f800000);
+	nv_wo32(ctx, 0x380c, 0x3f800000);
+	nv_wo32(ctx, 0x3820, 0x3f800000);
+	nv_wo32(ctx, 0x384c, 0x40000000);
+	nv_wo32(ctx, 0x3850, 0x3f800000);
+	nv_wo32(ctx, 0x3854, 0x3f000000);
+	nv_wo32(ctx, 0x385c, 0x40000000);
+	nv_wo32(ctx, 0x3860, 0x3f800000);
+	nv_wo32(ctx, 0x3868, 0xbf800000);
+	nv_wo32(ctx, 0x3870, 0xbf800000);
 }
 
 int
@@ -372,7 +372,7 @@ nv20_graph_create_context(struct nouveau_channel *chan)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pgraph_engine *pgraph = &dev_priv->engine.graph;
 	void (*ctx_init)(struct drm_device *, struct nouveau_gpuobj *);
-	unsigned int idoffs = 0x28/4;
+	unsigned int idoffs = 0x28;
 	int ret;
 
 	switch (dev_priv->chipset) {
@@ -403,21 +403,19 @@ nv20_graph_create_context(struct nouveau_channel *chan)
 		BUG_ON(1);
 	}
 
-	ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, pgraph->grctx_size,
-				     16, NVOBJ_FLAG_ZERO_ALLOC,
-				     &chan->ramin_grctx);
+	ret = nouveau_gpuobj_new(dev, chan, pgraph->grctx_size, 16,
+				 NVOBJ_FLAG_ZERO_ALLOC, &chan->ramin_grctx);
 	if (ret)
 		return ret;
 
 	/* Initialise default context values */
-	ctx_init(dev, chan->ramin_grctx->gpuobj);
+	ctx_init(dev, chan->ramin_grctx);
 
 	/* nv20: nv_wo32(dev, chan->ramin_grctx->gpuobj, 10, chan->id<<24); */
-	nv_wo32(dev, chan->ramin_grctx->gpuobj, idoffs,
-					(chan->id << 24) | 0x1); /* CTX_USER */
+	nv_wo32(chan->ramin_grctx, idoffs,
+		(chan->id << 24) | 0x1); /* CTX_USER */
 
-	nv_wo32(dev, pgraph->ctx_table->gpuobj, chan->id,
-		     chan->ramin_grctx->instance >> 4);
+	nv_wo32(pgraph->ctx_table, chan->id * 4, chan->ramin_grctx->pinst >> 4);
 	return 0;
 }
 
@@ -428,10 +426,8 @@ nv20_graph_destroy_context(struct nouveau_channel *chan)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pgraph_engine *pgraph = &dev_priv->engine.graph;
 
-	if (chan->ramin_grctx)
-		nouveau_gpuobj_ref_del(dev, &chan->ramin_grctx);
-
-	nv_wo32(dev, pgraph->ctx_table->gpuobj, chan->id, 0);
+	nouveau_gpuobj_ref(NULL, &chan->ramin_grctx);
+	nv_wo32(pgraph->ctx_table, chan->id * 4, 0);
 }
 
 int
@@ -442,7 +438,7 @@ nv20_graph_load_context(struct nouveau_channel *chan)
 
 	if (!chan->ramin_grctx)
 		return -EINVAL;
-	inst = chan->ramin_grctx->instance >> 4;
+	inst = chan->ramin_grctx->pinst >> 4;
 
 	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_POINTER, inst);
 	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_XFER,
@@ -465,7 +461,7 @@ nv20_graph_unload_context(struct drm_device *dev)
 	chan = pgraph->channel(dev);
 	if (!chan)
 		return 0;
-	inst = chan->ramin_grctx->instance >> 4;
+	inst = chan->ramin_grctx->pinst >> 4;
 
 	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_POINTER, inst);
 	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_XFER,
@@ -552,15 +548,15 @@ nv20_graph_init(struct drm_device *dev)
 
 	if (!pgraph->ctx_table) {
 		/* Create Context Pointer Table */
-		ret = nouveau_gpuobj_new_ref(dev, NULL, NULL, 0, 32 * 4, 16,
-						  NVOBJ_FLAG_ZERO_ALLOC,
-						  &pgraph->ctx_table);
+		ret = nouveau_gpuobj_new(dev, NULL, 32 * 4, 16,
+					 NVOBJ_FLAG_ZERO_ALLOC,
+					 &pgraph->ctx_table);
 		if (ret)
 			return ret;
 	}
 
 	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_TABLE,
-		     pgraph->ctx_table->instance >> 4);
+		     pgraph->ctx_table->pinst >> 4);
 
 	nv20_graph_rdi(dev);
 
@@ -646,7 +642,7 @@ nv20_graph_takedown(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pgraph_engine *pgraph = &dev_priv->engine.graph;
 
-	nouveau_gpuobj_ref_del(dev, &pgraph->ctx_table);
+	nouveau_gpuobj_ref(NULL, &pgraph->ctx_table);
 }
 
 int
@@ -681,15 +677,15 @@ nv30_graph_init(struct drm_device *dev)
 
 	if (!pgraph->ctx_table) {
 		/* Create Context Pointer Table */
-		ret = nouveau_gpuobj_new_ref(dev, NULL, NULL, 0, 32 * 4, 16,
-						  NVOBJ_FLAG_ZERO_ALLOC,
-						  &pgraph->ctx_table);
+		ret = nouveau_gpuobj_new(dev, NULL, 32 * 4, 16,
+					 NVOBJ_FLAG_ZERO_ALLOC,
+					 &pgraph->ctx_table);
 		if (ret)
 			return ret;
 	}
 
 	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_TABLE,
-		     pgraph->ctx_table->instance >> 4);
+		     pgraph->ctx_table->pinst >> 4);
 
 	nv_wr32(dev, NV03_PGRAPH_INTR   , 0xFFFFFFFF);
 	nv_wr32(dev, NV03_PGRAPH_INTR_EN, 0xFFFFFFFF);
diff --git a/drivers/gpu/drm/nouveau/nv40_fifo.c b/drivers/gpu/drm/nouveau/nv40_fifo.c
index 2b67f1835c39..d337b8b28cdd 100644
--- a/drivers/gpu/drm/nouveau/nv40_fifo.c
+++ b/drivers/gpu/drm/nouveau/nv40_fifo.c
@@ -27,8 +27,9 @@
 #include "drmP.h"
 #include "nouveau_drv.h"
 #include "nouveau_drm.h"
+#include "nouveau_ramht.h"
 
-#define NV40_RAMFC(c) (dev_priv->ramfc_offset + ((c) * NV40_RAMFC__SIZE))
+#define NV40_RAMFC(c) (dev_priv->ramfc->pinst + ((c) * NV40_RAMFC__SIZE))
 #define NV40_RAMFC__SIZE 128
 
 int
@@ -42,7 +43,7 @@ nv40_fifo_create_context(struct nouveau_channel *chan)
 
 	ret = nouveau_gpuobj_new_fake(dev, NV40_RAMFC(chan->id), ~0,
 				      NV40_RAMFC__SIZE, NVOBJ_FLAG_ZERO_ALLOC |
-				      NVOBJ_FLAG_ZERO_FREE, NULL, &chan->ramfc);
+				      NVOBJ_FLAG_ZERO_FREE, &chan->ramfc);
 	if (ret)
 		return ret;
 
@@ -50,7 +51,7 @@ nv40_fifo_create_context(struct nouveau_channel *chan)
 
 	nv_wi32(dev, fc +  0, chan->pushbuf_base);
 	nv_wi32(dev, fc +  4, chan->pushbuf_base);
-	nv_wi32(dev, fc + 12, chan->pushbuf->instance >> 4);
+	nv_wi32(dev, fc + 12, chan->pushbuf->pinst >> 4);
 	nv_wi32(dev, fc + 24, NV_PFIFO_CACHE1_DMA_FETCH_TRIG_128_BYTES |
 			      NV_PFIFO_CACHE1_DMA_FETCH_SIZE_128_BYTES |
 			      NV_PFIFO_CACHE1_DMA_FETCH_MAX_REQS_8 |
@@ -58,7 +59,7 @@ nv40_fifo_create_context(struct nouveau_channel *chan)
 			      NV_PFIFO_CACHE1_BIG_ENDIAN |
 #endif
 			      0x30000000 /* no idea.. */);
-	nv_wi32(dev, fc + 56, chan->ramin_grctx->instance >> 4);
+	nv_wi32(dev, fc + 56, chan->ramin_grctx->pinst >> 4);
 	nv_wi32(dev, fc + 60, 0x0001FFFF);
 
 	/* enable the fifo dma operation */
@@ -77,8 +78,7 @@ nv40_fifo_destroy_context(struct nouveau_channel *chan)
 	nv_wr32(dev, NV04_PFIFO_MODE,
 		nv_rd32(dev, NV04_PFIFO_MODE) & ~(1 << chan->id));
 
-	if (chan->ramfc)
-		nouveau_gpuobj_ref_del(dev, &chan->ramfc);
+	nouveau_gpuobj_ref(NULL, &chan->ramfc);
 }
 
 static void
@@ -241,9 +241,9 @@ nv40_fifo_init_ramxx(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 
 	nv_wr32(dev, NV03_PFIFO_RAMHT, (0x03 << 24) /* search 128 */ |
-				       ((dev_priv->ramht_bits - 9) << 16) |
-				       (dev_priv->ramht_offset >> 8));
-	nv_wr32(dev, NV03_PFIFO_RAMRO, dev_priv->ramro_offset>>8);
+				       ((dev_priv->ramht->bits - 9) << 16) |
+				       (dev_priv->ramht->gpuobj->pinst >> 8));
+	nv_wr32(dev, NV03_PFIFO_RAMRO, dev_priv->ramro->pinst >> 8);
 
 	switch (dev_priv->chipset) {
 	case 0x47:
@@ -271,7 +271,7 @@ nv40_fifo_init_ramxx(struct drm_device *dev)
 		nv_wr32(dev, 0x2230, 0);
 		nv_wr32(dev, NV40_PFIFO_RAMFC,
 			((dev_priv->vram_size - 512 * 1024 +
-			  dev_priv->ramfc_offset) >> 16) | (3 << 16));
+			  dev_priv->ramfc->pinst) >> 16) | (3 << 16));
 		break;
 	}
 }
diff --git a/drivers/gpu/drm/nouveau/nv40_graph.c b/drivers/gpu/drm/nouveau/nv40_graph.c
index fd7d2b501316..7ee1b91569b8 100644
--- a/drivers/gpu/drm/nouveau/nv40_graph.c
+++ b/drivers/gpu/drm/nouveau/nv40_graph.c
@@ -45,7 +45,7 @@ nv40_graph_channel(struct drm_device *dev)
 		struct nouveau_channel *chan = dev_priv->fifos[i];
 
 		if (chan && chan->ramin_grctx &&
-		    chan->ramin_grctx->instance == inst)
+		    chan->ramin_grctx->pinst == inst)
 			return chan;
 	}
 
@@ -61,27 +61,25 @@ nv40_graph_create_context(struct nouveau_channel *chan)
 	struct nouveau_grctx ctx = {};
 	int ret;
 
-	ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, pgraph->grctx_size,
-				     16, NVOBJ_FLAG_ZERO_ALLOC,
-				     &chan->ramin_grctx);
+	ret = nouveau_gpuobj_new(dev, chan, pgraph->grctx_size, 16,
+				 NVOBJ_FLAG_ZERO_ALLOC, &chan->ramin_grctx);
 	if (ret)
 		return ret;
 
 	/* Initialise default context values */
 	ctx.dev = chan->dev;
 	ctx.mode = NOUVEAU_GRCTX_VALS;
-	ctx.data = chan->ramin_grctx->gpuobj;
+	ctx.data = chan->ramin_grctx;
 	nv40_grctx_init(&ctx);
 
-	nv_wo32(dev, chan->ramin_grctx->gpuobj, 0,
-		     chan->ramin_grctx->gpuobj->im_pramin->start);
+	nv_wo32(chan->ramin_grctx, 0, chan->ramin_grctx->pinst);
 	return 0;
 }
 
 void
 nv40_graph_destroy_context(struct nouveau_channel *chan)
 {
-	nouveau_gpuobj_ref_del(chan->dev, &chan->ramin_grctx);
+	nouveau_gpuobj_ref(NULL, &chan->ramin_grctx);
 }
 
 static int
@@ -135,7 +133,7 @@ nv40_graph_load_context(struct nouveau_channel *chan)
 
 	if (!chan->ramin_grctx)
 		return -EINVAL;
-	inst = chan->ramin_grctx->instance >> 4;
+	inst = chan->ramin_grctx->pinst >> 4;
 
 	ret = nv40_graph_transfer_context(dev, inst, 0);
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/nv40_grctx.c b/drivers/gpu/drm/nouveau/nv40_grctx.c
index 9b5c97469588..ce585093264e 100644
--- a/drivers/gpu/drm/nouveau/nv40_grctx.c
+++ b/drivers/gpu/drm/nouveau/nv40_grctx.c
@@ -596,13 +596,13 @@ nv40_graph_construct_shader(struct nouveau_grctx *ctx)
 
 	offset += 0x0280/4;
 	for (i = 0; i < 16; i++, offset += 2)
-		nv_wo32(dev, obj, offset, 0x3f800000);
+		nv_wo32(obj, offset * 4, 0x3f800000);
 
 	for (vs = 0; vs < vs_nr; vs++, offset += vs_len) {
 		for (i = 0; i < vs_nr_b0 * 6; i += 6)
-			nv_wo32(dev, obj, offset + b0_offset + i, 0x00000001);
+			nv_wo32(obj, (offset + b0_offset + i) * 4, 0x00000001);
 		for (i = 0; i < vs_nr_b1 * 4; i += 4)
-			nv_wo32(dev, obj, offset + b1_offset + i, 0x3f800000);
+			nv_wo32(obj, (offset + b1_offset + i) * 4, 0x3f800000);
 	}
 }
 
diff --git a/drivers/gpu/drm/nouveau/nv50_crtc.c b/drivers/gpu/drm/nouveau/nv50_crtc.c
index bfd4ca2fe7ef..3f2fb4ec63ab 100644
--- a/drivers/gpu/drm/nouveau/nv50_crtc.c
+++ b/drivers/gpu/drm/nouveau/nv50_crtc.c
@@ -104,8 +104,7 @@ nv50_crtc_blank(struct nouveau_crtc *nv_crtc, bool blanked)
 		OUT_RING(evo, nv_crtc->lut.depth == 8 ?
 				NV50_EVO_CRTC_CLUT_MODE_OFF :
 				NV50_EVO_CRTC_CLUT_MODE_ON);
-		OUT_RING(evo, (nv_crtc->lut.nvbo->bo.mem.mm_node->start <<
-				 PAGE_SHIFT) >> 8);
+		OUT_RING(evo, (nv_crtc->lut.nvbo->bo.mem.start << PAGE_SHIFT) >> 8);
 		if (dev_priv->chipset != 0x50) {
 			BEGIN_RING(evo, 0, NV84_EVO_CRTC(index, CLUT_DMA), 1);
 			OUT_RING(evo, NvEvoVRAM);
@@ -266,15 +265,10 @@ nv50_crtc_set_clock(struct drm_device *dev, int head, int pclk)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct pll_lims pll;
-	uint32_t reg, reg1, reg2;
+	uint32_t reg1, reg2;
 	int ret, N1, M1, N2, M2, P;
 
-	if (dev_priv->chipset < NV_C0)
-		reg = NV50_PDISPLAY_CRTC_CLK_CTRL1(head);
-	else
-		reg = 0x614140 + (head * 0x800);
-
-	ret = get_pll_limits(dev, reg, &pll);
+	ret = get_pll_limits(dev, PLL_VPLL0 + head, &pll);
 	if (ret)
 		return ret;
 
@@ -286,11 +280,11 @@ nv50_crtc_set_clock(struct drm_device *dev, int head, int pclk)
 		NV_DEBUG(dev, "pclk %d out %d NM1 %d %d NM2 %d %d P %d\n",
 			 pclk, ret, N1, M1, N2, M2, P);
 
-		reg1 = nv_rd32(dev, reg + 4) & 0xff00ff00;
-		reg2 = nv_rd32(dev, reg + 8) & 0x8000ff00;
-		nv_wr32(dev, reg, 0x10000611);
-		nv_wr32(dev, reg + 4, reg1 | (M1 << 16) | N1);
-		nv_wr32(dev, reg + 8, reg2 | (P << 28) | (M2 << 16) | N2);
+		reg1 = nv_rd32(dev, pll.reg + 4) & 0xff00ff00;
+		reg2 = nv_rd32(dev, pll.reg + 8) & 0x8000ff00;
+		nv_wr32(dev, pll.reg + 0, 0x10000611);
+		nv_wr32(dev, pll.reg + 4, reg1 | (M1 << 16) | N1);
+		nv_wr32(dev, pll.reg + 8, reg2 | (P << 28) | (M2 << 16) | N2);
 	} else
 	if (dev_priv->chipset < NV_C0) {
 		ret = nv50_calc_pll2(dev, &pll, pclk, &N1, &N2, &M1, &P);
@@ -300,10 +294,10 @@ nv50_crtc_set_clock(struct drm_device *dev, int head, int pclk)
 		NV_DEBUG(dev, "pclk %d out %d N %d fN 0x%04x M %d P %d\n",
 			 pclk, ret, N1, N2, M1, P);
 
-		reg1 = nv_rd32(dev, reg + 4) & 0xffc00000;
-		nv_wr32(dev, reg, 0x50000610);
-		nv_wr32(dev, reg + 4, reg1 | (P << 16) | (M1 << 8) | N1);
-		nv_wr32(dev, reg + 8, N2);
+		reg1 = nv_rd32(dev, pll.reg + 4) & 0xffc00000;
+		nv_wr32(dev, pll.reg + 0, 0x50000610);
+		nv_wr32(dev, pll.reg + 4, reg1 | (P << 16) | (M1 << 8) | N1);
+		nv_wr32(dev, pll.reg + 8, N2);
 	} else {
 		ret = nv50_calc_pll2(dev, &pll, pclk, &N1, &N2, &M1, &P);
 		if (ret <= 0)
@@ -312,9 +306,9 @@ nv50_crtc_set_clock(struct drm_device *dev, int head, int pclk)
 		NV_DEBUG(dev, "pclk %d out %d N %d fN 0x%04x M %d P %d\n",
 			 pclk, ret, N1, N2, M1, P);
 
-		nv_mask(dev, reg + 0x0c, 0x00000000, 0x00000100);
-		nv_wr32(dev, reg + 0x04, (P << 16) | (N1 << 8) | M1);
-		nv_wr32(dev, reg + 0x10, N2 << 16);
+		nv_mask(dev, pll.reg + 0x0c, 0x00000000, 0x00000100);
+		nv_wr32(dev, pll.reg + 0x04, (P << 16) | (N1 << 8) | M1);
+		nv_wr32(dev, pll.reg + 0x10, N2 << 16);
 	}
 
 	return 0;
@@ -338,7 +332,9 @@ nv50_crtc_destroy(struct drm_crtc *crtc)
 
 	nv50_cursor_fini(nv_crtc);
 
+	nouveau_bo_unmap(nv_crtc->lut.nvbo);
 	nouveau_bo_ref(NULL, &nv_crtc->lut.nvbo);
+	nouveau_bo_unmap(nv_crtc->cursor.nvbo);
 	nouveau_bo_ref(NULL, &nv_crtc->cursor.nvbo);
 	kfree(nv_crtc->mode);
 	kfree(nv_crtc);
diff --git a/drivers/gpu/drm/nouveau/nv50_cursor.c b/drivers/gpu/drm/nouveau/nv50_cursor.c
index 03ad7ab14f09..1b9ce3021aa3 100644
--- a/drivers/gpu/drm/nouveau/nv50_cursor.c
+++ b/drivers/gpu/drm/nouveau/nv50_cursor.c
@@ -147,7 +147,7 @@ nv50_cursor_fini(struct nouveau_crtc *nv_crtc)
 	NV_DEBUG_KMS(dev, "\n");
 
 	nv_wr32(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(idx), 0);
-	if (!nv_wait(NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(idx),
+	if (!nv_wait(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(idx),
 		     NV50_PDISPLAY_CURSOR_CURSOR_CTRL2_STATUS, 0)) {
 		NV_ERROR(dev, "timeout: CURSOR_CTRL2_STATUS == 0\n");
 		NV_ERROR(dev, "CURSOR_CTRL2 = 0x%08x\n",
diff --git a/drivers/gpu/drm/nouveau/nv50_dac.c b/drivers/gpu/drm/nouveau/nv50_dac.c
index 1bc085962945..875414b09ade 100644
--- a/drivers/gpu/drm/nouveau/nv50_dac.c
+++ b/drivers/gpu/drm/nouveau/nv50_dac.c
@@ -79,7 +79,7 @@ nv50_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector)
 
 	nv_wr32(dev, NV50_PDISPLAY_DAC_DPMS_CTRL(or),
 		0x00150000 | NV50_PDISPLAY_DAC_DPMS_CTRL_PENDING);
-	if (!nv_wait(NV50_PDISPLAY_DAC_DPMS_CTRL(or),
+	if (!nv_wait(dev, NV50_PDISPLAY_DAC_DPMS_CTRL(or),
 		     NV50_PDISPLAY_DAC_DPMS_CTRL_PENDING, 0)) {
 		NV_ERROR(dev, "timeout: DAC_DPMS_CTRL_PENDING(%d) == 0\n", or);
 		NV_ERROR(dev, "DAC_DPMS_CTRL(%d) = 0x%08x\n", or,
@@ -130,7 +130,7 @@ nv50_dac_dpms(struct drm_encoder *encoder, int mode)
 	NV_DEBUG_KMS(dev, "or %d mode %d\n", or, mode);
 
 	/* wait for it to be done */
-	if (!nv_wait(NV50_PDISPLAY_DAC_DPMS_CTRL(or),
+	if (!nv_wait(dev, NV50_PDISPLAY_DAC_DPMS_CTRL(or),
 		     NV50_PDISPLAY_DAC_DPMS_CTRL_PENDING, 0)) {
 		NV_ERROR(dev, "timeout: DAC_DPMS_CTRL_PENDING(%d) == 0\n", or);
 		NV_ERROR(dev, "DAC_DPMS_CTRL(%d) = 0x%08x\n", or,
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 612fa6d6a0cb..55c9663ef2bf 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -30,8 +30,22 @@
 #include "nouveau_connector.h"
 #include "nouveau_fb.h"
 #include "nouveau_fbcon.h"
+#include "nouveau_ramht.h"
 #include "drm_crtc_helper.h"
 
+static inline int
+nv50_sor_nr(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->chipset  < 0x90 ||
+	    dev_priv->chipset == 0x92 ||
+	    dev_priv->chipset == 0xa0)
+		return 2;
+
+	return 4;
+}
+
 static void
 nv50_evo_channel_del(struct nouveau_channel **pchan)
 {
@@ -42,6 +56,7 @@ nv50_evo_channel_del(struct nouveau_channel **pchan)
 	*pchan = NULL;
 
 	nouveau_gpuobj_channel_takedown(chan);
+	nouveau_bo_unmap(chan->pushbuf_bo);
 	nouveau_bo_ref(NULL, &chan->pushbuf_bo);
 
 	if (chan->user)
@@ -65,23 +80,23 @@ nv50_evo_dmaobj_new(struct nouveau_channel *evo, uint32_t class, uint32_t name,
 		return ret;
 	obj->engine = NVOBJ_ENGINE_DISPLAY;
 
-	ret = nouveau_gpuobj_ref_add(dev, evo, name, obj, NULL);
+	nv_wo32(obj,  0, (tile_flags << 22) | (magic_flags << 16) | class);
+	nv_wo32(obj,  4, limit);
+	nv_wo32(obj,  8, offset);
+	nv_wo32(obj, 12, 0x00000000);
+	nv_wo32(obj, 16, 0x00000000);
+	if (dev_priv->card_type < NV_C0)
+		nv_wo32(obj, 20, 0x00010000);
+	else
+		nv_wo32(obj, 20, 0x00020000);
+	dev_priv->engine.instmem.flush(dev);
+
+	ret = nouveau_ramht_insert(evo, name, obj);
+	nouveau_gpuobj_ref(NULL, &obj);
 	if (ret) {
-		nouveau_gpuobj_del(dev, &obj);
 		return ret;
 	}
 
-	nv_wo32(dev, obj, 0, (tile_flags << 22) | (magic_flags << 16) | class);
-	nv_wo32(dev, obj, 1, limit);
-	nv_wo32(dev, obj, 2, offset);
-	nv_wo32(dev, obj, 3, 0x00000000);
-	nv_wo32(dev, obj, 4, 0x00000000);
-	if (dev_priv->card_type < NV_C0)
-		nv_wo32(dev, obj, 5, 0x00010000);
-	else
-		nv_wo32(dev, obj, 5, 0x00020000);
-	dev_priv->engine.instmem.flush(dev);
-
 	return 0;
 }
 
@@ -89,6 +104,7 @@ static int
 nv50_evo_channel_new(struct drm_device *dev, struct nouveau_channel **pchan)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpuobj *ramht = NULL;
 	struct nouveau_channel *chan;
 	int ret;
 
@@ -102,32 +118,35 @@ nv50_evo_channel_new(struct drm_device *dev, struct nouveau_channel **pchan)
 	chan->user_get = 4;
 	chan->user_put = 0;
 
-	INIT_LIST_HEAD(&chan->ramht_refs);
-
-	ret = nouveau_gpuobj_new_ref(dev, NULL, NULL, 0, 32768, 0x1000,
-				     NVOBJ_FLAG_ZERO_ALLOC, &chan->ramin);
+	ret = nouveau_gpuobj_new(dev, NULL, 32768, 0x1000,
+				 NVOBJ_FLAG_ZERO_ALLOC, &chan->ramin);
 	if (ret) {
 		NV_ERROR(dev, "Error allocating EVO channel memory: %d\n", ret);
 		nv50_evo_channel_del(pchan);
 		return ret;
 	}
 
-	ret = drm_mm_init(&chan->ramin_heap,
-			  chan->ramin->gpuobj->im_pramin->start, 32768);
+	ret = drm_mm_init(&chan->ramin_heap, 0, 32768);
 	if (ret) {
 		NV_ERROR(dev, "Error initialising EVO PRAMIN heap: %d\n", ret);
 		nv50_evo_channel_del(pchan);
 		return ret;
 	}
 
-	ret = nouveau_gpuobj_new_ref(dev, chan, chan, 0, 4096, 16,
-				     0, &chan->ramht);
+	ret = nouveau_gpuobj_new(dev, chan, 4096, 16, 0, &ramht);
 	if (ret) {
 		NV_ERROR(dev, "Unable to allocate EVO RAMHT: %d\n", ret);
 		nv50_evo_channel_del(pchan);
 		return ret;
 	}
 
+	ret = nouveau_ramht_new(dev, ramht, &chan->ramht);
+	nouveau_gpuobj_ref(NULL, &ramht);
+	if (ret) {
+		nv50_evo_channel_del(pchan);
+		return ret;
+	}
+
 	if (dev_priv->chipset != 0x50) {
 		ret = nv50_evo_dmaobj_new(chan, 0x3d, NvEvoFB16, 0x70, 0x19,
 					  0, 0xffffffff);
@@ -227,11 +246,11 @@ nv50_display_init(struct drm_device *dev)
 		nv_wr32(dev, 0x006101d0 + (i * 0x04), val);
 	}
 	/* SOR */
-	for (i = 0; i < 4; i++) {
+	for (i = 0; i < nv50_sor_nr(dev); i++) {
 		val = nv_rd32(dev, 0x0061c000 + (i * 0x800));
 		nv_wr32(dev, 0x006101e0 + (i * 0x04), val);
 	}
-	/* Something not yet in use, tv-out maybe. */
+	/* EXT */
 	for (i = 0; i < 3; i++) {
 		val = nv_rd32(dev, 0x0061e000 + (i * 0x800));
 		nv_wr32(dev, 0x006101f0 + (i * 0x04), val);
@@ -260,7 +279,7 @@ nv50_display_init(struct drm_device *dev)
 	if (nv_rd32(dev, NV50_PDISPLAY_INTR_1) & 0x100) {
 		nv_wr32(dev, NV50_PDISPLAY_INTR_1, 0x100);
 		nv_wr32(dev, 0x006194e8, nv_rd32(dev, 0x006194e8) & ~1);
-		if (!nv_wait(0x006194e8, 2, 0)) {
+		if (!nv_wait(dev, 0x006194e8, 2, 0)) {
 			NV_ERROR(dev, "timeout: (0x6194e8 & 2) != 0\n");
 			NV_ERROR(dev, "0x6194e8 = 0x%08x\n",
 						nv_rd32(dev, 0x6194e8));
@@ -291,7 +310,8 @@ nv50_display_init(struct drm_device *dev)
 
 	nv_wr32(dev, NV50_PDISPLAY_CTRL_STATE, NV50_PDISPLAY_CTRL_STATE_ENABLE);
 	nv_wr32(dev, NV50_PDISPLAY_CHANNEL_STAT(0), 0x1000b03);
-	if (!nv_wait(NV50_PDISPLAY_CHANNEL_STAT(0), 0x40000000, 0x40000000)) {
+	if (!nv_wait(dev, NV50_PDISPLAY_CHANNEL_STAT(0),
+		     0x40000000, 0x40000000)) {
 		NV_ERROR(dev, "timeout: (0x610200 & 0x40000000) == 0x40000000\n");
 		NV_ERROR(dev, "0x610200 = 0x%08x\n",
 			  nv_rd32(dev, NV50_PDISPLAY_CHANNEL_STAT(0)));
@@ -300,7 +320,7 @@ nv50_display_init(struct drm_device *dev)
 
 	for (i = 0; i < 2; i++) {
 		nv_wr32(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i), 0x2000);
-		if (!nv_wait(NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i),
+		if (!nv_wait(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i),
 			     NV50_PDISPLAY_CURSOR_CURSOR_CTRL2_STATUS, 0)) {
 			NV_ERROR(dev, "timeout: CURSOR_CTRL2_STATUS == 0\n");
 			NV_ERROR(dev, "CURSOR_CTRL2 = 0x%08x\n",
@@ -310,7 +330,7 @@ nv50_display_init(struct drm_device *dev)
 
 		nv_wr32(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i),
 			NV50_PDISPLAY_CURSOR_CURSOR_CTRL2_ON);
-		if (!nv_wait(NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i),
+		if (!nv_wait(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i),
 			     NV50_PDISPLAY_CURSOR_CURSOR_CTRL2_STATUS,
 			     NV50_PDISPLAY_CURSOR_CURSOR_CTRL2_STATUS_ACTIVE)) {
 			NV_ERROR(dev, "timeout: "
@@ -321,16 +341,16 @@ nv50_display_init(struct drm_device *dev)
 		}
 	}
 
-	nv_wr32(dev, NV50_PDISPLAY_OBJECTS, (evo->ramin->instance >> 8) | 9);
+	nv_wr32(dev, NV50_PDISPLAY_OBJECTS, (evo->ramin->vinst >> 8) | 9);
 
 	/* initialise fifo */
 	nv_wr32(dev, NV50_PDISPLAY_CHANNEL_DMA_CB(0),
-		((evo->pushbuf_bo->bo.mem.mm_node->start << PAGE_SHIFT) >> 8) |
+		((evo->pushbuf_bo->bo.mem.start << PAGE_SHIFT) >> 8) |
 		NV50_PDISPLAY_CHANNEL_DMA_CB_LOCATION_VRAM |
 		NV50_PDISPLAY_CHANNEL_DMA_CB_VALID);
 	nv_wr32(dev, NV50_PDISPLAY_CHANNEL_UNK2(0), 0x00010000);
 	nv_wr32(dev, NV50_PDISPLAY_CHANNEL_UNK3(0), 0x00000002);
-	if (!nv_wait(0x610200, 0x80000000, 0x00000000)) {
+	if (!nv_wait(dev, 0x610200, 0x80000000, 0x00000000)) {
 		NV_ERROR(dev, "timeout: (0x610200 & 0x80000000) == 0\n");
 		NV_ERROR(dev, "0x610200 = 0x%08x\n", nv_rd32(dev, 0x610200));
 		return -EBUSY;
@@ -370,7 +390,7 @@ nv50_display_init(struct drm_device *dev)
 	BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, UNK082C), 1);
 	OUT_RING(evo, 0);
 	FIRE_RING(evo);
-	if (!nv_wait(0x640004, 0xffffffff, evo->dma.put << 2))
+	if (!nv_wait(dev, 0x640004, 0xffffffff, evo->dma.put << 2))
 		NV_ERROR(dev, "evo pushbuf stalled\n");
 
 	/* enable clock change interrupts. */
@@ -424,7 +444,7 @@ static int nv50_display_disable(struct drm_device *dev)
 			continue;
 
 		nv_wr32(dev, NV50_PDISPLAY_INTR_1, mask);
-		if (!nv_wait(NV50_PDISPLAY_INTR_1, mask, mask)) {
+		if (!nv_wait(dev, NV50_PDISPLAY_INTR_1, mask, mask)) {
 			NV_ERROR(dev, "timeout: (0x610024 & 0x%08x) == "
 				      "0x%08x\n", mask, mask);
 			NV_ERROR(dev, "0x610024 = 0x%08x\n",
@@ -434,14 +454,14 @@ static int nv50_display_disable(struct drm_device *dev)
 
 	nv_wr32(dev, NV50_PDISPLAY_CHANNEL_STAT(0), 0);
 	nv_wr32(dev, NV50_PDISPLAY_CTRL_STATE, 0);
-	if (!nv_wait(NV50_PDISPLAY_CHANNEL_STAT(0), 0x1e0000, 0)) {
+	if (!nv_wait(dev, NV50_PDISPLAY_CHANNEL_STAT(0), 0x1e0000, 0)) {
 		NV_ERROR(dev, "timeout: (0x610200 & 0x1e0000) == 0\n");
 		NV_ERROR(dev, "0x610200 = 0x%08x\n",
 			  nv_rd32(dev, NV50_PDISPLAY_CHANNEL_STAT(0)));
 	}
 
 	for (i = 0; i < 3; i++) {
-		if (!nv_wait(NV50_PDISPLAY_SOR_DPMS_STATE(i),
+		if (!nv_wait(dev, NV50_PDISPLAY_SOR_DPMS_STATE(i),
 			     NV50_PDISPLAY_SOR_DPMS_STATE_WAIT, 0)) {
 			NV_ERROR(dev, "timeout: SOR_DPMS_STATE_WAIT(%d) == 0\n", i);
 			NV_ERROR(dev, "SOR_DPMS_STATE(%d) = 0x%08x\n", i,
@@ -710,7 +730,7 @@ nv50_display_unk10_handler(struct drm_device *dev)
 		or = i;
 	}
 
-	for (i = 0; type == OUTPUT_ANY && i < 4; i++) {
+	for (i = 0; type == OUTPUT_ANY && i < nv50_sor_nr(dev); i++) {
 		if (dev_priv->chipset  < 0x90 ||
 		    dev_priv->chipset == 0x92 ||
 		    dev_priv->chipset == 0xa0)
@@ -841,7 +861,7 @@ nv50_display_unk20_handler(struct drm_device *dev)
 		or = i;
 	}
 
-	for (i = 0; type == OUTPUT_ANY && i < 4; i++) {
+	for (i = 0; type == OUTPUT_ANY && i < nv50_sor_nr(dev); i++) {
 		if (dev_priv->chipset  < 0x90 ||
 		    dev_priv->chipset == 0x92 ||
 		    dev_priv->chipset == 0xa0)
diff --git a/drivers/gpu/drm/nouveau/nv50_fb.c b/drivers/gpu/drm/nouveau/nv50_fb.c
index 32611bd30e6d..cd1988b15d2c 100644
--- a/drivers/gpu/drm/nouveau/nv50_fb.c
+++ b/drivers/gpu/drm/nouveau/nv50_fb.c
@@ -20,6 +20,7 @@ nv50_fb_init(struct drm_device *dev)
 	case 0x50:
 		nv_wr32(dev, 0x100c90, 0x0707ff);
 		break;
+	case 0xa3:
 	case 0xa5:
 	case 0xa8:
 		nv_wr32(dev, 0x100c90, 0x0d0fff);
@@ -36,3 +37,42 @@ void
 nv50_fb_takedown(struct drm_device *dev)
 {
 }
+
+void
+nv50_fb_vm_trap(struct drm_device *dev, int display, const char *name)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 trap[6], idx, chinst;
+	int i, ch;
+
+	idx = nv_rd32(dev, 0x100c90);
+	if (!(idx & 0x80000000))
+		return;
+	idx &= 0x00ffffff;
+
+	for (i = 0; i < 6; i++) {
+		nv_wr32(dev, 0x100c90, idx | i << 24);
+		trap[i] = nv_rd32(dev, 0x100c94);
+	}
+	nv_wr32(dev, 0x100c90, idx | 0x80000000);
+
+	if (!display)
+		return;
+
+	chinst = (trap[2] << 16) | trap[1];
+	for (ch = 0; ch < dev_priv->engine.fifo.channels; ch++) {
+		struct nouveau_channel *chan = dev_priv->fifos[ch];
+
+		if (!chan || !chan->ramin)
+			continue;
+
+		if (chinst == chan->ramin->vinst >> 12)
+			break;
+	}
+
+	NV_INFO(dev, "%s - VM: Trapped %s at %02x%04x%04x status %08x "
+		     "channel %d (0x%08x)\n",
+		name, (trap[5] & 0x100 ? "read" : "write"),
+		trap[5] & 0xff, trap[4] & 0xffff, trap[3] & 0xffff,
+		trap[0], ch, chinst);
+}
diff --git a/drivers/gpu/drm/nouveau/nv50_fbcon.c b/drivers/gpu/drm/nouveau/nv50_fbcon.c
index 6bf025c6fc6f..6dcf048eddbc 100644
--- a/drivers/gpu/drm/nouveau/nv50_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nv50_fbcon.c
@@ -1,6 +1,7 @@
 #include "drmP.h"
 #include "nouveau_drv.h"
 #include "nouveau_dma.h"
+#include "nouveau_ramht.h"
 #include "nouveau_fbcon.h"
 
 void
@@ -193,7 +194,8 @@ nv50_fbcon_accel_init(struct fb_info *info)
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_ref_add(dev, dev_priv->channel, Nv2D, eng2d, NULL);
+	ret = nouveau_ramht_insert(dev_priv->channel, Nv2D, eng2d);
+	nouveau_gpuobj_ref(NULL, &eng2d);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nv50_fifo.c b/drivers/gpu/drm/nouveau/nv50_fifo.c
index fb0281ae8f90..a46a961102f3 100644
--- a/drivers/gpu/drm/nouveau/nv50_fifo.c
+++ b/drivers/gpu/drm/nouveau/nv50_fifo.c
@@ -27,13 +27,14 @@
 #include "drmP.h"
 #include "drm.h"
 #include "nouveau_drv.h"
+#include "nouveau_ramht.h"
 
 static void
 nv50_fifo_playlist_update(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_fifo_engine *pfifo = &dev_priv->engine.fifo;
-	struct nouveau_gpuobj_ref *cur;
+	struct nouveau_gpuobj *cur;
 	int i, nr;
 
 	NV_DEBUG(dev, "\n");
@@ -43,12 +44,14 @@ nv50_fifo_playlist_update(struct drm_device *dev)
 
 	/* We never schedule channel 0 or 127 */
 	for (i = 1, nr = 0; i < 127; i++) {
-		if (dev_priv->fifos[i] && dev_priv->fifos[i]->ramfc)
-			nv_wo32(dev, cur->gpuobj, nr++, i);
+		if (dev_priv->fifos[i] && dev_priv->fifos[i]->ramfc) {
+			nv_wo32(cur, (nr * 4), i);
+			nr++;
+		}
 	}
 	dev_priv->engine.instmem.flush(dev);
 
-	nv_wr32(dev, 0x32f4, cur->instance >> 12);
+	nv_wr32(dev, 0x32f4, cur->vinst >> 12);
 	nv_wr32(dev, 0x32ec, nr);
 	nv_wr32(dev, 0x2500, 0x101);
 }
@@ -63,9 +66,9 @@ nv50_fifo_channel_enable(struct drm_device *dev, int channel)
 	NV_DEBUG(dev, "ch%d\n", channel);
 
 	if (dev_priv->chipset == 0x50)
-		inst = chan->ramfc->instance >> 12;
+		inst = chan->ramfc->vinst >> 12;
 	else
-		inst = chan->ramfc->instance >> 8;
+		inst = chan->ramfc->vinst >> 8;
 
 	nv_wr32(dev, NV50_PFIFO_CTX_TABLE(channel), inst |
 		     NV50_PFIFO_CTX_TABLE_CHANNEL_ENABLED);
@@ -163,19 +166,19 @@ nv50_fifo_init(struct drm_device *dev)
 		goto just_reset;
 	}
 
-	ret = nouveau_gpuobj_new_ref(dev, NULL, NULL, 0, 128*4, 0x1000,
-				     NVOBJ_FLAG_ZERO_ALLOC,
-				     &pfifo->playlist[0]);
+	ret = nouveau_gpuobj_new(dev, NULL, 128*4, 0x1000,
+				 NVOBJ_FLAG_ZERO_ALLOC,
+				 &pfifo->playlist[0]);
 	if (ret) {
 		NV_ERROR(dev, "error creating playlist 0: %d\n", ret);
 		return ret;
 	}
 
-	ret = nouveau_gpuobj_new_ref(dev, NULL, NULL, 0, 128*4, 0x1000,
-				     NVOBJ_FLAG_ZERO_ALLOC,
-				     &pfifo->playlist[1]);
+	ret = nouveau_gpuobj_new(dev, NULL, 128*4, 0x1000,
+				 NVOBJ_FLAG_ZERO_ALLOC,
+				 &pfifo->playlist[1]);
 	if (ret) {
-		nouveau_gpuobj_ref_del(dev, &pfifo->playlist[0]);
+		nouveau_gpuobj_ref(NULL, &pfifo->playlist[0]);
 		NV_ERROR(dev, "error creating playlist 1: %d\n", ret);
 		return ret;
 	}
@@ -203,8 +206,8 @@ nv50_fifo_takedown(struct drm_device *dev)
 	if (!pfifo->playlist[0])
 		return;
 
-	nouveau_gpuobj_ref_del(dev, &pfifo->playlist[0]);
-	nouveau_gpuobj_ref_del(dev, &pfifo->playlist[1]);
+	nouveau_gpuobj_ref(NULL, &pfifo->playlist[0]);
+	nouveau_gpuobj_ref(NULL, &pfifo->playlist[1]);
 }
 
 int
@@ -226,59 +229,54 @@ nv50_fifo_create_context(struct nouveau_channel *chan)
 	NV_DEBUG(dev, "ch%d\n", chan->id);
 
 	if (dev_priv->chipset == 0x50) {
-		uint32_t ramin_poffset = chan->ramin->gpuobj->im_pramin->start;
-		uint32_t ramin_voffset = chan->ramin->gpuobj->im_backing_start;
-
-		ret = nouveau_gpuobj_new_fake(dev, ramin_poffset, ramin_voffset,
-					      0x100, NVOBJ_FLAG_ZERO_ALLOC |
-					      NVOBJ_FLAG_ZERO_FREE, &ramfc,
+		ret = nouveau_gpuobj_new_fake(dev, chan->ramin->pinst,
+					      chan->ramin->vinst, 0x100,
+					      NVOBJ_FLAG_ZERO_ALLOC |
+					      NVOBJ_FLAG_ZERO_FREE,
 					      &chan->ramfc);
 		if (ret)
 			return ret;
 
-		ret = nouveau_gpuobj_new_fake(dev, ramin_poffset + 0x0400,
-					      ramin_voffset + 0x0400, 4096,
-					      0, NULL, &chan->cache);
+		ret = nouveau_gpuobj_new_fake(dev, chan->ramin->pinst + 0x0400,
+					      chan->ramin->vinst + 0x0400,
+					      4096, 0, &chan->cache);
 		if (ret)
 			return ret;
 	} else {
-		ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, 0x100, 256,
-					     NVOBJ_FLAG_ZERO_ALLOC |
-					     NVOBJ_FLAG_ZERO_FREE,
-					     &chan->ramfc);
+		ret = nouveau_gpuobj_new(dev, chan, 0x100, 256,
+					 NVOBJ_FLAG_ZERO_ALLOC |
+					 NVOBJ_FLAG_ZERO_FREE, &chan->ramfc);
 		if (ret)
 			return ret;
-		ramfc = chan->ramfc->gpuobj;
 
-		ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, 4096, 1024,
-					     0, &chan->cache);
+		ret = nouveau_gpuobj_new(dev, chan, 4096, 1024,
+					 0, &chan->cache);
 		if (ret)
 			return ret;
 	}
+	ramfc = chan->ramfc;
 
 	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
 
-	nv_wo32(dev, ramfc, 0x48/4, chan->pushbuf->instance >> 4);
-	nv_wo32(dev, ramfc, 0x80/4, (0 << 27) /* 4KiB */ |
-				    (4 << 24) /* SEARCH_FULL */ |
-				    (chan->ramht->instance >> 4));
-	nv_wo32(dev, ramfc, 0x44/4, 0x2101ffff);
-	nv_wo32(dev, ramfc, 0x60/4, 0x7fffffff);
-	nv_wo32(dev, ramfc, 0x40/4, 0x00000000);
-	nv_wo32(dev, ramfc, 0x7c/4, 0x30000001);
-	nv_wo32(dev, ramfc, 0x78/4, 0x00000000);
-	nv_wo32(dev, ramfc, 0x3c/4, 0x403f6078);
-	nv_wo32(dev, ramfc, 0x50/4, chan->pushbuf_base +
-				    chan->dma.ib_base * 4);
-	nv_wo32(dev, ramfc, 0x54/4, drm_order(chan->dma.ib_max + 1) << 16);
+	nv_wo32(ramfc, 0x48, chan->pushbuf->cinst >> 4);
+	nv_wo32(ramfc, 0x80, ((chan->ramht->bits - 9) << 27) |
+			     (4 << 24) /* SEARCH_FULL */ |
+			     (chan->ramht->gpuobj->cinst >> 4));
+	nv_wo32(ramfc, 0x44, 0x2101ffff);
+	nv_wo32(ramfc, 0x60, 0x7fffffff);
+	nv_wo32(ramfc, 0x40, 0x00000000);
+	nv_wo32(ramfc, 0x7c, 0x30000001);
+	nv_wo32(ramfc, 0x78, 0x00000000);
+	nv_wo32(ramfc, 0x3c, 0x403f6078);
+	nv_wo32(ramfc, 0x50, chan->pushbuf_base + chan->dma.ib_base * 4);
+	nv_wo32(ramfc, 0x54, drm_order(chan->dma.ib_max + 1) << 16);
 
 	if (dev_priv->chipset != 0x50) {
-		nv_wo32(dev, chan->ramin->gpuobj, 0, chan->id);
-		nv_wo32(dev, chan->ramin->gpuobj, 1,
-						chan->ramfc->instance >> 8);
+		nv_wo32(chan->ramin, 0, chan->id);
+		nv_wo32(chan->ramin, 4, chan->ramfc->vinst >> 8);
 
-		nv_wo32(dev, ramfc, 0x88/4, chan->cache->instance >> 10);
-		nv_wo32(dev, ramfc, 0x98/4, chan->ramin->instance >> 12);
+		nv_wo32(ramfc, 0x88, chan->cache->vinst >> 10);
+		nv_wo32(ramfc, 0x98, chan->ramin->vinst >> 12);
 	}
 
 	dev_priv->engine.instmem.flush(dev);
@@ -293,12 +291,13 @@ void
 nv50_fifo_destroy_context(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->dev;
-	struct nouveau_gpuobj_ref *ramfc = chan->ramfc;
+	struct nouveau_gpuobj *ramfc = NULL;
 
 	NV_DEBUG(dev, "ch%d\n", chan->id);
 
 	/* This will ensure the channel is seen as disabled. */
-	chan->ramfc = NULL;
+	nouveau_gpuobj_ref(chan->ramfc, &ramfc);
+	nouveau_gpuobj_ref(NULL, &chan->ramfc);
 	nv50_fifo_channel_disable(dev, chan->id);
 
 	/* Dummy channel, also used on ch 127 */
@@ -306,8 +305,8 @@ nv50_fifo_destroy_context(struct nouveau_channel *chan)
 		nv50_fifo_channel_disable(dev, 127);
 	nv50_fifo_playlist_update(dev);
 
-	nouveau_gpuobj_ref_del(dev, &ramfc);
-	nouveau_gpuobj_ref_del(dev, &chan->cache);
+	nouveau_gpuobj_ref(NULL, &ramfc);
+	nouveau_gpuobj_ref(NULL, &chan->cache);
 }
 
 int
@@ -315,63 +314,63 @@ nv50_fifo_load_context(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *ramfc = chan->ramfc->gpuobj;
-	struct nouveau_gpuobj *cache = chan->cache->gpuobj;
+	struct nouveau_gpuobj *ramfc = chan->ramfc;
+	struct nouveau_gpuobj *cache = chan->cache;
 	int ptr, cnt;
 
 	NV_DEBUG(dev, "ch%d\n", chan->id);
 
-	nv_wr32(dev, 0x3330, nv_ro32(dev, ramfc, 0x00/4));
-	nv_wr32(dev, 0x3334, nv_ro32(dev, ramfc, 0x04/4));
-	nv_wr32(dev, 0x3240, nv_ro32(dev, ramfc, 0x08/4));
-	nv_wr32(dev, 0x3320, nv_ro32(dev, ramfc, 0x0c/4));
-	nv_wr32(dev, 0x3244, nv_ro32(dev, ramfc, 0x10/4));
-	nv_wr32(dev, 0x3328, nv_ro32(dev, ramfc, 0x14/4));
-	nv_wr32(dev, 0x3368, nv_ro32(dev, ramfc, 0x18/4));
-	nv_wr32(dev, 0x336c, nv_ro32(dev, ramfc, 0x1c/4));
-	nv_wr32(dev, 0x3370, nv_ro32(dev, ramfc, 0x20/4));
-	nv_wr32(dev, 0x3374, nv_ro32(dev, ramfc, 0x24/4));
-	nv_wr32(dev, 0x3378, nv_ro32(dev, ramfc, 0x28/4));
-	nv_wr32(dev, 0x337c, nv_ro32(dev, ramfc, 0x2c/4));
-	nv_wr32(dev, 0x3228, nv_ro32(dev, ramfc, 0x30/4));
-	nv_wr32(dev, 0x3364, nv_ro32(dev, ramfc, 0x34/4));
-	nv_wr32(dev, 0x32a0, nv_ro32(dev, ramfc, 0x38/4));
-	nv_wr32(dev, 0x3224, nv_ro32(dev, ramfc, 0x3c/4));
-	nv_wr32(dev, 0x324c, nv_ro32(dev, ramfc, 0x40/4));
-	nv_wr32(dev, 0x2044, nv_ro32(dev, ramfc, 0x44/4));
-	nv_wr32(dev, 0x322c, nv_ro32(dev, ramfc, 0x48/4));
-	nv_wr32(dev, 0x3234, nv_ro32(dev, ramfc, 0x4c/4));
-	nv_wr32(dev, 0x3340, nv_ro32(dev, ramfc, 0x50/4));
-	nv_wr32(dev, 0x3344, nv_ro32(dev, ramfc, 0x54/4));
-	nv_wr32(dev, 0x3280, nv_ro32(dev, ramfc, 0x58/4));
-	nv_wr32(dev, 0x3254, nv_ro32(dev, ramfc, 0x5c/4));
-	nv_wr32(dev, 0x3260, nv_ro32(dev, ramfc, 0x60/4));
-	nv_wr32(dev, 0x3264, nv_ro32(dev, ramfc, 0x64/4));
-	nv_wr32(dev, 0x3268, nv_ro32(dev, ramfc, 0x68/4));
-	nv_wr32(dev, 0x326c, nv_ro32(dev, ramfc, 0x6c/4));
-	nv_wr32(dev, 0x32e4, nv_ro32(dev, ramfc, 0x70/4));
-	nv_wr32(dev, 0x3248, nv_ro32(dev, ramfc, 0x74/4));
-	nv_wr32(dev, 0x2088, nv_ro32(dev, ramfc, 0x78/4));
-	nv_wr32(dev, 0x2058, nv_ro32(dev, ramfc, 0x7c/4));
-	nv_wr32(dev, 0x2210, nv_ro32(dev, ramfc, 0x80/4));
+	nv_wr32(dev, 0x3330, nv_ro32(ramfc, 0x00));
+	nv_wr32(dev, 0x3334, nv_ro32(ramfc, 0x04));
+	nv_wr32(dev, 0x3240, nv_ro32(ramfc, 0x08));
+	nv_wr32(dev, 0x3320, nv_ro32(ramfc, 0x0c));
+	nv_wr32(dev, 0x3244, nv_ro32(ramfc, 0x10));
+	nv_wr32(dev, 0x3328, nv_ro32(ramfc, 0x14));
+	nv_wr32(dev, 0x3368, nv_ro32(ramfc, 0x18));
+	nv_wr32(dev, 0x336c, nv_ro32(ramfc, 0x1c));
+	nv_wr32(dev, 0x3370, nv_ro32(ramfc, 0x20));
+	nv_wr32(dev, 0x3374, nv_ro32(ramfc, 0x24));
+	nv_wr32(dev, 0x3378, nv_ro32(ramfc, 0x28));
+	nv_wr32(dev, 0x337c, nv_ro32(ramfc, 0x2c));
+	nv_wr32(dev, 0x3228, nv_ro32(ramfc, 0x30));
+	nv_wr32(dev, 0x3364, nv_ro32(ramfc, 0x34));
+	nv_wr32(dev, 0x32a0, nv_ro32(ramfc, 0x38));
+	nv_wr32(dev, 0x3224, nv_ro32(ramfc, 0x3c));
+	nv_wr32(dev, 0x324c, nv_ro32(ramfc, 0x40));
+	nv_wr32(dev, 0x2044, nv_ro32(ramfc, 0x44));
+	nv_wr32(dev, 0x322c, nv_ro32(ramfc, 0x48));
+	nv_wr32(dev, 0x3234, nv_ro32(ramfc, 0x4c));
+	nv_wr32(dev, 0x3340, nv_ro32(ramfc, 0x50));
+	nv_wr32(dev, 0x3344, nv_ro32(ramfc, 0x54));
+	nv_wr32(dev, 0x3280, nv_ro32(ramfc, 0x58));
+	nv_wr32(dev, 0x3254, nv_ro32(ramfc, 0x5c));
+	nv_wr32(dev, 0x3260, nv_ro32(ramfc, 0x60));
+	nv_wr32(dev, 0x3264, nv_ro32(ramfc, 0x64));
+	nv_wr32(dev, 0x3268, nv_ro32(ramfc, 0x68));
+	nv_wr32(dev, 0x326c, nv_ro32(ramfc, 0x6c));
+	nv_wr32(dev, 0x32e4, nv_ro32(ramfc, 0x70));
+	nv_wr32(dev, 0x3248, nv_ro32(ramfc, 0x74));
+	nv_wr32(dev, 0x2088, nv_ro32(ramfc, 0x78));
+	nv_wr32(dev, 0x2058, nv_ro32(ramfc, 0x7c));
+	nv_wr32(dev, 0x2210, nv_ro32(ramfc, 0x80));
 
-	cnt = nv_ro32(dev, ramfc, 0x84/4);
+	cnt = nv_ro32(ramfc, 0x84);
 	for (ptr = 0; ptr < cnt; ptr++) {
 		nv_wr32(dev, NV40_PFIFO_CACHE1_METHOD(ptr),
-			nv_ro32(dev, cache, (ptr * 2) + 0));
+			nv_ro32(cache, (ptr * 8) + 0));
 		nv_wr32(dev, NV40_PFIFO_CACHE1_DATA(ptr),
-			nv_ro32(dev, cache, (ptr * 2) + 1));
+			nv_ro32(cache, (ptr * 8) + 4));
 	}
 	nv_wr32(dev, NV03_PFIFO_CACHE1_PUT, cnt << 2);
 	nv_wr32(dev, NV03_PFIFO_CACHE1_GET, 0);
 
 	/* guessing that all the 0x34xx regs aren't on NV50 */
 	if (dev_priv->chipset != 0x50) {
-		nv_wr32(dev, 0x340c, nv_ro32(dev, ramfc, 0x88/4));
-		nv_wr32(dev, 0x3400, nv_ro32(dev, ramfc, 0x8c/4));
-		nv_wr32(dev, 0x3404, nv_ro32(dev, ramfc, 0x90/4));
-		nv_wr32(dev, 0x3408, nv_ro32(dev, ramfc, 0x94/4));
-		nv_wr32(dev, 0x3410, nv_ro32(dev, ramfc, 0x98/4));
+		nv_wr32(dev, 0x340c, nv_ro32(ramfc, 0x88));
+		nv_wr32(dev, 0x3400, nv_ro32(ramfc, 0x8c));
+		nv_wr32(dev, 0x3404, nv_ro32(ramfc, 0x90));
+		nv_wr32(dev, 0x3408, nv_ro32(ramfc, 0x94));
+		nv_wr32(dev, 0x3410, nv_ro32(ramfc, 0x98));
 	}
 
 	nv_wr32(dev, NV03_PFIFO_CACHE1_PUSH1, chan->id | (1<<16));
@@ -399,62 +398,63 @@ nv50_fifo_unload_context(struct drm_device *dev)
 		return -EINVAL;
 	}
 	NV_DEBUG(dev, "ch%d\n", chan->id);
-	ramfc = chan->ramfc->gpuobj;
-	cache = chan->cache->gpuobj;
+	ramfc = chan->ramfc;
+	cache = chan->cache;
 
-	nv_wo32(dev, ramfc, 0x00/4, nv_rd32(dev, 0x3330));
-	nv_wo32(dev, ramfc, 0x04/4, nv_rd32(dev, 0x3334));
-	nv_wo32(dev, ramfc, 0x08/4, nv_rd32(dev, 0x3240));
-	nv_wo32(dev, ramfc, 0x0c/4, nv_rd32(dev, 0x3320));
-	nv_wo32(dev, ramfc, 0x10/4, nv_rd32(dev, 0x3244));
-	nv_wo32(dev, ramfc, 0x14/4, nv_rd32(dev, 0x3328));
-	nv_wo32(dev, ramfc, 0x18/4, nv_rd32(dev, 0x3368));
-	nv_wo32(dev, ramfc, 0x1c/4, nv_rd32(dev, 0x336c));
-	nv_wo32(dev, ramfc, 0x20/4, nv_rd32(dev, 0x3370));
-	nv_wo32(dev, ramfc, 0x24/4, nv_rd32(dev, 0x3374));
-	nv_wo32(dev, ramfc, 0x28/4, nv_rd32(dev, 0x3378));
-	nv_wo32(dev, ramfc, 0x2c/4, nv_rd32(dev, 0x337c));
-	nv_wo32(dev, ramfc, 0x30/4, nv_rd32(dev, 0x3228));
-	nv_wo32(dev, ramfc, 0x34/4, nv_rd32(dev, 0x3364));
-	nv_wo32(dev, ramfc, 0x38/4, nv_rd32(dev, 0x32a0));
-	nv_wo32(dev, ramfc, 0x3c/4, nv_rd32(dev, 0x3224));
-	nv_wo32(dev, ramfc, 0x40/4, nv_rd32(dev, 0x324c));
-	nv_wo32(dev, ramfc, 0x44/4, nv_rd32(dev, 0x2044));
-	nv_wo32(dev, ramfc, 0x48/4, nv_rd32(dev, 0x322c));
-	nv_wo32(dev, ramfc, 0x4c/4, nv_rd32(dev, 0x3234));
-	nv_wo32(dev, ramfc, 0x50/4, nv_rd32(dev, 0x3340));
-	nv_wo32(dev, ramfc, 0x54/4, nv_rd32(dev, 0x3344));
-	nv_wo32(dev, ramfc, 0x58/4, nv_rd32(dev, 0x3280));
-	nv_wo32(dev, ramfc, 0x5c/4, nv_rd32(dev, 0x3254));
-	nv_wo32(dev, ramfc, 0x60/4, nv_rd32(dev, 0x3260));
-	nv_wo32(dev, ramfc, 0x64/4, nv_rd32(dev, 0x3264));
-	nv_wo32(dev, ramfc, 0x68/4, nv_rd32(dev, 0x3268));
-	nv_wo32(dev, ramfc, 0x6c/4, nv_rd32(dev, 0x326c));
-	nv_wo32(dev, ramfc, 0x70/4, nv_rd32(dev, 0x32e4));
-	nv_wo32(dev, ramfc, 0x74/4, nv_rd32(dev, 0x3248));
-	nv_wo32(dev, ramfc, 0x78/4, nv_rd32(dev, 0x2088));
-	nv_wo32(dev, ramfc, 0x7c/4, nv_rd32(dev, 0x2058));
-	nv_wo32(dev, ramfc, 0x80/4, nv_rd32(dev, 0x2210));
+	nv_wo32(ramfc, 0x00, nv_rd32(dev, 0x3330));
+	nv_wo32(ramfc, 0x04, nv_rd32(dev, 0x3334));
+	nv_wo32(ramfc, 0x08, nv_rd32(dev, 0x3240));
+	nv_wo32(ramfc, 0x0c, nv_rd32(dev, 0x3320));
+	nv_wo32(ramfc, 0x10, nv_rd32(dev, 0x3244));
+	nv_wo32(ramfc, 0x14, nv_rd32(dev, 0x3328));
+	nv_wo32(ramfc, 0x18, nv_rd32(dev, 0x3368));
+	nv_wo32(ramfc, 0x1c, nv_rd32(dev, 0x336c));
+	nv_wo32(ramfc, 0x20, nv_rd32(dev, 0x3370));
+	nv_wo32(ramfc, 0x24, nv_rd32(dev, 0x3374));
+	nv_wo32(ramfc, 0x28, nv_rd32(dev, 0x3378));
+	nv_wo32(ramfc, 0x2c, nv_rd32(dev, 0x337c));
+	nv_wo32(ramfc, 0x30, nv_rd32(dev, 0x3228));
+	nv_wo32(ramfc, 0x34, nv_rd32(dev, 0x3364));
+	nv_wo32(ramfc, 0x38, nv_rd32(dev, 0x32a0));
+	nv_wo32(ramfc, 0x3c, nv_rd32(dev, 0x3224));
+	nv_wo32(ramfc, 0x40, nv_rd32(dev, 0x324c));
+	nv_wo32(ramfc, 0x44, nv_rd32(dev, 0x2044));
+	nv_wo32(ramfc, 0x48, nv_rd32(dev, 0x322c));
+	nv_wo32(ramfc, 0x4c, nv_rd32(dev, 0x3234));
+	nv_wo32(ramfc, 0x50, nv_rd32(dev, 0x3340));
+	nv_wo32(ramfc, 0x54, nv_rd32(dev, 0x3344));
+	nv_wo32(ramfc, 0x58, nv_rd32(dev, 0x3280));
+	nv_wo32(ramfc, 0x5c, nv_rd32(dev, 0x3254));
+	nv_wo32(ramfc, 0x60, nv_rd32(dev, 0x3260));
+	nv_wo32(ramfc, 0x64, nv_rd32(dev, 0x3264));
+	nv_wo32(ramfc, 0x68, nv_rd32(dev, 0x3268));
+	nv_wo32(ramfc, 0x6c, nv_rd32(dev, 0x326c));
+	nv_wo32(ramfc, 0x70, nv_rd32(dev, 0x32e4));
+	nv_wo32(ramfc, 0x74, nv_rd32(dev, 0x3248));
+	nv_wo32(ramfc, 0x78, nv_rd32(dev, 0x2088));
+	nv_wo32(ramfc, 0x7c, nv_rd32(dev, 0x2058));
+	nv_wo32(ramfc, 0x80, nv_rd32(dev, 0x2210));
 
 	put = (nv_rd32(dev, NV03_PFIFO_CACHE1_PUT) & 0x7ff) >> 2;
 	get = (nv_rd32(dev, NV03_PFIFO_CACHE1_GET) & 0x7ff) >> 2;
 	ptr = 0;
 	while (put != get) {
-		nv_wo32(dev, cache, ptr++,
-			    nv_rd32(dev, NV40_PFIFO_CACHE1_METHOD(get)));
-		nv_wo32(dev, cache, ptr++,
-			    nv_rd32(dev, NV40_PFIFO_CACHE1_DATA(get)));
+		nv_wo32(cache, ptr + 0,
+			nv_rd32(dev, NV40_PFIFO_CACHE1_METHOD(get)));
+		nv_wo32(cache, ptr + 4,
+			nv_rd32(dev, NV40_PFIFO_CACHE1_DATA(get)));
 		get = (get + 1) & 0x1ff;
+		ptr += 8;
 	}
 
 	/* guessing that all the 0x34xx regs aren't on NV50 */
 	if (dev_priv->chipset != 0x50) {
-		nv_wo32(dev, ramfc, 0x84/4, ptr >> 1);
-		nv_wo32(dev, ramfc, 0x88/4, nv_rd32(dev, 0x340c));
-		nv_wo32(dev, ramfc, 0x8c/4, nv_rd32(dev, 0x3400));
-		nv_wo32(dev, ramfc, 0x90/4, nv_rd32(dev, 0x3404));
-		nv_wo32(dev, ramfc, 0x94/4, nv_rd32(dev, 0x3408));
-		nv_wo32(dev, ramfc, 0x98/4, nv_rd32(dev, 0x3410));
+		nv_wo32(ramfc, 0x84, ptr >> 3);
+		nv_wo32(ramfc, 0x88, nv_rd32(dev, 0x340c));
+		nv_wo32(ramfc, 0x8c, nv_rd32(dev, 0x3400));
+		nv_wo32(ramfc, 0x90, nv_rd32(dev, 0x3404));
+		nv_wo32(ramfc, 0x94, nv_rd32(dev, 0x3408));
+		nv_wo32(ramfc, 0x98, nv_rd32(dev, 0x3410));
 	}
 
 	dev_priv->engine.instmem.flush(dev);
diff --git a/drivers/gpu/drm/nouveau/nv50_graph.c b/drivers/gpu/drm/nouveau/nv50_graph.c
index 1413028e1580..cbf5ae2f67d4 100644
--- a/drivers/gpu/drm/nouveau/nv50_graph.c
+++ b/drivers/gpu/drm/nouveau/nv50_graph.c
@@ -27,7 +27,7 @@
 #include "drmP.h"
 #include "drm.h"
 #include "nouveau_drv.h"
-
+#include "nouveau_ramht.h"
 #include "nouveau_grctx.h"
 
 static void
@@ -181,7 +181,7 @@ nv50_graph_channel(struct drm_device *dev)
 	/* Be sure we're not in the middle of a context switch or bad things
 	 * will happen, such as unloading the wrong pgraph context.
 	 */
-	if (!nv_wait(0x400300, 0x00000001, 0x00000000))
+	if (!nv_wait(dev, 0x400300, 0x00000001, 0x00000000))
 		NV_ERROR(dev, "Ctxprog is still running\n");
 
 	inst = nv_rd32(dev, NV50_PGRAPH_CTXCTL_CUR);
@@ -192,7 +192,7 @@ nv50_graph_channel(struct drm_device *dev)
 	for (i = 0; i < dev_priv->engine.fifo.channels; i++) {
 		struct nouveau_channel *chan = dev_priv->fifos[i];
 
-		if (chan && chan->ramin && chan->ramin->instance == inst)
+		if (chan && chan->ramin && chan->ramin->vinst == inst)
 			return chan;
 	}
 
@@ -204,36 +204,34 @@ nv50_graph_create_context(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *ramin = chan->ramin->gpuobj;
-	struct nouveau_gpuobj *obj;
+	struct nouveau_gpuobj *ramin = chan->ramin;
 	struct nouveau_pgraph_engine *pgraph = &dev_priv->engine.graph;
 	struct nouveau_grctx ctx = {};
 	int hdr, ret;
 
 	NV_DEBUG(dev, "ch%d\n", chan->id);
 
-	ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, pgraph->grctx_size,
-				     0x1000, NVOBJ_FLAG_ZERO_ALLOC |
-				     NVOBJ_FLAG_ZERO_FREE, &chan->ramin_grctx);
+	ret = nouveau_gpuobj_new(dev, chan, pgraph->grctx_size, 0x1000,
+				 NVOBJ_FLAG_ZERO_ALLOC |
+				 NVOBJ_FLAG_ZERO_FREE, &chan->ramin_grctx);
 	if (ret)
 		return ret;
-	obj = chan->ramin_grctx->gpuobj;
 
 	hdr = (dev_priv->chipset == 0x50) ? 0x200 : 0x20;
-	nv_wo32(dev, ramin, (hdr + 0x00)/4, 0x00190002);
-	nv_wo32(dev, ramin, (hdr + 0x04)/4, chan->ramin_grctx->instance +
-					   pgraph->grctx_size - 1);
-	nv_wo32(dev, ramin, (hdr + 0x08)/4, chan->ramin_grctx->instance);
-	nv_wo32(dev, ramin, (hdr + 0x0c)/4, 0);
-	nv_wo32(dev, ramin, (hdr + 0x10)/4, 0);
-	nv_wo32(dev, ramin, (hdr + 0x14)/4, 0x00010000);
+	nv_wo32(ramin, hdr + 0x00, 0x00190002);
+	nv_wo32(ramin, hdr + 0x04, chan->ramin_grctx->vinst +
+				   pgraph->grctx_size - 1);
+	nv_wo32(ramin, hdr + 0x08, chan->ramin_grctx->vinst);
+	nv_wo32(ramin, hdr + 0x0c, 0);
+	nv_wo32(ramin, hdr + 0x10, 0);
+	nv_wo32(ramin, hdr + 0x14, 0x00010000);
 
 	ctx.dev = chan->dev;
 	ctx.mode = NOUVEAU_GRCTX_VALS;
-	ctx.data = obj;
+	ctx.data = chan->ramin_grctx;
 	nv50_grctx_init(&ctx);
 
-	nv_wo32(dev, obj, 0x00000/4, chan->ramin->instance >> 12);
+	nv_wo32(chan->ramin_grctx, 0x00000, chan->ramin->vinst >> 12);
 
 	dev_priv->engine.instmem.flush(dev);
 	return 0;
@@ -248,14 +246,14 @@ nv50_graph_destroy_context(struct nouveau_channel *chan)
 
 	NV_DEBUG(dev, "ch%d\n", chan->id);
 
-	if (!chan->ramin || !chan->ramin->gpuobj)
+	if (!chan->ramin)
 		return;
 
 	for (i = hdr; i < hdr + 24; i += 4)
-		nv_wo32(dev, chan->ramin->gpuobj, i/4, 0);
+		nv_wo32(chan->ramin, i, 0);
 	dev_priv->engine.instmem.flush(dev);
 
-	nouveau_gpuobj_ref_del(dev, &chan->ramin_grctx);
+	nouveau_gpuobj_ref(NULL, &chan->ramin_grctx);
 }
 
 static int
@@ -282,7 +280,7 @@ nv50_graph_do_load_context(struct drm_device *dev, uint32_t inst)
 int
 nv50_graph_load_context(struct nouveau_channel *chan)
 {
-	uint32_t inst = chan->ramin->instance >> 12;
+	uint32_t inst = chan->ramin->vinst >> 12;
 
 	NV_DEBUG(chan->dev, "ch%d\n", chan->id);
 	return nv50_graph_do_load_context(chan->dev, inst);
@@ -327,15 +325,16 @@ static int
 nv50_graph_nvsw_dma_vblsem(struct nouveau_channel *chan, int grclass,
 			   int mthd, uint32_t data)
 {
-	struct nouveau_gpuobj_ref *ref = NULL;
+	struct nouveau_gpuobj *gpuobj;
 
-	if (nouveau_gpuobj_ref_find(chan, data, &ref))
+	gpuobj = nouveau_ramht_find(chan, data);
+	if (!gpuobj)
 		return -ENOENT;
 
-	if (nouveau_notifier_offset(ref->gpuobj, NULL))
+	if (nouveau_notifier_offset(gpuobj, NULL))
 		return -EINVAL;
 
-	chan->nvsw.vblsem = ref->gpuobj;
+	chan->nvsw.vblsem = gpuobj;
 	chan->nvsw.vblsem_offset = ~0;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_grctx.c b/drivers/gpu/drm/nouveau/nv50_grctx.c
index 42a8fb20c1e6..336aab2a24a6 100644
--- a/drivers/gpu/drm/nouveau/nv50_grctx.c
+++ b/drivers/gpu/drm/nouveau/nv50_grctx.c
@@ -103,6 +103,9 @@
 #include "nouveau_drv.h"
 #include "nouveau_grctx.h"
 
+#define IS_NVA3F(x) (((x) > 0xa0 && (x) < 0xaa) || (x) == 0xaf)
+#define IS_NVAAF(x) ((x) >= 0xaa && (x) <= 0xac)
+
 /*
  * This code deals with PGRAPH contexts on NV50 family cards. Like NV40, it's
  * the GPU itself that does context-switching, but it needs a special
@@ -182,6 +185,7 @@ nv50_grctx_init(struct nouveau_grctx *ctx)
 	case 0xa8:
 	case 0xaa:
 	case 0xac:
+	case 0xaf:
 		break;
 	default:
 		NV_ERROR(ctx->dev, "I don't know how to make a ctxprog for "
@@ -267,6 +271,9 @@ nv50_grctx_init(struct nouveau_grctx *ctx)
  * registers to save/restore and the default values for them.
  */
 
+static void
+nv50_graph_construct_mmio_ddata(struct nouveau_grctx *ctx);
+
 static void
 nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 {
@@ -286,7 +293,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 		gr_def(ctx, 0x400840, 0xffe806a8);
 	}
 	gr_def(ctx, 0x400844, 0x00000002);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
+	if (IS_NVA3F(dev_priv->chipset))
 		gr_def(ctx, 0x400894, 0x00001000);
 	gr_def(ctx, 0x4008e8, 0x00000003);
 	gr_def(ctx, 0x4008ec, 0x00001000);
@@ -299,13 +306,15 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 
 	if (dev_priv->chipset >= 0xa0)
 		cp_ctx(ctx, 0x400b00, 0x1);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
+	if (IS_NVA3F(dev_priv->chipset)) {
 		cp_ctx(ctx, 0x400b10, 0x1);
 		gr_def(ctx, 0x400b10, 0x0001629d);
 		cp_ctx(ctx, 0x400b20, 0x1);
 		gr_def(ctx, 0x400b20, 0x0001629d);
 	}
 
+	nv50_graph_construct_mmio_ddata(ctx);
+
 	/* 0C00: VFETCH */
 	cp_ctx(ctx, 0x400c08, 0x2);
 	gr_def(ctx, 0x400c08, 0x0000fe0c);
@@ -314,7 +323,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 	if (dev_priv->chipset < 0xa0) {
 		cp_ctx(ctx, 0x401008, 0x4);
 		gr_def(ctx, 0x401014, 0x00001000);
-	} else if (dev_priv->chipset == 0xa0 || dev_priv->chipset >= 0xaa) {
+	} else if (!IS_NVA3F(dev_priv->chipset)) {
 		cp_ctx(ctx, 0x401008, 0x5);
 		gr_def(ctx, 0x401018, 0x00001000);
 	} else {
@@ -368,10 +377,13 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 	case 0xa3:
 	case 0xa5:
 	case 0xa8:
+	case 0xaf:
 		gr_def(ctx, 0x401c00, 0x142500df);
 		break;
 	}
 
+	/* 2000 */
+
 	/* 2400 */
 	cp_ctx(ctx, 0x402400, 0x1);
 	if (dev_priv->chipset == 0x50)
@@ -380,12 +392,12 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 		cp_ctx(ctx, 0x402408, 0x2);
 	gr_def(ctx, 0x402408, 0x00000600);
 
-	/* 2800 */
+	/* 2800: CSCHED */
 	cp_ctx(ctx, 0x402800, 0x1);
 	if (dev_priv->chipset == 0x50)
 		gr_def(ctx, 0x402800, 0x00000006);
 
-	/* 2C00 */
+	/* 2C00: ZCULL */
 	cp_ctx(ctx, 0x402c08, 0x6);
 	if (dev_priv->chipset != 0x50)
 		gr_def(ctx, 0x402c14, 0x01000000);
@@ -396,23 +408,23 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 		cp_ctx(ctx, 0x402ca0, 0x2);
 	if (dev_priv->chipset < 0xa0)
 		gr_def(ctx, 0x402ca0, 0x00000400);
-	else if (dev_priv->chipset == 0xa0 || dev_priv->chipset >= 0xaa)
+	else if (!IS_NVA3F(dev_priv->chipset))
 		gr_def(ctx, 0x402ca0, 0x00000800);
 	else
 		gr_def(ctx, 0x402ca0, 0x00000400);
 	cp_ctx(ctx, 0x402cac, 0x4);
 
-	/* 3000 */
+	/* 3000: ENG2D */
 	cp_ctx(ctx, 0x403004, 0x1);
 	gr_def(ctx, 0x403004, 0x00000001);
 
-	/* 3404 */
+	/* 3400 */
 	if (dev_priv->chipset >= 0xa0) {
 		cp_ctx(ctx, 0x403404, 0x1);
 		gr_def(ctx, 0x403404, 0x00000001);
 	}
 
-	/* 5000 */
+	/* 5000: CCACHE */
 	cp_ctx(ctx, 0x405000, 0x1);
 	switch (dev_priv->chipset) {
 	case 0x50:
@@ -425,6 +437,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 	case 0xa8:
 	case 0xaa:
 	case 0xac:
+	case 0xaf:
 		gr_def(ctx, 0x405000, 0x000e0080);
 		break;
 	case 0x86:
@@ -441,210 +454,6 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 	cp_ctx(ctx, 0x405024, 0x1);
 	cp_ctx(ctx, 0x40502c, 0x1);
 
-	/* 5400 or maybe 4800 */
-	if (dev_priv->chipset == 0x50) {
-		offset = 0x405400;
-		cp_ctx(ctx, 0x405400, 0xea);
-	} else if (dev_priv->chipset < 0x94) {
-		offset = 0x405400;
-		cp_ctx(ctx, 0x405400, 0xcb);
-	} else if (dev_priv->chipset < 0xa0) {
-		offset = 0x405400;
-		cp_ctx(ctx, 0x405400, 0xcc);
-	} else if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		offset = 0x404800;
-		cp_ctx(ctx, 0x404800, 0xda);
-	} else {
-		offset = 0x405400;
-		cp_ctx(ctx, 0x405400, 0xd4);
-	}
-	gr_def(ctx, offset + 0x0c, 0x00000002);
-	gr_def(ctx, offset + 0x10, 0x00000001);
-	if (dev_priv->chipset >= 0x94)
-		offset += 4;
-	gr_def(ctx, offset + 0x1c, 0x00000001);
-	gr_def(ctx, offset + 0x20, 0x00000100);
-	gr_def(ctx, offset + 0x38, 0x00000002);
-	gr_def(ctx, offset + 0x3c, 0x00000001);
-	gr_def(ctx, offset + 0x40, 0x00000001);
-	gr_def(ctx, offset + 0x50, 0x00000001);
-	gr_def(ctx, offset + 0x54, 0x003fffff);
-	gr_def(ctx, offset + 0x58, 0x00001fff);
-	gr_def(ctx, offset + 0x60, 0x00000001);
-	gr_def(ctx, offset + 0x64, 0x00000001);
-	gr_def(ctx, offset + 0x6c, 0x00000001);
-	gr_def(ctx, offset + 0x70, 0x00000001);
-	gr_def(ctx, offset + 0x74, 0x00000001);
-	gr_def(ctx, offset + 0x78, 0x00000004);
-	gr_def(ctx, offset + 0x7c, 0x00000001);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		offset += 4;
-	gr_def(ctx, offset + 0x80, 0x00000001);
-	gr_def(ctx, offset + 0x84, 0x00000001);
-	gr_def(ctx, offset + 0x88, 0x00000007);
-	gr_def(ctx, offset + 0x8c, 0x00000001);
-	gr_def(ctx, offset + 0x90, 0x00000007);
-	gr_def(ctx, offset + 0x94, 0x00000001);
-	gr_def(ctx, offset + 0x98, 0x00000001);
-	gr_def(ctx, offset + 0x9c, 0x00000001);
-	if (dev_priv->chipset == 0x50) {
-		 gr_def(ctx, offset + 0xb0, 0x00000001);
-		 gr_def(ctx, offset + 0xb4, 0x00000001);
-		 gr_def(ctx, offset + 0xbc, 0x00000001);
-		 gr_def(ctx, offset + 0xc0, 0x0000000a);
-		 gr_def(ctx, offset + 0xd0, 0x00000040);
-		 gr_def(ctx, offset + 0xd8, 0x00000002);
-		 gr_def(ctx, offset + 0xdc, 0x00000100);
-		 gr_def(ctx, offset + 0xe0, 0x00000001);
-		 gr_def(ctx, offset + 0xe4, 0x00000100);
-		 gr_def(ctx, offset + 0x100, 0x00000001);
-		 gr_def(ctx, offset + 0x124, 0x00000004);
-		 gr_def(ctx, offset + 0x13c, 0x00000001);
-		 gr_def(ctx, offset + 0x140, 0x00000100);
-		 gr_def(ctx, offset + 0x148, 0x00000001);
-		 gr_def(ctx, offset + 0x154, 0x00000100);
-		 gr_def(ctx, offset + 0x158, 0x00000001);
-		 gr_def(ctx, offset + 0x15c, 0x00000100);
-		 gr_def(ctx, offset + 0x164, 0x00000001);
-		 gr_def(ctx, offset + 0x170, 0x00000100);
-		 gr_def(ctx, offset + 0x174, 0x00000001);
-		 gr_def(ctx, offset + 0x17c, 0x00000001);
-		 gr_def(ctx, offset + 0x188, 0x00000002);
-		 gr_def(ctx, offset + 0x190, 0x00000001);
-		 gr_def(ctx, offset + 0x198, 0x00000001);
-		 gr_def(ctx, offset + 0x1ac, 0x00000003);
-		 offset += 0xd0;
-	} else {
-		gr_def(ctx, offset + 0xb0, 0x00000001);
-		gr_def(ctx, offset + 0xb4, 0x00000100);
-		gr_def(ctx, offset + 0xbc, 0x00000001);
-		gr_def(ctx, offset + 0xc8, 0x00000100);
-		gr_def(ctx, offset + 0xcc, 0x00000001);
-		gr_def(ctx, offset + 0xd0, 0x00000100);
-		gr_def(ctx, offset + 0xd8, 0x00000001);
-		gr_def(ctx, offset + 0xe4, 0x00000100);
-	}
-	gr_def(ctx, offset + 0xf8, 0x00000004);
-	gr_def(ctx, offset + 0xfc, 0x00000070);
-	gr_def(ctx, offset + 0x100, 0x00000080);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		offset += 4;
-	gr_def(ctx, offset + 0x114, 0x0000000c);
-	if (dev_priv->chipset == 0x50)
-		offset -= 4;
-	gr_def(ctx, offset + 0x11c, 0x00000008);
-	gr_def(ctx, offset + 0x120, 0x00000014);
-	if (dev_priv->chipset == 0x50) {
-		gr_def(ctx, offset + 0x124, 0x00000026);
-		offset -= 0x18;
-	} else {
-		gr_def(ctx, offset + 0x128, 0x00000029);
-		gr_def(ctx, offset + 0x12c, 0x00000027);
-		gr_def(ctx, offset + 0x130, 0x00000026);
-		gr_def(ctx, offset + 0x134, 0x00000008);
-		gr_def(ctx, offset + 0x138, 0x00000004);
-		gr_def(ctx, offset + 0x13c, 0x00000027);
-	}
-	gr_def(ctx, offset + 0x148, 0x00000001);
-	gr_def(ctx, offset + 0x14c, 0x00000002);
-	gr_def(ctx, offset + 0x150, 0x00000003);
-	gr_def(ctx, offset + 0x154, 0x00000004);
-	gr_def(ctx, offset + 0x158, 0x00000005);
-	gr_def(ctx, offset + 0x15c, 0x00000006);
-	gr_def(ctx, offset + 0x160, 0x00000007);
-	gr_def(ctx, offset + 0x164, 0x00000001);
-	gr_def(ctx, offset + 0x1a8, 0x000000cf);
-	if (dev_priv->chipset == 0x50)
-		offset -= 4;
-	gr_def(ctx, offset + 0x1d8, 0x00000080);
-	gr_def(ctx, offset + 0x1dc, 0x00000004);
-	gr_def(ctx, offset + 0x1e0, 0x00000004);
-	if (dev_priv->chipset == 0x50)
-		offset -= 4;
-	else
-		gr_def(ctx, offset + 0x1e4, 0x00000003);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		gr_def(ctx, offset + 0x1ec, 0x00000003);
-		offset += 8;
-	}
-	gr_def(ctx, offset + 0x1e8, 0x00000001);
-	if (dev_priv->chipset == 0x50)
-		offset -= 4;
-	gr_def(ctx, offset + 0x1f4, 0x00000012);
-	gr_def(ctx, offset + 0x1f8, 0x00000010);
-	gr_def(ctx, offset + 0x1fc, 0x0000000c);
-	gr_def(ctx, offset + 0x200, 0x00000001);
-	gr_def(ctx, offset + 0x210, 0x00000004);
-	gr_def(ctx, offset + 0x214, 0x00000002);
-	gr_def(ctx, offset + 0x218, 0x00000004);
-	if (dev_priv->chipset >= 0xa0)
-		offset += 4;
-	gr_def(ctx, offset + 0x224, 0x003fffff);
-	gr_def(ctx, offset + 0x228, 0x00001fff);
-	if (dev_priv->chipset == 0x50)
-		offset -= 0x20;
-	else if (dev_priv->chipset >= 0xa0) {
-		gr_def(ctx, offset + 0x250, 0x00000001);
-		gr_def(ctx, offset + 0x254, 0x00000001);
-		gr_def(ctx, offset + 0x258, 0x00000002);
-		offset += 0x10;
-	}
-	gr_def(ctx, offset + 0x250, 0x00000004);
-	gr_def(ctx, offset + 0x254, 0x00000014);
-	gr_def(ctx, offset + 0x258, 0x00000001);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		offset += 4;
-	gr_def(ctx, offset + 0x264, 0x00000002);
-	if (dev_priv->chipset >= 0xa0)
-		offset += 8;
-	gr_def(ctx, offset + 0x270, 0x00000001);
-	gr_def(ctx, offset + 0x278, 0x00000002);
-	gr_def(ctx, offset + 0x27c, 0x00001000);
-	if (dev_priv->chipset == 0x50)
-		offset -= 0xc;
-	else {
-		gr_def(ctx, offset + 0x280, 0x00000e00);
-		gr_def(ctx, offset + 0x284, 0x00001000);
-		gr_def(ctx, offset + 0x288, 0x00001e00);
-	}
-	gr_def(ctx, offset + 0x290, 0x00000001);
-	gr_def(ctx, offset + 0x294, 0x00000001);
-	gr_def(ctx, offset + 0x298, 0x00000001);
-	gr_def(ctx, offset + 0x29c, 0x00000001);
-	gr_def(ctx, offset + 0x2a0, 0x00000001);
-	gr_def(ctx, offset + 0x2b0, 0x00000200);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		gr_def(ctx, offset + 0x2b4, 0x00000200);
-		offset += 4;
-	}
-	if (dev_priv->chipset < 0xa0) {
-		gr_def(ctx, offset + 0x2b8, 0x00000001);
-		gr_def(ctx, offset + 0x2bc, 0x00000070);
-		gr_def(ctx, offset + 0x2c0, 0x00000080);
-		gr_def(ctx, offset + 0x2cc, 0x00000001);
-		gr_def(ctx, offset + 0x2d0, 0x00000070);
-		gr_def(ctx, offset + 0x2d4, 0x00000080);
-	} else {
-		gr_def(ctx, offset + 0x2b8, 0x00000001);
-		gr_def(ctx, offset + 0x2bc, 0x000000f0);
-		gr_def(ctx, offset + 0x2c0, 0x000000ff);
-		gr_def(ctx, offset + 0x2cc, 0x00000001);
-		gr_def(ctx, offset + 0x2d0, 0x000000f0);
-		gr_def(ctx, offset + 0x2d4, 0x000000ff);
-		gr_def(ctx, offset + 0x2dc, 0x00000009);
-		offset += 4;
-	}
-	gr_def(ctx, offset + 0x2e4, 0x00000001);
-	gr_def(ctx, offset + 0x2e8, 0x000000cf);
-	gr_def(ctx, offset + 0x2f0, 0x00000001);
-	gr_def(ctx, offset + 0x300, 0x000000cf);
-	gr_def(ctx, offset + 0x308, 0x00000002);
-	gr_def(ctx, offset + 0x310, 0x00000001);
-	gr_def(ctx, offset + 0x318, 0x00000001);
-	gr_def(ctx, offset + 0x320, 0x000000cf);
-	gr_def(ctx, offset + 0x324, 0x000000cf);
-	gr_def(ctx, offset + 0x328, 0x00000001);
-
 	/* 6000? */
 	if (dev_priv->chipset == 0x50)
 		cp_ctx(ctx, 0x4063e0, 0x1);
@@ -661,7 +470,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 			gr_def(ctx, 0x406818, 0x00000f80);
 		else
 			gr_def(ctx, 0x406818, 0x00001f80);
-		if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
+		if (IS_NVA3F(dev_priv->chipset))
 			gr_def(ctx, 0x40681c, 0x00000030);
 		cp_ctx(ctx, 0x406830, 0x3);
 	}
@@ -706,7 +515,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 
 			if (dev_priv->chipset < 0xa0)
 				cp_ctx(ctx, 0x407094 + (i<<8), 1);
-			else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa)
+			else if (!IS_NVA3F(dev_priv->chipset))
 				cp_ctx(ctx, 0x407094 + (i<<8), 3);
 			else {
 				cp_ctx(ctx, 0x407094 + (i<<8), 4);
@@ -799,6 +608,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 				case 0xa8:
 				case 0xaa:
 				case 0xac:
+				case 0xaf:
 					gr_def(ctx, offset + 0x1c, 0x300c0000);
 					break;
 				}
@@ -825,7 +635,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 				gr_def(ctx, base + 0x304, 0x00007070);
 			else if (dev_priv->chipset < 0xa0)
 				gr_def(ctx, base + 0x304, 0x00027070);
-			else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa)
+			else if (!IS_NVA3F(dev_priv->chipset))
 				gr_def(ctx, base + 0x304, 0x01127070);
 			else
 				gr_def(ctx, base + 0x304, 0x05127070);
@@ -849,7 +659,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 			if (dev_priv->chipset < 0xa0) {
 				cp_ctx(ctx, base + 0x340, 9);
 				offset = base + 0x340;
-			} else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) {
+			} else if (!IS_NVA3F(dev_priv->chipset)) {
 				cp_ctx(ctx, base + 0x33c, 0xb);
 				offset = base + 0x344;
 			} else {
@@ -880,7 +690,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 			gr_def(ctx, offset + 0x0, 0x000001f0);
 			gr_def(ctx, offset + 0x4, 0x00000001);
 			gr_def(ctx, offset + 0x8, 0x00000003);
-			if (dev_priv->chipset == 0x50 || dev_priv->chipset >= 0xaa)
+			if (dev_priv->chipset == 0x50 || IS_NVAAF(dev_priv->chipset))
 				gr_def(ctx, offset + 0xc, 0x00008000);
 			gr_def(ctx, offset + 0x14, 0x00039e00);
 			cp_ctx(ctx, offset + 0x1c, 2);
@@ -892,7 +702,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 
 			if (dev_priv->chipset >= 0xa0) {
 				cp_ctx(ctx, base + 0x54c, 2);
-				if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa)
+				if (!IS_NVA3F(dev_priv->chipset))
 					gr_def(ctx, base + 0x54c, 0x003fe006);
 				else
 					gr_def(ctx, base + 0x54c, 0x003fe007);
@@ -948,6 +758,336 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 	}
 }
 
+static void
+dd_emit(struct nouveau_grctx *ctx, int num, uint32_t val) {
+	int i;
+	if (val && ctx->mode == NOUVEAU_GRCTX_VALS)
+		for (i = 0; i < num; i++)
+			nv_wo32(ctx->data, 4 * (ctx->ctxvals_pos + i), val);
+	ctx->ctxvals_pos += num;
+}
+
+static void
+nv50_graph_construct_mmio_ddata(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	int base, num;
+	base = ctx->ctxvals_pos;
+
+	/* tesla state */
+	dd_emit(ctx, 1, 0);	/* 00000001 UNK0F90 */
+	dd_emit(ctx, 1, 0);	/* 00000001 UNK135C */
+
+	/* SRC_TIC state */
+	dd_emit(ctx, 1, 0);	/* 00000007 SRC_TILE_MODE_Z */
+	dd_emit(ctx, 1, 2);	/* 00000007 SRC_TILE_MODE_Y */
+	dd_emit(ctx, 1, 1);	/* 00000001 SRC_LINEAR #1 */
+	dd_emit(ctx, 1, 0);	/* 000000ff SRC_ADDRESS_HIGH */
+	dd_emit(ctx, 1, 0);	/* 00000001 SRC_SRGB */
+	if (dev_priv->chipset >= 0x94)
+		dd_emit(ctx, 1, 0);	/* 00000003 eng2d UNK0258 */
+	dd_emit(ctx, 1, 1);	/* 00000fff SRC_DEPTH */
+	dd_emit(ctx, 1, 0x100);	/* 0000ffff SRC_HEIGHT */
+
+	/* turing state */
+	dd_emit(ctx, 1, 0);		/* 0000000f TEXTURES_LOG2 */
+	dd_emit(ctx, 1, 0);		/* 0000000f SAMPLERS_LOG2 */
+	dd_emit(ctx, 1, 0);		/* 000000ff CB_DEF_ADDRESS_HIGH */
+	dd_emit(ctx, 1, 0);		/* ffffffff CB_DEF_ADDRESS_LOW */
+	dd_emit(ctx, 1, 0);		/* ffffffff SHARED_SIZE */
+	dd_emit(ctx, 1, 2);		/* ffffffff REG_MODE */
+	dd_emit(ctx, 1, 1);		/* 0000ffff BLOCK_ALLOC_THREADS */
+	dd_emit(ctx, 1, 1);		/* 00000001 LANES32 */
+	dd_emit(ctx, 1, 0);		/* 000000ff UNK370 */
+	dd_emit(ctx, 1, 0);		/* 000000ff USER_PARAM_UNK */
+	dd_emit(ctx, 1, 0);		/* 000000ff USER_PARAM_COUNT */
+	dd_emit(ctx, 1, 1);		/* 000000ff UNK384 bits 8-15 */
+	dd_emit(ctx, 1, 0x3fffff);	/* 003fffff TIC_LIMIT */
+	dd_emit(ctx, 1, 0x1fff);	/* 000fffff TSC_LIMIT */
+	dd_emit(ctx, 1, 0);		/* 0000ffff CB_ADDR_INDEX */
+	dd_emit(ctx, 1, 1);		/* 000007ff BLOCKDIM_X */
+	dd_emit(ctx, 1, 1);		/* 000007ff BLOCKDIM_XMY */
+	dd_emit(ctx, 1, 0);		/* 00000001 BLOCKDIM_XMY_OVERFLOW */
+	dd_emit(ctx, 1, 1);		/* 0003ffff BLOCKDIM_XMYMZ */
+	dd_emit(ctx, 1, 1);		/* 000007ff BLOCKDIM_Y */
+	dd_emit(ctx, 1, 1);		/* 0000007f BLOCKDIM_Z */
+	dd_emit(ctx, 1, 4);		/* 000000ff CP_REG_ALLOC_TEMP */
+	dd_emit(ctx, 1, 1);		/* 00000001 BLOCKDIM_DIRTY */
+	if (IS_NVA3F(dev_priv->chipset))
+		dd_emit(ctx, 1, 0);	/* 00000003 UNK03E8 */
+	dd_emit(ctx, 1, 1);		/* 0000007f BLOCK_ALLOC_HALFWARPS */
+	dd_emit(ctx, 1, 1);		/* 00000007 LOCAL_WARPS_NO_CLAMP */
+	dd_emit(ctx, 1, 7);		/* 00000007 LOCAL_WARPS_LOG_ALLOC */
+	dd_emit(ctx, 1, 1);		/* 00000007 STACK_WARPS_NO_CLAMP */
+	dd_emit(ctx, 1, 7);		/* 00000007 STACK_WARPS_LOG_ALLOC */
+	dd_emit(ctx, 1, 1);		/* 00001fff BLOCK_ALLOC_REGSLOTS_PACKED */
+	dd_emit(ctx, 1, 1);		/* 00001fff BLOCK_ALLOC_REGSLOTS_STRIDED */
+	dd_emit(ctx, 1, 1);		/* 000007ff BLOCK_ALLOC_THREADS */
+
+	/* compat 2d state */
+	if (dev_priv->chipset == 0x50) {
+		dd_emit(ctx, 4, 0);		/* 0000ffff clip X, Y, W, H */
+
+		dd_emit(ctx, 1, 1);		/* ffffffff chroma COLOR_FORMAT */
+
+		dd_emit(ctx, 1, 1);		/* ffffffff pattern COLOR_FORMAT */
+		dd_emit(ctx, 1, 0);		/* ffffffff pattern SHAPE */
+		dd_emit(ctx, 1, 1);		/* ffffffff pattern PATTERN_SELECT */
+
+		dd_emit(ctx, 1, 0xa);		/* ffffffff surf2d SRC_FORMAT */
+		dd_emit(ctx, 1, 0);		/* ffffffff surf2d DMA_SRC */
+		dd_emit(ctx, 1, 0);		/* 000000ff surf2d SRC_ADDRESS_HIGH */
+		dd_emit(ctx, 1, 0);		/* ffffffff surf2d SRC_ADDRESS_LOW */
+		dd_emit(ctx, 1, 0x40);		/* 0000ffff surf2d SRC_PITCH */
+		dd_emit(ctx, 1, 0);		/* 0000000f surf2d SRC_TILE_MODE_Z */
+		dd_emit(ctx, 1, 2);		/* 0000000f surf2d SRC_TILE_MODE_Y */
+		dd_emit(ctx, 1, 0x100);		/* ffffffff surf2d SRC_HEIGHT */
+		dd_emit(ctx, 1, 1);		/* 00000001 surf2d SRC_LINEAR */
+		dd_emit(ctx, 1, 0x100);		/* ffffffff surf2d SRC_WIDTH */
+
+		dd_emit(ctx, 1, 0);		/* 0000ffff gdirect CLIP_B_X */
+		dd_emit(ctx, 1, 0);		/* 0000ffff gdirect CLIP_B_Y */
+		dd_emit(ctx, 1, 0);		/* 0000ffff gdirect CLIP_C_X */
+		dd_emit(ctx, 1, 0);		/* 0000ffff gdirect CLIP_C_Y */
+		dd_emit(ctx, 1, 0);		/* 0000ffff gdirect CLIP_D_X */
+		dd_emit(ctx, 1, 0);		/* 0000ffff gdirect CLIP_D_Y */
+		dd_emit(ctx, 1, 1);		/* ffffffff gdirect COLOR_FORMAT */
+		dd_emit(ctx, 1, 0);		/* ffffffff gdirect OPERATION */
+		dd_emit(ctx, 1, 0);		/* 0000ffff gdirect POINT_X */
+		dd_emit(ctx, 1, 0);		/* 0000ffff gdirect POINT_Y */
+
+		dd_emit(ctx, 1, 0);		/* 0000ffff blit SRC_Y */
+		dd_emit(ctx, 1, 0);		/* ffffffff blit OPERATION */
+
+		dd_emit(ctx, 1, 0);		/* ffffffff ifc OPERATION */
+
+		dd_emit(ctx, 1, 0);		/* ffffffff iifc INDEX_FORMAT */
+		dd_emit(ctx, 1, 0);		/* ffffffff iifc LUT_OFFSET */
+		dd_emit(ctx, 1, 4);		/* ffffffff iifc COLOR_FORMAT */
+		dd_emit(ctx, 1, 0);		/* ffffffff iifc OPERATION */
+	}
+
+	/* m2mf state */
+	dd_emit(ctx, 1, 0);		/* ffffffff m2mf LINE_COUNT */
+	dd_emit(ctx, 1, 0);		/* ffffffff m2mf LINE_LENGTH_IN */
+	dd_emit(ctx, 2, 0);		/* ffffffff m2mf OFFSET_IN, OFFSET_OUT */
+	dd_emit(ctx, 1, 1);		/* ffffffff m2mf TILING_DEPTH_OUT */
+	dd_emit(ctx, 1, 0x100);		/* ffffffff m2mf TILING_HEIGHT_OUT */
+	dd_emit(ctx, 1, 0);		/* ffffffff m2mf TILING_POSITION_OUT_Z */
+	dd_emit(ctx, 1, 1);		/* 00000001 m2mf LINEAR_OUT */
+	dd_emit(ctx, 2, 0);		/* 0000ffff m2mf TILING_POSITION_OUT_X, Y */
+	dd_emit(ctx, 1, 0x100);		/* ffffffff m2mf TILING_PITCH_OUT */
+	dd_emit(ctx, 1, 1);		/* ffffffff m2mf TILING_DEPTH_IN */
+	dd_emit(ctx, 1, 0x100);		/* ffffffff m2mf TILING_HEIGHT_IN */
+	dd_emit(ctx, 1, 0);		/* ffffffff m2mf TILING_POSITION_IN_Z */
+	dd_emit(ctx, 1, 1);		/* 00000001 m2mf LINEAR_IN */
+	dd_emit(ctx, 2, 0);		/* 0000ffff m2mf TILING_POSITION_IN_X, Y */
+	dd_emit(ctx, 1, 0x100);		/* ffffffff m2mf TILING_PITCH_IN */
+
+	/* more compat 2d state */
+	if (dev_priv->chipset == 0x50) {
+		dd_emit(ctx, 1, 1);		/* ffffffff line COLOR_FORMAT */
+		dd_emit(ctx, 1, 0);		/* ffffffff line OPERATION */
+
+		dd_emit(ctx, 1, 1);		/* ffffffff triangle COLOR_FORMAT */
+		dd_emit(ctx, 1, 0);		/* ffffffff triangle OPERATION */
+
+		dd_emit(ctx, 1, 0);		/* 0000000f sifm TILE_MODE_Z */
+		dd_emit(ctx, 1, 2);		/* 0000000f sifm TILE_MODE_Y */
+		dd_emit(ctx, 1, 0);		/* 000000ff sifm FORMAT_FILTER */
+		dd_emit(ctx, 1, 1);		/* 000000ff sifm FORMAT_ORIGIN */
+		dd_emit(ctx, 1, 0);		/* 0000ffff sifm SRC_PITCH */
+		dd_emit(ctx, 1, 1);		/* 00000001 sifm SRC_LINEAR */
+		dd_emit(ctx, 1, 0);		/* 000000ff sifm SRC_OFFSET_HIGH */
+		dd_emit(ctx, 1, 0);		/* ffffffff sifm SRC_OFFSET */
+		dd_emit(ctx, 1, 0);		/* 0000ffff sifm SRC_HEIGHT */
+		dd_emit(ctx, 1, 0);		/* 0000ffff sifm SRC_WIDTH */
+		dd_emit(ctx, 1, 3);		/* ffffffff sifm COLOR_FORMAT */
+		dd_emit(ctx, 1, 0);		/* ffffffff sifm OPERATION */
+
+		dd_emit(ctx, 1, 0);		/* ffffffff sifc OPERATION */
+	}
+
+	/* tesla state */
+	dd_emit(ctx, 1, 0);		/* 0000000f GP_TEXTURES_LOG2 */
+	dd_emit(ctx, 1, 0);		/* 0000000f GP_SAMPLERS_LOG2 */
+	dd_emit(ctx, 1, 0);		/* 000000ff */
+	dd_emit(ctx, 1, 0);		/* ffffffff */
+	dd_emit(ctx, 1, 4);		/* 000000ff UNK12B0_0 */
+	dd_emit(ctx, 1, 0x70);		/* 000000ff UNK12B0_1 */
+	dd_emit(ctx, 1, 0x80);		/* 000000ff UNK12B0_3 */
+	dd_emit(ctx, 1, 0);		/* 000000ff UNK12B0_2 */
+	dd_emit(ctx, 1, 0);		/* 0000000f FP_TEXTURES_LOG2 */
+	dd_emit(ctx, 1, 0);		/* 0000000f FP_SAMPLERS_LOG2 */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		dd_emit(ctx, 1, 0);	/* ffffffff */
+		dd_emit(ctx, 1, 0);	/* 0000007f MULTISAMPLE_SAMPLES_LOG2 */
+	} else {
+		dd_emit(ctx, 1, 0);	/* 0000000f MULTISAMPLE_SAMPLES_LOG2 */
+	} 
+	dd_emit(ctx, 1, 0xc);		/* 000000ff SEMANTIC_COLOR.BFC0_ID */
+	if (dev_priv->chipset != 0x50)
+		dd_emit(ctx, 1, 0);	/* 00000001 SEMANTIC_COLOR.CLMP_EN */
+	dd_emit(ctx, 1, 8);		/* 000000ff SEMANTIC_COLOR.COLR_NR */
+	dd_emit(ctx, 1, 0x14);		/* 000000ff SEMANTIC_COLOR.FFC0_ID */
+	if (dev_priv->chipset == 0x50) {
+		dd_emit(ctx, 1, 0);	/* 000000ff SEMANTIC_LAYER */
+		dd_emit(ctx, 1, 0);	/* 00000001 */
+	} else {
+		dd_emit(ctx, 1, 0);	/* 00000001 SEMANTIC_PTSZ.ENABLE */
+		dd_emit(ctx, 1, 0x29);	/* 000000ff SEMANTIC_PTSZ.PTSZ_ID */
+		dd_emit(ctx, 1, 0x27);	/* 000000ff SEMANTIC_PRIM */
+		dd_emit(ctx, 1, 0x26);	/* 000000ff SEMANTIC_LAYER */
+		dd_emit(ctx, 1, 8);	/* 0000000f SMENATIC_CLIP.CLIP_HIGH */
+		dd_emit(ctx, 1, 4);	/* 000000ff SEMANTIC_CLIP.CLIP_LO */
+		dd_emit(ctx, 1, 0x27);	/* 000000ff UNK0FD4 */
+		dd_emit(ctx, 1, 0);	/* 00000001 UNK1900 */
+	}
+	dd_emit(ctx, 1, 0);		/* 00000007 RT_CONTROL_MAP0 */
+	dd_emit(ctx, 1, 1);		/* 00000007 RT_CONTROL_MAP1 */
+	dd_emit(ctx, 1, 2);		/* 00000007 RT_CONTROL_MAP2 */
+	dd_emit(ctx, 1, 3);		/* 00000007 RT_CONTROL_MAP3 */
+	dd_emit(ctx, 1, 4);		/* 00000007 RT_CONTROL_MAP4 */
+	dd_emit(ctx, 1, 5);		/* 00000007 RT_CONTROL_MAP5 */
+	dd_emit(ctx, 1, 6);		/* 00000007 RT_CONTROL_MAP6 */
+	dd_emit(ctx, 1, 7);		/* 00000007 RT_CONTROL_MAP7 */
+	dd_emit(ctx, 1, 1);		/* 0000000f RT_CONTROL_COUNT */
+	dd_emit(ctx, 8, 0);		/* 00000001 RT_HORIZ_UNK */
+	dd_emit(ctx, 8, 0);		/* ffffffff RT_ADDRESS_LOW */
+	dd_emit(ctx, 1, 0xcf);		/* 000000ff RT_FORMAT */
+	dd_emit(ctx, 7, 0);		/* 000000ff RT_FORMAT */
+	if (dev_priv->chipset != 0x50)
+		dd_emit(ctx, 3, 0);	/* 1, 1, 1 */
+	else
+		dd_emit(ctx, 2, 0);	/* 1, 1 */
+	dd_emit(ctx, 1, 0);		/* ffffffff GP_ENABLE */
+	dd_emit(ctx, 1, 0x80);		/* 0000ffff GP_VERTEX_OUTPUT_COUNT*/
+	dd_emit(ctx, 1, 4);		/* 000000ff GP_REG_ALLOC_RESULT */
+	dd_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		dd_emit(ctx, 1, 3);	/* 00000003 */
+		dd_emit(ctx, 1, 0);	/* 00000001 UNK1418. Alone. */
+	}
+	if (dev_priv->chipset != 0x50)
+		dd_emit(ctx, 1, 3);	/* 00000003 UNK15AC */
+	dd_emit(ctx, 1, 1);		/* ffffffff RASTERIZE_ENABLE */
+	dd_emit(ctx, 1, 0);		/* 00000001 FP_CONTROL.EXPORTS_Z */
+	if (dev_priv->chipset != 0x50)
+		dd_emit(ctx, 1, 0);	/* 00000001 FP_CONTROL.MULTIPLE_RESULTS */
+	dd_emit(ctx, 1, 0x12);		/* 000000ff FP_INTERPOLANT_CTRL.COUNT */
+	dd_emit(ctx, 1, 0x10);		/* 000000ff FP_INTERPOLANT_CTRL.COUNT_NONFLAT */
+	dd_emit(ctx, 1, 0xc);		/* 000000ff FP_INTERPOLANT_CTRL.OFFSET */
+	dd_emit(ctx, 1, 1);		/* 00000001 FP_INTERPOLANT_CTRL.UMASK.W */
+	dd_emit(ctx, 1, 0);		/* 00000001 FP_INTERPOLANT_CTRL.UMASK.X */
+	dd_emit(ctx, 1, 0);		/* 00000001 FP_INTERPOLANT_CTRL.UMASK.Y */
+	dd_emit(ctx, 1, 0);		/* 00000001 FP_INTERPOLANT_CTRL.UMASK.Z */
+	dd_emit(ctx, 1, 4);		/* 000000ff FP_RESULT_COUNT */
+	dd_emit(ctx, 1, 2);		/* ffffffff REG_MODE */
+	dd_emit(ctx, 1, 4);		/* 000000ff FP_REG_ALLOC_TEMP */
+	if (dev_priv->chipset >= 0xa0)
+		dd_emit(ctx, 1, 0);	/* ffffffff */
+	dd_emit(ctx, 1, 0);		/* 00000001 GP_BUILTIN_RESULT_EN.LAYER_IDX */
+	dd_emit(ctx, 1, 0);		/* ffffffff STRMOUT_ENABLE */
+	dd_emit(ctx, 1, 0x3fffff);	/* 003fffff TIC_LIMIT */
+	dd_emit(ctx, 1, 0x1fff);	/* 000fffff TSC_LIMIT */
+	dd_emit(ctx, 1, 0);		/* 00000001 VERTEX_TWO_SIDE_ENABLE*/
+	if (dev_priv->chipset != 0x50)
+		dd_emit(ctx, 8, 0);	/* 00000001 */
+	if (dev_priv->chipset >= 0xa0) {
+		dd_emit(ctx, 1, 1);	/* 00000007 VTX_ATTR_DEFINE.COMP */
+		dd_emit(ctx, 1, 1);	/* 00000007 VTX_ATTR_DEFINE.SIZE */
+		dd_emit(ctx, 1, 2);	/* 00000007 VTX_ATTR_DEFINE.TYPE */
+		dd_emit(ctx, 1, 0);	/* 000000ff VTX_ATTR_DEFINE.ATTR */
+	}
+	dd_emit(ctx, 1, 4);		/* 0000007f VP_RESULT_MAP_SIZE */
+	dd_emit(ctx, 1, 0x14);		/* 0000001f ZETA_FORMAT */
+	dd_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	dd_emit(ctx, 1, 0);		/* 0000000f VP_TEXTURES_LOG2 */
+	dd_emit(ctx, 1, 0);		/* 0000000f VP_SAMPLERS_LOG2 */
+	if (IS_NVA3F(dev_priv->chipset))
+		dd_emit(ctx, 1, 0);	/* 00000001 */
+	dd_emit(ctx, 1, 2);		/* 00000003 POLYGON_MODE_BACK */
+	if (dev_priv->chipset >= 0xa0)
+		dd_emit(ctx, 1, 0);	/* 00000003 VTX_ATTR_DEFINE.SIZE - 1 */
+	dd_emit(ctx, 1, 0);		/* 0000ffff CB_ADDR_INDEX */
+	if (dev_priv->chipset >= 0xa0)
+		dd_emit(ctx, 1, 0);	/* 00000003 */
+	dd_emit(ctx, 1, 0);		/* 00000001 CULL_FACE_ENABLE */
+	dd_emit(ctx, 1, 1);		/* 00000003 CULL_FACE */
+	dd_emit(ctx, 1, 0);		/* 00000001 FRONT_FACE */
+	dd_emit(ctx, 1, 2);		/* 00000003 POLYGON_MODE_FRONT */
+	dd_emit(ctx, 1, 0x1000);	/* 00007fff UNK141C */
+	if (dev_priv->chipset != 0x50) {
+		dd_emit(ctx, 1, 0xe00);		/* 7fff */
+		dd_emit(ctx, 1, 0x1000);	/* 7fff */
+		dd_emit(ctx, 1, 0x1e00);	/* 7fff */
+	}
+	dd_emit(ctx, 1, 0);		/* 00000001 BEGIN_END_ACTIVE */
+	dd_emit(ctx, 1, 1);		/* 00000001 POLYGON_MODE_??? */
+	dd_emit(ctx, 1, 1);		/* 000000ff GP_REG_ALLOC_TEMP / 4 rounded up */
+	dd_emit(ctx, 1, 1);		/* 000000ff FP_REG_ALLOC_TEMP... without /4? */
+	dd_emit(ctx, 1, 1);		/* 000000ff VP_REG_ALLOC_TEMP / 4 rounded up */
+	dd_emit(ctx, 1, 1);		/* 00000001 */
+	dd_emit(ctx, 1, 0);		/* 00000001 */
+	dd_emit(ctx, 1, 0);		/* 00000001 VTX_ATTR_MASK_UNK0 nonempty */
+	dd_emit(ctx, 1, 0);		/* 00000001 VTX_ATTR_MASK_UNK1 nonempty */
+	dd_emit(ctx, 1, 0x200);		/* 0003ffff GP_VERTEX_OUTPUT_COUNT*GP_REG_ALLOC_RESULT */
+	if (IS_NVA3F(dev_priv->chipset))
+		dd_emit(ctx, 1, 0x200);
+	dd_emit(ctx, 1, 0);		/* 00000001 */
+	if (dev_priv->chipset < 0xa0) {
+		dd_emit(ctx, 1, 1);	/* 00000001 */
+		dd_emit(ctx, 1, 0x70);	/* 000000ff */
+		dd_emit(ctx, 1, 0x80);	/* 000000ff */
+		dd_emit(ctx, 1, 0);	/* 000000ff */
+		dd_emit(ctx, 1, 0);	/* 00000001 */
+		dd_emit(ctx, 1, 1);	/* 00000001 */
+		dd_emit(ctx, 1, 0x70);	/* 000000ff */
+		dd_emit(ctx, 1, 0x80);	/* 000000ff */
+		dd_emit(ctx, 1, 0);	/* 000000ff */
+	} else {
+		dd_emit(ctx, 1, 1);	/* 00000001 */
+		dd_emit(ctx, 1, 0xf0);	/* 000000ff */
+		dd_emit(ctx, 1, 0xff);	/* 000000ff */
+		dd_emit(ctx, 1, 0);	/* 000000ff */
+		dd_emit(ctx, 1, 0);	/* 00000001 */
+		dd_emit(ctx, 1, 1);	/* 00000001 */
+		dd_emit(ctx, 1, 0xf0);	/* 000000ff */
+		dd_emit(ctx, 1, 0xff);	/* 000000ff */
+		dd_emit(ctx, 1, 0);	/* 000000ff */
+		dd_emit(ctx, 1, 9);	/* 0000003f UNK114C.COMP,SIZE */
+	}
+
+	/* eng2d state */
+	dd_emit(ctx, 1, 0);		/* 00000001 eng2d COLOR_KEY_ENABLE */
+	dd_emit(ctx, 1, 0);		/* 00000007 eng2d COLOR_KEY_FORMAT */
+	dd_emit(ctx, 1, 1);		/* ffffffff eng2d DST_DEPTH */
+	dd_emit(ctx, 1, 0xcf);		/* 000000ff eng2d DST_FORMAT */
+	dd_emit(ctx, 1, 0);		/* ffffffff eng2d DST_LAYER */
+	dd_emit(ctx, 1, 1);		/* 00000001 eng2d DST_LINEAR */
+	dd_emit(ctx, 1, 0);		/* 00000007 eng2d PATTERN_COLOR_FORMAT */
+	dd_emit(ctx, 1, 0);		/* 00000007 eng2d OPERATION */
+	dd_emit(ctx, 1, 0);		/* 00000003 eng2d PATTERN_SELECT */
+	dd_emit(ctx, 1, 0xcf);		/* 000000ff eng2d SIFC_FORMAT */
+	dd_emit(ctx, 1, 0);		/* 00000001 eng2d SIFC_BITMAP_ENABLE */
+	dd_emit(ctx, 1, 2);		/* 00000003 eng2d SIFC_BITMAP_UNK808 */
+	dd_emit(ctx, 1, 0);		/* ffffffff eng2d BLIT_DU_DX_FRACT */
+	dd_emit(ctx, 1, 1);		/* ffffffff eng2d BLIT_DU_DX_INT */
+	dd_emit(ctx, 1, 0);		/* ffffffff eng2d BLIT_DV_DY_FRACT */
+	dd_emit(ctx, 1, 1);		/* ffffffff eng2d BLIT_DV_DY_INT */
+	dd_emit(ctx, 1, 0);		/* 00000001 eng2d BLIT_CONTROL_FILTER */
+	dd_emit(ctx, 1, 0xcf);		/* 000000ff eng2d DRAW_COLOR_FORMAT */
+	dd_emit(ctx, 1, 0xcf);		/* 000000ff eng2d SRC_FORMAT */
+	dd_emit(ctx, 1, 1);		/* 00000001 eng2d SRC_LINEAR #2 */
+
+	num = ctx->ctxvals_pos - base;
+	ctx->ctxvals_pos = base;
+	if (IS_NVA3F(dev_priv->chipset))
+		cp_ctx(ctx, 0x404800, num);
+	else
+		cp_ctx(ctx, 0x405400, num);
+}
+
 /*
  * xfer areas. These are a pain.
  *
@@ -990,28 +1130,33 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
  * without the help of ctxprog.
  */
 
-static inline void
+static void
 xf_emit(struct nouveau_grctx *ctx, int num, uint32_t val) {
 	int i;
 	if (val && ctx->mode == NOUVEAU_GRCTX_VALS)
 		for (i = 0; i < num; i++)
-			nv_wo32(ctx->dev, ctx->data, ctx->ctxvals_pos + (i << 3), val);
+			nv_wo32(ctx->data, 4 * (ctx->ctxvals_pos + (i << 3)), val);
 	ctx->ctxvals_pos += num << 3;
 }
 
 /* Gene declarations... */
 
+static void nv50_graph_construct_gene_dispatch(struct nouveau_grctx *ctx);
 static void nv50_graph_construct_gene_m2mf(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk1(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk2(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk3(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk4(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk5(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk6(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk7(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk8(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk9(struct nouveau_grctx *ctx);
-static void nv50_graph_construct_gene_unk10(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_ccache(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_unk10xx(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_unk14xx(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_zcull(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_clipid(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_unk24xx(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_vfetch(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_eng2d(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_csched(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_unk1cxx(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_strmout(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_unk34xx(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_ropm1(struct nouveau_grctx *ctx);
+static void nv50_graph_construct_gene_ropm2(struct nouveau_grctx *ctx);
 static void nv50_graph_construct_gene_ropc(struct nouveau_grctx *ctx);
 static void nv50_graph_construct_xfer_tp(struct nouveau_grctx *ctx);
 
@@ -1030,102 +1175,32 @@ nv50_graph_construct_xfer1(struct nouveau_grctx *ctx)
 	if (dev_priv->chipset < 0xa0) {
 		/* Strand 0 */
 		ctx->ctxvals_pos = offset;
-		switch (dev_priv->chipset) {
-		case 0x50:
-			xf_emit(ctx, 0x99, 0);
-			break;
-		case 0x84:
-		case 0x86:
-			xf_emit(ctx, 0x384, 0);
-			break;
-		case 0x92:
-		case 0x94:
-		case 0x96:
-		case 0x98:
-			xf_emit(ctx, 0x380, 0);
-			break;
-		}
-		nv50_graph_construct_gene_m2mf (ctx);
-		switch (dev_priv->chipset) {
-		case 0x50:
-		case 0x84:
-		case 0x86:
-		case 0x98:
-			xf_emit(ctx, 0x4c4, 0);
-			break;
-		case 0x92:
-		case 0x94:
-		case 0x96:
-			xf_emit(ctx, 0x984, 0);
-			break;
-		}
-		nv50_graph_construct_gene_unk5(ctx);
-		if (dev_priv->chipset == 0x50)
-			xf_emit(ctx, 0xa, 0);
-		else
-			xf_emit(ctx, 0xb, 0);
-		nv50_graph_construct_gene_unk4(ctx);
-		nv50_graph_construct_gene_unk3(ctx);
+		nv50_graph_construct_gene_dispatch(ctx);
+		nv50_graph_construct_gene_m2mf(ctx);
+		nv50_graph_construct_gene_unk24xx(ctx);
+		nv50_graph_construct_gene_clipid(ctx);
+		nv50_graph_construct_gene_zcull(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
 		/* Strand 1 */
 		ctx->ctxvals_pos = offset + 0x1;
-		nv50_graph_construct_gene_unk6(ctx);
-		nv50_graph_construct_gene_unk7(ctx);
-		nv50_graph_construct_gene_unk8(ctx);
-		switch (dev_priv->chipset) {
-		case 0x50:
-		case 0x92:
-			xf_emit(ctx, 0xfb, 0);
-			break;
-		case 0x84:
-			xf_emit(ctx, 0xd3, 0);
-			break;
-		case 0x94:
-		case 0x96:
-			xf_emit(ctx, 0xab, 0);
-			break;
-		case 0x86:
-		case 0x98:
-			xf_emit(ctx, 0x6b, 0);
-			break;
-		}
-		xf_emit(ctx, 2, 0x4e3bfdf);
-		xf_emit(ctx, 4, 0);
-		xf_emit(ctx, 1, 0x0fac6881);
-		xf_emit(ctx, 0xb, 0);
-		xf_emit(ctx, 2, 0x4e3bfdf);
+		nv50_graph_construct_gene_vfetch(ctx);
+		nv50_graph_construct_gene_eng2d(ctx);
+		nv50_graph_construct_gene_csched(ctx);
+		nv50_graph_construct_gene_ropm1(ctx);
+		nv50_graph_construct_gene_ropm2(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
 		/* Strand 2 */
 		ctx->ctxvals_pos = offset + 0x2;
-		switch (dev_priv->chipset) {
-		case 0x50:
-		case 0x92:
-			xf_emit(ctx, 0xa80, 0);
-			break;
-		case 0x84:
-			xf_emit(ctx, 0xa7e, 0);
-			break;
-		case 0x94:
-		case 0x96:
-			xf_emit(ctx, 0xa7c, 0);
-			break;
-		case 0x86:
-		case 0x98:
-			xf_emit(ctx, 0xa7a, 0);
-			break;
-		}
-		xf_emit(ctx, 1, 0x3fffff);
-		xf_emit(ctx, 2, 0);
-		xf_emit(ctx, 1, 0x1fff);
-		xf_emit(ctx, 0xe, 0);
-		nv50_graph_construct_gene_unk9(ctx);
-		nv50_graph_construct_gene_unk2(ctx);
-		nv50_graph_construct_gene_unk1(ctx);
-		nv50_graph_construct_gene_unk10(ctx);
+		nv50_graph_construct_gene_ccache(ctx);
+		nv50_graph_construct_gene_unk1cxx(ctx);
+		nv50_graph_construct_gene_strmout(ctx);
+		nv50_graph_construct_gene_unk14xx(ctx);
+		nv50_graph_construct_gene_unk10xx(ctx);
+		nv50_graph_construct_gene_unk34xx(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
@@ -1150,86 +1225,46 @@ nv50_graph_construct_xfer1(struct nouveau_grctx *ctx)
 	} else {
 		/* Strand 0 */
 		ctx->ctxvals_pos = offset;
-		if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-			xf_emit(ctx, 0x385, 0);
-		else
-			xf_emit(ctx, 0x384, 0);
+		nv50_graph_construct_gene_dispatch(ctx);
 		nv50_graph_construct_gene_m2mf(ctx);
-		xf_emit(ctx, 0x950, 0);
-		nv50_graph_construct_gene_unk10(ctx);
-		xf_emit(ctx, 1, 0x0fac6881);
-		if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-			xf_emit(ctx, 1, 1);
-			xf_emit(ctx, 3, 0);
-		}
-		nv50_graph_construct_gene_unk8(ctx);
-		if (dev_priv->chipset == 0xa0)
-			xf_emit(ctx, 0x189, 0);
-		else if (dev_priv->chipset == 0xa3)
-			xf_emit(ctx, 0xd5, 0);
-		else if (dev_priv->chipset == 0xa5)
-			xf_emit(ctx, 0x99, 0);
-		else if (dev_priv->chipset == 0xaa)
-			xf_emit(ctx, 0x65, 0);
-		else
-			xf_emit(ctx, 0x6d, 0);
-		nv50_graph_construct_gene_unk9(ctx);
+		nv50_graph_construct_gene_unk34xx(ctx);
+		nv50_graph_construct_gene_csched(ctx);
+		nv50_graph_construct_gene_unk1cxx(ctx);
+		nv50_graph_construct_gene_strmout(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
 		/* Strand 1 */
 		ctx->ctxvals_pos = offset + 1;
-		nv50_graph_construct_gene_unk1(ctx);
+		nv50_graph_construct_gene_unk10xx(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
 		/* Strand 2 */
 		ctx->ctxvals_pos = offset + 2;
-		if (dev_priv->chipset == 0xa0) {
-			nv50_graph_construct_gene_unk2(ctx);
-		}
-		xf_emit(ctx, 0x36, 0);
-		nv50_graph_construct_gene_unk5(ctx);
+		if (dev_priv->chipset == 0xa0)
+			nv50_graph_construct_gene_unk14xx(ctx);
+		nv50_graph_construct_gene_unk24xx(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
 		/* Strand 3 */
 		ctx->ctxvals_pos = offset + 3;
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 1, 1);
-		nv50_graph_construct_gene_unk6(ctx);
+		nv50_graph_construct_gene_vfetch(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
 		/* Strand 4 */
 		ctx->ctxvals_pos = offset + 4;
-		if (dev_priv->chipset == 0xa0)
-			xf_emit(ctx, 0xa80, 0);
-		else if (dev_priv->chipset == 0xa3)
-			xf_emit(ctx, 0xa7c, 0);
-		else
-			xf_emit(ctx, 0xa7a, 0);
-		xf_emit(ctx, 1, 0x3fffff);
-		xf_emit(ctx, 2, 0);
-		xf_emit(ctx, 1, 0x1fff);
+		nv50_graph_construct_gene_ccache(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
 		/* Strand 5 */
 		ctx->ctxvals_pos = offset + 5;
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 1, 0x0fac6881);
-		xf_emit(ctx, 0xb, 0);
-		xf_emit(ctx, 2, 0x4e3bfdf);
-		xf_emit(ctx, 3, 0);
-		if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-			xf_emit(ctx, 1, 0x11);
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 2, 0x4e3bfdf);
-		xf_emit(ctx, 2, 0);
-		if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-			xf_emit(ctx, 1, 0x11);
-		xf_emit(ctx, 1, 0);
+		nv50_graph_construct_gene_ropm2(ctx);
+		nv50_graph_construct_gene_ropm1(ctx);
+		/* per-ROP context */
 		for (i = 0; i < 8; i++)
 			if (units & (1<<(i+16)))
 				nv50_graph_construct_gene_ropc(ctx);
@@ -1238,10 +1273,9 @@ nv50_graph_construct_xfer1(struct nouveau_grctx *ctx)
 
 		/* Strand 6 */
 		ctx->ctxvals_pos = offset + 6;
-		nv50_graph_construct_gene_unk3(ctx);
-		xf_emit(ctx, 0xb, 0);
-		nv50_graph_construct_gene_unk4(ctx);
-		nv50_graph_construct_gene_unk7(ctx);
+		nv50_graph_construct_gene_zcull(ctx);
+		nv50_graph_construct_gene_clipid(ctx);
+		nv50_graph_construct_gene_eng2d(ctx);
 		if (units & (1 << 0))
 			nv50_graph_construct_xfer_tp(ctx);
 		if (units & (1 << 1))
@@ -1269,7 +1303,7 @@ nv50_graph_construct_xfer1(struct nouveau_grctx *ctx)
 			if (units & (1 << 9))
 				nv50_graph_construct_xfer_tp(ctx);
 		} else {
-			nv50_graph_construct_gene_unk2(ctx);
+			nv50_graph_construct_gene_unk14xx(ctx);
 		}
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
@@ -1289,10 +1323,71 @@ nv50_graph_construct_xfer1(struct nouveau_grctx *ctx)
  * non-trivial demagiced parts of ctx init go here
  */
 
+static void
+nv50_graph_construct_gene_dispatch(struct nouveau_grctx *ctx)
+{
+	/* start of strand 0 */
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	/* SEEK */
+	if (dev_priv->chipset == 0x50)
+		xf_emit(ctx, 5, 0);
+	else if (!IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 6, 0);
+	else
+		xf_emit(ctx, 4, 0);
+	/* SEEK */
+	/* the PGRAPH's internal FIFO */
+	if (dev_priv->chipset == 0x50)
+		xf_emit(ctx, 8*3, 0);
+	else
+		xf_emit(ctx, 0x100*3, 0);
+	/* and another bonus slot?!? */
+	xf_emit(ctx, 3, 0);
+	/* and YET ANOTHER bonus slot? */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 3, 0);
+	/* SEEK */
+	/* CTX_SWITCH: caches of gr objects bound to subchannels. 8 values, last used index */
+	xf_emit(ctx, 9, 0);
+	/* SEEK */
+	xf_emit(ctx, 9, 0);
+	/* SEEK */
+	xf_emit(ctx, 9, 0);
+	/* SEEK */
+	xf_emit(ctx, 9, 0);
+	/* SEEK */
+	if (dev_priv->chipset < 0x90)
+		xf_emit(ctx, 4, 0);
+	/* SEEK */
+	xf_emit(ctx, 2, 0);
+	/* SEEK */
+	xf_emit(ctx, 6*2, 0);
+	xf_emit(ctx, 2, 0);
+	/* SEEK */
+	xf_emit(ctx, 2, 0);
+	/* SEEK */
+	xf_emit(ctx, 6*2, 0);
+	xf_emit(ctx, 2, 0);
+	/* SEEK */
+	if (dev_priv->chipset == 0x50)
+		xf_emit(ctx, 0x1c, 0);
+	else if (dev_priv->chipset < 0xa0)
+		xf_emit(ctx, 0x1e, 0);
+	else
+		xf_emit(ctx, 0x22, 0);
+	/* SEEK */
+	xf_emit(ctx, 0x15, 0);
+}
+
 static void
 nv50_graph_construct_gene_m2mf(struct nouveau_grctx *ctx)
 {
-	/* m2mf state */
+	/* Strand 0, right after dispatch */
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	int smallm2mf = 0;
+	if (dev_priv->chipset < 0x92 || dev_priv->chipset == 0x98)
+		smallm2mf = 1;
+	/* SEEK */
 	xf_emit (ctx, 1, 0);		/* DMA_NOTIFY instance >> 4 */
 	xf_emit (ctx, 1, 0);		/* DMA_BUFFER_IN instance >> 4 */
 	xf_emit (ctx, 1, 0);		/* DMA_BUFFER_OUT instance >> 4 */
@@ -1319,427 +1414,975 @@ nv50_graph_construct_gene_m2mf(struct nouveau_grctx *ctx)
 	xf_emit (ctx, 1, 0);		/* TILING_POSITION_OUT */
 	xf_emit (ctx, 1, 0);		/* OFFSET_IN_HIGH */
 	xf_emit (ctx, 1, 0);		/* OFFSET_OUT_HIGH */
+	/* SEEK */
+	if (smallm2mf)
+		xf_emit(ctx, 0x40, 0);	/* 20 * ffffffff, 3ffff */
+	else
+		xf_emit(ctx, 0x100, 0);	/* 80 * ffffffff, 3ffff */
+	xf_emit(ctx, 4, 0);		/* 1f/7f, 0, 1f/7f, 0 [1f for smallm2mf, 7f otherwise] */
+	/* SEEK */
+	if (smallm2mf)
+		xf_emit(ctx, 0x400, 0);	/* ffffffff */
+	else
+		xf_emit(ctx, 0x800, 0);	/* ffffffff */
+	xf_emit(ctx, 4, 0);		/* ff/1ff, 0, 0, 0 [ff for smallm2mf, 1ff otherwise] */
+	/* SEEK */
+	xf_emit(ctx, 0x40, 0);		/* 20 * bits ffffffff, 3ffff */
+	xf_emit(ctx, 0x6, 0);		/* 1f, 0, 1f, 0, 1f, 0 */
 }
 
 static void
-nv50_graph_construct_gene_unk1(struct nouveau_grctx *ctx)
+nv50_graph_construct_gene_ccache(struct nouveau_grctx *ctx)
 {
 	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	xf_emit(ctx, 2, 0);		/* RO */
+	xf_emit(ctx, 0x800, 0);		/* ffffffff */
+	switch (dev_priv->chipset) {
+	case 0x50:
+	case 0x92:
+	case 0xa0:
+		xf_emit(ctx, 0x2b, 0);
+		break;
+	case 0x84:
+		xf_emit(ctx, 0x29, 0);
+		break;
+	case 0x94:
+	case 0x96:
+	case 0xa3:
+		xf_emit(ctx, 0x27, 0);
+		break;
+	case 0x86:
+	case 0x98:
+	case 0xa5:
+	case 0xa8:
+	case 0xaa:
+	case 0xac:
+	case 0xaf:
+		xf_emit(ctx, 0x25, 0);
+		break;
+	}
+	/* CB bindings, 0x80 of them. first word is address >> 8, second is
+	 * size >> 4 | valid << 24 */
+	xf_emit(ctx, 0x100, 0);		/* ffffffff CB_DEF */
+	xf_emit(ctx, 1, 0);		/* 0000007f CB_ADDR_BUFFER */
+	xf_emit(ctx, 1, 0);		/* 0 */
+	xf_emit(ctx, 0x30, 0);		/* ff SET_PROGRAM_CB */
+	xf_emit(ctx, 1, 0);		/* 3f last SET_PROGRAM_CB */
+	xf_emit(ctx, 4, 0);		/* RO */
+	xf_emit(ctx, 0x100, 0);		/* ffffffff */
+	xf_emit(ctx, 8, 0);		/* 1f, 0, 0, ... */
+	xf_emit(ctx, 8, 0);		/* ffffffff */
+	xf_emit(ctx, 4, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 3 */
+	xf_emit(ctx, 1, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_CODE_CB */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_TIC */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_TSC */
+	xf_emit(ctx, 1, 0);		/* 00000001 LINKED_TSC */
+	xf_emit(ctx, 1, 0);		/* 000000ff TIC_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff TIC_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0x3fffff);	/* 003fffff TIC_LIMIT */
+	xf_emit(ctx, 1, 0);		/* 000000ff TSC_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff TSC_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0x1fff);	/* 000fffff TSC_LIMIT */
+	xf_emit(ctx, 1, 0);		/* 000000ff VP_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff VP_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0);		/* 00ffffff VP_START_ID */
+	xf_emit(ctx, 1, 0);		/* 000000ff CB_DEF_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff CB_DEF_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 000000ff GP_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff GP_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0);		/* 00ffffff GP_START_ID */
+	xf_emit(ctx, 1, 0);		/* 000000ff FP_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff FP_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0);		/* 00ffffff FP_START_ID */
+}
+
+static void
+nv50_graph_construct_gene_unk10xx(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	int i;
 	/* end of area 2 on pre-NVA0, area 1 on NVAx */
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x80);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0x80c14);
-	xf_emit(ctx, 1, 0);
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 4);		/* 0000007f VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0x80);		/* 0000ffff GP_VERTEX_OUTPUT_COUNT */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_REG_ALLOC_RESULT */
+	xf_emit(ctx, 1, 0x80c14);	/* 01ffffff SEMANTIC_COLOR */
+	xf_emit(ctx, 1, 0);		/* 00000001 VERTEX_TWO_SIDE_ENABLE */
 	if (dev_priv->chipset == 0x50)
 		xf_emit(ctx, 1, 0x3ff);
 	else
-		xf_emit(ctx, 1, 0x7ff);
-	switch (dev_priv->chipset) {
-	case 0x50:
-	case 0x86:
-	case 0x98:
-	case 0xaa:
-	case 0xac:
-		xf_emit(ctx, 0x542, 0);
-		break;
-	case 0x84:
-	case 0x92:
-	case 0x94:
-	case 0x96:
-		xf_emit(ctx, 0x942, 0);
-		break;
-	case 0xa0:
-	case 0xa3:
-		xf_emit(ctx, 0x2042, 0);
-		break;
-	case 0xa5:
-	case 0xa8:
-		xf_emit(ctx, 0x842, 0);
-		break;
+		xf_emit(ctx, 1, 0x7ff);	/* 000007ff */
+	xf_emit(ctx, 1, 0);		/* 111/113 */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	for (i = 0; i < 8; i++) {
+		switch (dev_priv->chipset) {
+		case 0x50:
+		case 0x86:
+		case 0x98:
+		case 0xaa:
+		case 0xac:
+			xf_emit(ctx, 0xa0, 0);	/* ffffffff */
+			break;
+		case 0x84:
+		case 0x92:
+		case 0x94:
+		case 0x96:
+			xf_emit(ctx, 0x120, 0);
+			break;
+		case 0xa5:
+		case 0xa8:
+			xf_emit(ctx, 0x100, 0);	/* ffffffff */
+			break;
+		case 0xa0:
+		case 0xa3:
+		case 0xaf:
+			xf_emit(ctx, 0x400, 0);	/* ffffffff */
+			break;
+		}
+		xf_emit(ctx, 4, 0);	/* 3f, 0, 0, 0 */
+		xf_emit(ctx, 4, 0);	/* ffffffff */
 	}
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x80);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x27);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x26);
-	xf_emit(ctx, 3, 0);
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 4);		/* 0000007f VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0x80);		/* 0000ffff GP_VERTEX_OUTPUT_COUNT */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_REG_ALLOC_TEMP */
+	xf_emit(ctx, 1, 1);		/* 00000001 RASTERIZE_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1900 */
+	xf_emit(ctx, 1, 0x27);		/* 000000ff UNK0FD4 */
+	xf_emit(ctx, 1, 0);		/* 0001ffff GP_BUILTIN_RESULT_EN */
+	xf_emit(ctx, 1, 0x26);		/* 000000ff SEMANTIC_LAYER */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
 }
 
 static void
-nv50_graph_construct_gene_unk10(struct nouveau_grctx *ctx)
+nv50_graph_construct_gene_unk34xx(struct nouveau_grctx *ctx)
 {
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
 	/* end of area 2 on pre-NVA0, area 1 on NVAx */
-	xf_emit(ctx, 0x10, 0x04000000);
-	xf_emit(ctx, 0x24, 0);
-	xf_emit(ctx, 2, 0x04e3bfdf);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x1fe21);
+	xf_emit(ctx, 1, 0);		/* 00000001 VIEWPORT_CLIP_RECTS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000003 VIEWPORT_CLIP_MODE */
+	xf_emit(ctx, 0x10, 0x04000000);	/* 07ffffff VIEWPORT_CLIP_HORIZ*8, VIEWPORT_CLIP_VERT*8 */
+	xf_emit(ctx, 1, 0);		/* 00000001 POLYGON_STIPPLE_ENABLE */
+	xf_emit(ctx, 0x20, 0);		/* ffffffff POLYGON_STIPPLE */
+	xf_emit(ctx, 2, 0);		/* 00007fff WINDOW_OFFSET_XY */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, 0x04e3bfdf);	/* ffffffff UNK0D64 */
+	xf_emit(ctx, 1, 0x04e3bfdf);	/* ffffffff UNK0DF4 */
+	xf_emit(ctx, 1, 0);		/* 00000003 WINDOW_ORIGIN */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	xf_emit(ctx, 1, 0x1fe21);	/* 0001ffff tesla UNK0FAC */
+	if (dev_priv->chipset >= 0xa0)
+		xf_emit(ctx, 1, 0x0fac6881);
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 1, 1);
+		xf_emit(ctx, 3, 0);
+	}
 }
 
 static void
-nv50_graph_construct_gene_unk2(struct nouveau_grctx *ctx)
+nv50_graph_construct_gene_unk14xx(struct nouveau_grctx *ctx)
 {
 	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
 	/* middle of area 2 on pre-NVA0, beginning of area 2 on NVA0, area 7 on >NVA0 */
 	if (dev_priv->chipset != 0x50) {
-		xf_emit(ctx, 5, 0);
-		xf_emit(ctx, 1, 0x80c14);
-		xf_emit(ctx, 2, 0);
-		xf_emit(ctx, 1, 0x804);
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 2, 4);
-		xf_emit(ctx, 1, 0x8100c12);
+		xf_emit(ctx, 5, 0);		/* ffffffff */
+		xf_emit(ctx, 1, 0x80c14);	/* 01ffffff SEMANTIC_COLOR */
+		xf_emit(ctx, 1, 0);		/* 00000001 */
+		xf_emit(ctx, 1, 0);		/* 000003ff */
+		xf_emit(ctx, 1, 0x804);		/* 00000fff SEMANTIC_CLIP */
+		xf_emit(ctx, 1, 0);		/* 00000001 */
+		xf_emit(ctx, 2, 4);		/* 7f, ff */
+		xf_emit(ctx, 1, 0x8100c12);	/* 1fffffff FP_INTERPOLANT_CTRL */
 	}
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x10);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 3, 0);
-	else
-		xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 0x804);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x1a);
+	xf_emit(ctx, 1, 0);			/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 4);			/* 0000007f VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 4);			/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);			/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0x10);			/* 7f/ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 1, 0);			/* 000000ff VP_CLIP_DISTANCE_ENABLE */
 	if (dev_priv->chipset != 0x50)
-		xf_emit(ctx, 1, 0x7f);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x80c14);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x8100c12);
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x10);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x8100c12);
-	xf_emit(ctx, 6, 0);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 1, 0x3ff);
-	else
-		xf_emit(ctx, 1, 0x7ff);
-	xf_emit(ctx, 1, 0x80c14);
-	xf_emit(ctx, 0x38, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x10);
-	xf_emit(ctx, 0x38, 0);
-	xf_emit(ctx, 2, 0x88);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 0x16, 0);
-	xf_emit(ctx, 1, 0x26);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x3f800000);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 4, 0);
-	else
-		xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x1a);
-	xf_emit(ctx, 1, 0x10);
+		xf_emit(ctx, 1, 0);		/* 3ff */
+	xf_emit(ctx, 1, 0);			/* 000000ff tesla UNK1940 */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK0D7C */
+	xf_emit(ctx, 1, 0x804);			/* 00000fff SEMANTIC_CLIP */
+	xf_emit(ctx, 1, 1);			/* 00000001 VIEWPORT_TRANSFORM_EN */
+	xf_emit(ctx, 1, 0x1a);			/* 0000001f POLYGON_MODE */
 	if (dev_priv->chipset != 0x50)
-		xf_emit(ctx, 0x28, 0);
-	else
-		xf_emit(ctx, 0x25, 0);
-	xf_emit(ctx, 1, 0x52);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x26);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x1a);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x00ffff00);
-	xf_emit(ctx, 1, 0);
-}
-
-static void
-nv50_graph_construct_gene_unk3(struct nouveau_grctx *ctx)
-{
-	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	/* end of area 0 on pre-NVA0, beginning of area 6 on NVAx */
-	xf_emit(ctx, 1, 0x3f);
-	xf_emit(ctx, 0xa, 0);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 2, 0x04000000);
-	xf_emit(ctx, 8, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 4);
+		xf_emit(ctx, 1, 0x7f);		/* 000000ff tesla UNK0FFC */
+	xf_emit(ctx, 1, 0);			/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 1);			/* 00000001 SHADE_MODEL */
+	xf_emit(ctx, 1, 0x80c14);		/* 01ffffff SEMANTIC_COLOR */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK1900 */
+	xf_emit(ctx, 1, 0x8100c12);		/* 1fffffff FP_INTERPOLANT_CTRL */
+	xf_emit(ctx, 1, 4);			/* 0000007f VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 4);			/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);			/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0x10);			/* 7f/ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK0D7C */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK0F8C */
+	xf_emit(ctx, 1, 0);			/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 1);			/* 00000001 VIEWPORT_TRANSFORM_EN */
+	xf_emit(ctx, 1, 0x8100c12);		/* 1fffffff FP_INTERPOLANT_CTRL */
+	xf_emit(ctx, 4, 0);			/* ffffffff NOPERSPECTIVE_BITMAP */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK1900 */
+	xf_emit(ctx, 1, 0);			/* 0000000f */
 	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 0x10, 0);
+		xf_emit(ctx, 1, 0x3ff);		/* 000003ff tesla UNK0D68 */
 	else
-		xf_emit(ctx, 0x11, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x1001);
-	xf_emit(ctx, 4, 0xffff);
-	xf_emit(ctx, 0x20, 0);
-	xf_emit(ctx, 0x10, 0x3f800000);
-	xf_emit(ctx, 1, 0x10);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 1, 0);
-	else
-		xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 3);
-	xf_emit(ctx, 2, 0);
-}
-
-static void
-nv50_graph_construct_gene_unk4(struct nouveau_grctx *ctx)
-{
-	/* middle of area 0 on pre-NVA0, middle of area 6 on NVAx */
-	xf_emit(ctx, 2, 0x04000000);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x80);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x80);
-	xf_emit(ctx, 1, 0);
-}
-
-static void
-nv50_graph_construct_gene_unk5(struct nouveau_grctx *ctx)
-{
-	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	/* middle of area 0 on pre-NVA0 [after m2mf], end of area 2 on NVAx */
-	xf_emit(ctx, 2, 4);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 0x1c4d, 0);
-	else
-		xf_emit(ctx, 0x1c4b, 0);
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0x8100c12);
-	if (dev_priv->chipset != 0x50)
-		xf_emit(ctx, 1, 3);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x8100c12);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x80c14);
-	xf_emit(ctx, 1, 1);
-	if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0x80c14);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x8100c12);
-	xf_emit(ctx, 1, 0x27);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0x3c1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0x16, 0);
-	xf_emit(ctx, 1, 0x8100c12);
-	xf_emit(ctx, 1, 0);
-}
-
-static void
-nv50_graph_construct_gene_unk6(struct nouveau_grctx *ctx)
-{
-	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	/* beginning of area 1 on pre-NVA0 [after m2mf], area 3 on NVAx */
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 0xf);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 8, 0);
-	else
-		xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 0x20);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 0x11, 0);
-	else if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 0xf, 0);
-	else
-		xf_emit(ctx, 0xe, 0);
-	xf_emit(ctx, 1, 0x1a);
-	xf_emit(ctx, 0xd, 0);
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 8);
-	xf_emit(ctx, 1, 0);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 1, 0x3ff);
-	else
-		xf_emit(ctx, 1, 0x7ff);
-	if (dev_priv->chipset == 0xa8)
-		xf_emit(ctx, 1, 0x1e00);
-	xf_emit(ctx, 0xc, 0);
-	xf_emit(ctx, 1, 0xf);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 0x125, 0);
-	else if (dev_priv->chipset < 0xa0)
-		xf_emit(ctx, 0x126, 0);
-	else if (dev_priv->chipset == 0xa0 || dev_priv->chipset >= 0xaa)
-		xf_emit(ctx, 0x124, 0);
-	else
-		xf_emit(ctx, 0x1f7, 0);
-	xf_emit(ctx, 1, 0xf);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 3, 0);
-	else
-		xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 0xa1, 0);
-	else
-		xf_emit(ctx, 0x5a, 0);
-	xf_emit(ctx, 1, 0xf);
-	if (dev_priv->chipset < 0xa0)
-		xf_emit(ctx, 0x834, 0);
-	else if (dev_priv->chipset == 0xa0)
-		xf_emit(ctx, 0x1873, 0);
-	else if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 0x8ba, 0);
-	else
-		xf_emit(ctx, 0x833, 0);
-	xf_emit(ctx, 1, 0xf);
-	xf_emit(ctx, 0xf, 0);
-}
-
-static void
-nv50_graph_construct_gene_unk7(struct nouveau_grctx *ctx)
-{
-	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	/* middle of area 1 on pre-NVA0 [after m2mf], middle of area 6 on NVAx */
-	xf_emit(ctx, 2, 0);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 2, 1);
-	else
-		xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 2, 0x100);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 8);
-	xf_emit(ctx, 5, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 3, 1);
-	xf_emit(ctx, 1, 0xcf);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 6, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 3, 1);
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x15);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x4444480);
-	xf_emit(ctx, 0x37, 0);
-}
-
-static void
-nv50_graph_construct_gene_unk8(struct nouveau_grctx *ctx)
-{
-	/* middle of area 1 on pre-NVA0 [after m2mf], middle of area 0 on NVAx */
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 0x8100c12);
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 0x100);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x10001);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x10001);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x10001);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 2);
-}
-
-static void
-nv50_graph_construct_gene_unk9(struct nouveau_grctx *ctx)
-{
-	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	/* middle of area 2 on pre-NVA0 [after m2mf], end of area 0 on NVAx */
-	xf_emit(ctx, 1, 0x3f800000);
-	xf_emit(ctx, 6, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0x1a);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0x12, 0);
-	xf_emit(ctx, 1, 0x00ffff00);
-	xf_emit(ctx, 6, 0);
-	xf_emit(ctx, 1, 0xf);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x0fac6881);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 0xf, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 2, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 3);
-	else if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 2, 0x04000000);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 5);
-	xf_emit(ctx, 1, 0x52);
-	if (dev_priv->chipset == 0x50) {
-		xf_emit(ctx, 0x13, 0);
-	} else {
-		xf_emit(ctx, 4, 0);
-		xf_emit(ctx, 1, 1);
-		if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-			xf_emit(ctx, 0x11, 0);
-		else
-			xf_emit(ctx, 0x10, 0);
-	}
-	xf_emit(ctx, 0x10, 0x3f800000);
-	xf_emit(ctx, 1, 0x10);
-	xf_emit(ctx, 0x26, 0);
-	xf_emit(ctx, 1, 0x8100c12);
-	xf_emit(ctx, 1, 5);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 4, 0xffff);
-	if (dev_priv->chipset != 0x50)
-		xf_emit(ctx, 1, 3);
-	if (dev_priv->chipset < 0xa0)
-		xf_emit(ctx, 0x1f, 0);
-	else if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 0xc, 0);
-	else
-		xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x00ffff00);
-	xf_emit(ctx, 1, 0x1a);
+		xf_emit(ctx, 1, 0x7ff);		/* 000007ff tesla UNK0D68 */
+	xf_emit(ctx, 1, 0x80c14);		/* 01ffffff SEMANTIC_COLOR */
+	xf_emit(ctx, 1, 0);			/* 00000001 VERTEX_TWO_SIDE_ENABLE */
+	xf_emit(ctx, 0x30, 0);			/* ffffffff VIEWPORT_SCALE: X0, Y0, Z0, X1, Y1, ... */
+	xf_emit(ctx, 3, 0);			/* f, 0, 0 */
+	xf_emit(ctx, 3, 0);			/* ffffffff last VIEWPORT_SCALE? */
+	xf_emit(ctx, 1, 0);			/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 1);			/* 00000001 VIEWPORT_TRANSFORM_EN */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK1900 */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK1924 */
+	xf_emit(ctx, 1, 0x10);			/* 000000ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 1, 0);			/* 00000001 */
+	xf_emit(ctx, 0x30, 0);			/* ffffffff VIEWPORT_TRANSLATE */
+	xf_emit(ctx, 3, 0);			/* f, 0, 0 */
+	xf_emit(ctx, 3, 0);			/* ffffffff */
+	xf_emit(ctx, 1, 0);			/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 2, 0x88);			/* 000001ff tesla UNK19D8 */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK1924 */
+	xf_emit(ctx, 1, 0);			/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 4);			/* 0000000f CULL_MODE */
+	xf_emit(ctx, 2, 0);			/* 07ffffff SCREEN_SCISSOR */
+	xf_emit(ctx, 2, 0);			/* 00007fff WINDOW_OFFSET_XY */
+	xf_emit(ctx, 1, 0);			/* 00000003 WINDOW_ORIGIN */
+	xf_emit(ctx, 0x10, 0);			/* 00000001 SCISSOR_ENABLE */
+	xf_emit(ctx, 1, 0);			/* 0001ffff GP_BUILTIN_RESULT_EN */
+	xf_emit(ctx, 1, 0x26);			/* 000000ff SEMANTIC_LAYER */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK1900 */
+	xf_emit(ctx, 1, 0);			/* 0000000f */
+	xf_emit(ctx, 1, 0x3f800000);		/* ffffffff LINE_WIDTH */
+	xf_emit(ctx, 1, 0);			/* 00000001 LINE_STIPPLE_ENABLE */
+	xf_emit(ctx, 1, 0);			/* 00000001 LINE_SMOOTH_ENABLE */
+	xf_emit(ctx, 1, 0);			/* 00000007 MULTISAMPLE_SAMPLES_LOG2 */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 0);		/* 00000001 */
+	xf_emit(ctx, 1, 0x1a);			/* 0000001f POLYGON_MODE */
+	xf_emit(ctx, 1, 0x10);			/* 000000ff VIEW_VOLUME_CLIP_CTRL */
 	if (dev_priv->chipset != 0x50) {
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 1, 3);
+		xf_emit(ctx, 1, 0);		/* ffffffff */
+		xf_emit(ctx, 1, 0);		/* 00000001 */
+		xf_emit(ctx, 1, 0);		/* 000003ff */
+	}
+	xf_emit(ctx, 0x20, 0);			/* 10xbits ffffffff, 3fffff. SCISSOR_* */
+	xf_emit(ctx, 1, 0);			/* f */
+	xf_emit(ctx, 1, 0);			/* 0? */
+	xf_emit(ctx, 1, 0);			/* ffffffff */
+	xf_emit(ctx, 1, 0);			/* 003fffff */
+	xf_emit(ctx, 1, 0);			/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 0x52);			/* 000001ff SEMANTIC_PTSZ */
+	xf_emit(ctx, 1, 0);			/* 0001ffff GP_BUILTIN_RESULT_EN */
+	xf_emit(ctx, 1, 0x26);			/* 000000ff SEMANTIC_LAYER */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK1900 */
+	xf_emit(ctx, 1, 4);			/* 0000007f VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 4);			/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);			/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0x1a);			/* 0000001f POLYGON_MODE */
+	xf_emit(ctx, 1, 0);			/* 00000001 LINE_SMOOTH_ENABLE */
+	xf_emit(ctx, 1, 0);			/* 00000001 LINE_STIPPLE_ENABLE */
+	xf_emit(ctx, 1, 0x00ffff00);		/* 00ffffff LINE_STIPPLE_PATTERN */
+	xf_emit(ctx, 1, 0);			/* 0000000f */
+}
+
+static void
+nv50_graph_construct_gene_zcull(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	/* end of strand 0 on pre-NVA0, beginning of strand 6 on NVAx */
+	/* SEEK */
+	xf_emit(ctx, 1, 0x3f);		/* 0000003f UNK1590 */
+	xf_emit(ctx, 1, 0);		/* 00000001 ALPHA_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000007 MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_BACK_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_FUNC_REF */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_MASK */
+	xf_emit(ctx, 3, 0);		/* 00000007 STENCIL_BACK_OP_FAIL, ZFAIL, ZPASS */
+	xf_emit(ctx, 1, 2);		/* 00000003 tesla UNK143C */
+	xf_emit(ctx, 2, 0x04000000);	/* 07ffffff tesla UNK0D6C */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, 0);		/* 00000001 CLIPID_ENABLE */
+	xf_emit(ctx, 2, 0);		/* ffffffff DEPTH_BOUNDS */
+	xf_emit(ctx, 1, 0);		/* 00000001 */
+	xf_emit(ctx, 1, 0);		/* 00000007 DEPTH_TEST_FUNC */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 0000000f CULL_MODE */
+	xf_emit(ctx, 1, 0);		/* 0000ffff */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK0FB0 */
+	xf_emit(ctx, 1, 0);		/* 00000001 POLYGON_STIPPLE_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 00000007 FP_CONTROL */
+	xf_emit(ctx, 1, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 0001ffff GP_BUILTIN_RESULT_EN */
+	xf_emit(ctx, 1, 0);		/* 000000ff CLEAR_STENCIL */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_FRONT_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_FUNC_REF */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_MASK */
+	xf_emit(ctx, 3, 0);		/* 00000007 STENCIL_FRONT_OP_FAIL, ZFAIL, ZPASS */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_BACK_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff CLEAR_DEPTH */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	if (dev_priv->chipset != 0x50)
+		xf_emit(ctx, 1, 0);	/* 00000003 tesla UNK1108 */
+	xf_emit(ctx, 1, 0);		/* 00000001 SAMPLECNT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0x1001);	/* 00001fff ZETA_ARRAY_MODE */
+	/* SEEK */
+	xf_emit(ctx, 4, 0xffff);	/* 0000ffff MSAA_MASK */
+	xf_emit(ctx, 0x10, 0);		/* 00000001 SCISSOR_ENABLE */
+	xf_emit(ctx, 0x10, 0);		/* ffffffff DEPTH_RANGE_NEAR */
+	xf_emit(ctx, 0x10, 0x3f800000);	/* ffffffff DEPTH_RANGE_FAR */
+	xf_emit(ctx, 1, 0x10);		/* 7f/ff/3ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 1, 0);		/* 00000001 VIEWPORT_CLIP_RECTS_EN */
+	xf_emit(ctx, 1, 3);		/* 00000003 FP_CTRL_UNK196C */
+	xf_emit(ctx, 1, 0);		/* 00000003 tesla UNK1968 */
+	if (dev_priv->chipset != 0x50)
+		xf_emit(ctx, 1, 0);	/* 0fffffff tesla UNK1104 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK151C */
+}
+
+static void
+nv50_graph_construct_gene_clipid(struct nouveau_grctx *ctx)
+{
+	/* middle of strand 0 on pre-NVA0 [after 24xx], middle of area 6 on NVAx */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 00000007 UNK0FB4 */
+	/* SEEK */
+	xf_emit(ctx, 4, 0);		/* 07ffffff CLIPID_REGION_HORIZ */
+	xf_emit(ctx, 4, 0);		/* 07ffffff CLIPID_REGION_VERT */
+	xf_emit(ctx, 2, 0);		/* 07ffffff SCREEN_SCISSOR */
+	xf_emit(ctx, 2, 0x04000000);	/* 07ffffff UNK1508 */
+	xf_emit(ctx, 1, 0);		/* 00000001 CLIPID_ENABLE */
+	xf_emit(ctx, 1, 0x80);		/* 00003fff CLIPID_WIDTH */
+	xf_emit(ctx, 1, 0);		/* 000000ff CLIPID_ID */
+	xf_emit(ctx, 1, 0);		/* 000000ff CLIPID_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff CLIPID_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0x80);		/* 00003fff CLIPID_HEIGHT */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_CLIPID */
+}
+
+static void
+nv50_graph_construct_gene_unk24xx(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	int i;
+	/* middle of strand 0 on pre-NVA0 [after m2mf], end of strand 2 on NVAx */
+	/* SEEK */
+	xf_emit(ctx, 0x33, 0);
+	/* SEEK */
+	xf_emit(ctx, 2, 0);
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 0000007f VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	/* SEEK */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 4, 0);	/* RO */
+		xf_emit(ctx, 0xe10, 0); /* 190 * 9: 8*ffffffff, 7ff */
+		xf_emit(ctx, 1, 0);	/* 1ff */
+		xf_emit(ctx, 8, 0);	/* 0? */
+		xf_emit(ctx, 9, 0);	/* ffffffff, 7ff */
+
+		xf_emit(ctx, 4, 0);	/* RO */
+		xf_emit(ctx, 0xe10, 0); /* 190 * 9: 8*ffffffff, 7ff */
+		xf_emit(ctx, 1, 0);	/* 1ff */
+		xf_emit(ctx, 8, 0);	/* 0? */
+		xf_emit(ctx, 9, 0);	/* ffffffff, 7ff */
 	}
-	if (dev_priv->chipset < 0xa0)
-		xf_emit(ctx, 0x26, 0);
 	else
-		xf_emit(ctx, 0x3c, 0);
-	xf_emit(ctx, 1, 0x102);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 4, 4);
+	{
+		xf_emit(ctx, 0xc, 0);	/* RO */
+		/* SEEK */
+		xf_emit(ctx, 0xe10, 0); /* 190 * 9: 8*ffffffff, 7ff */
+		xf_emit(ctx, 1, 0);	/* 1ff */
+		xf_emit(ctx, 8, 0);	/* 0? */
+
+		/* SEEK */
+		xf_emit(ctx, 0xc, 0);	/* RO */
+		/* SEEK */
+		xf_emit(ctx, 0xe10, 0); /* 190 * 9: 8*ffffffff, 7ff */
+		xf_emit(ctx, 1, 0);	/* 1ff */
+		xf_emit(ctx, 8, 0);	/* 0? */
+	}
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 4);		/* 0000007f VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0x8100c12);	/* 1fffffff FP_INTERPOLANT_CTRL */
+	if (dev_priv->chipset != 0x50)
+		xf_emit(ctx, 1, 3);	/* 00000003 tesla UNK1100 */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0x8100c12);	/* 1fffffff FP_INTERPOLANT_CTRL */
+	xf_emit(ctx, 1, 0);		/* 0000000f VP_GP_BUILTIN_ATTR_EN */
+	xf_emit(ctx, 1, 0x80c14);	/* 01ffffff SEMANTIC_COLOR */
+	xf_emit(ctx, 1, 1);		/* 00000001 */
+	/* SEEK */
 	if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 8, 0);
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 1, 0x3ff);
+		xf_emit(ctx, 2, 4);	/* 000000ff */
+	xf_emit(ctx, 1, 0x80c14);	/* 01ffffff SEMANTIC_COLOR */
+	xf_emit(ctx, 1, 0);		/* 00000001 VERTEX_TWO_SIDE_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 POINT_SPRITE_ENABLE */
+	xf_emit(ctx, 1, 0x8100c12);	/* 1fffffff FP_INTERPOLANT_CTRL */
+	xf_emit(ctx, 1, 0x27);		/* 000000ff SEMANTIC_PRIM_ID */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 0000000f */
+	xf_emit(ctx, 1, 1);		/* 00000001 */
+	for (i = 0; i < 10; i++) {
+		/* SEEK */
+		xf_emit(ctx, 0x40, 0);		/* ffffffff */
+		xf_emit(ctx, 0x10, 0);		/* 3, 0, 0.... */
+		xf_emit(ctx, 0x10, 0);		/* ffffffff */
+	}
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 00000001 POINT_SPRITE_CTRL */
+	xf_emit(ctx, 1, 1);		/* 00000001 */
+	xf_emit(ctx, 1, 0);		/* ffffffff */
+	xf_emit(ctx, 4, 0);		/* ffffffff NOPERSPECTIVE_BITMAP */
+	xf_emit(ctx, 0x10, 0);		/* 00ffffff POINT_COORD_REPLACE_MAP */
+	xf_emit(ctx, 1, 0);		/* 00000003 WINDOW_ORIGIN */
+	xf_emit(ctx, 1, 0x8100c12);	/* 1fffffff FP_INTERPOLANT_CTRL */
+	if (dev_priv->chipset != 0x50)
+		xf_emit(ctx, 1, 0);	/* 000003ff */
+}
+
+static void
+nv50_graph_construct_gene_vfetch(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	int acnt = 0x10, rep, i;
+	/* beginning of strand 1 on pre-NVA0, strand 3 on NVAx */
+	if (IS_NVA3F(dev_priv->chipset))
+		acnt = 0x20;
+	/* SEEK */
+	if (dev_priv->chipset >= 0xa0) {
+		xf_emit(ctx, 1, 0);	/* ffffffff tesla UNK13A4 */
+		xf_emit(ctx, 1, 1);	/* 00000fff tesla UNK1318 */
+	}
+	xf_emit(ctx, 1, 0);		/* ffffffff VERTEX_BUFFER_FIRST */
+	xf_emit(ctx, 1, 0);		/* 00000001 PRIMITIVE_RESTART_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK0DE8 */
+	xf_emit(ctx, 1, 0);		/* ffffffff PRIMITIVE_RESTART_INDEX */
+	xf_emit(ctx, 1, 0xf);		/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, (acnt/8)-1, 0);	/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, acnt/8, 0);	/* ffffffff VTX_ATR_MASK_UNK0DD0 */
+	xf_emit(ctx, 1, 0);		/* 0000000f VP_GP_BUILTIN_ATTR_EN */
+	xf_emit(ctx, 1, 0x20);		/* 0000ffff tesla UNK129C */
+	xf_emit(ctx, 1, 0);		/* 000000ff turing UNK370??? */
+	xf_emit(ctx, 1, 0);		/* 0000ffff turing USER_PARAM_COUNT */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	/* SEEK */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 0xb, 0);	/* RO */
+	else if (dev_priv->chipset >= 0xa0)
+		xf_emit(ctx, 0x9, 0);	/* RO */
 	else
-		xf_emit(ctx, 1, 0x7ff);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x102);
-	xf_emit(ctx, 9, 0);
-	xf_emit(ctx, 4, 4);
-	xf_emit(ctx, 0x2c, 0);
+		xf_emit(ctx, 0x8, 0);	/* RO */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 00000001 EDGE_FLAG */
+	xf_emit(ctx, 1, 0);		/* 00000001 PROVOKING_VERTEX_LAST */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0x1a);		/* 0000001f POLYGON_MODE */
+	/* SEEK */
+	xf_emit(ctx, 0xc, 0);		/* RO */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 7f/ff */
+	xf_emit(ctx, 1, 4);		/* 7f/ff VP_REG_ALLOC_RESULT */
+	xf_emit(ctx, 1, 4);		/* 7f/ff VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);		/* 0000000f VP_GP_BUILTIN_ATTR_EN */
+	xf_emit(ctx, 1, 4);		/* 000001ff UNK1A28 */
+	xf_emit(ctx, 1, 8);		/* 000001ff UNK0DF0 */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	if (dev_priv->chipset == 0x50)
+		xf_emit(ctx, 1, 0x3ff);	/* 3ff tesla UNK0D68 */
+	else
+		xf_emit(ctx, 1, 0x7ff);	/* 7ff tesla UNK0D68 */
+	if (dev_priv->chipset == 0xa8)
+		xf_emit(ctx, 1, 0x1e00);	/* 7fff */
+	/* SEEK */
+	xf_emit(ctx, 0xc, 0);		/* RO or close */
+	/* SEEK */
+	xf_emit(ctx, 1, 0xf);		/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, (acnt/8)-1, 0);	/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, 1, 0);		/* 0000000f VP_GP_BUILTIN_ATTR_EN */
+	if (dev_priv->chipset > 0x50 && dev_priv->chipset < 0xa0)
+		xf_emit(ctx, 2, 0);	/* ffffffff */
+	else
+		xf_emit(ctx, 1, 0);	/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 00000003 tesla UNK0FD8 */
+	/* SEEK */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 0x10, 0);	/* 0? */
+		xf_emit(ctx, 2, 0);	/* weird... */
+		xf_emit(ctx, 2, 0);	/* RO */
+	} else {
+		xf_emit(ctx, 8, 0);	/* 0? */
+		xf_emit(ctx, 1, 0);	/* weird... */
+		xf_emit(ctx, 2, 0);	/* RO */
+	}
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* ffffffff VB_ELEMENT_BASE */
+	xf_emit(ctx, 1, 0);		/* ffffffff UNK1438 */
+	xf_emit(ctx, acnt, 0);		/* 1 tesla UNK1000 */
+	if (dev_priv->chipset >= 0xa0)
+		xf_emit(ctx, 1, 0);	/* ffffffff tesla UNK1118? */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* ffffffff VERTEX_ARRAY_UNK90C */
+	xf_emit(ctx, 1, 0);		/* f/1f */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* ffffffff VERTEX_ARRAY_UNK90C */
+	xf_emit(ctx, 1, 0);		/* f/1f */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* RO */
+	xf_emit(ctx, 2, 0);		/* RO */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK111C? */
+	xf_emit(ctx, 1, 0);		/* RO */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 000000ff UNK15F4_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff UNK15F4_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0);		/* 000000ff UNK0F84_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff UNK0F84_ADDRESS_LOW */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* 00003fff VERTEX_ARRAY_ATTRIB_OFFSET */
+	xf_emit(ctx, 3, 0);		/* f/1f */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* 00000fff VERTEX_ARRAY_STRIDE */
+	xf_emit(ctx, 3, 0);		/* f/1f */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* ffffffff VERTEX_ARRAY_LOW */
+	xf_emit(ctx, 3, 0);		/* f/1f */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* 000000ff VERTEX_ARRAY_HIGH */
+	xf_emit(ctx, 3, 0);		/* f/1f */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* ffffffff VERTEX_LIMIT_LOW */
+	xf_emit(ctx, 3, 0);		/* f/1f */
+	/* SEEK */
+	xf_emit(ctx, acnt, 0);		/* 000000ff VERTEX_LIMIT_HIGH */
+	xf_emit(ctx, 3, 0);		/* f/1f */
+	/* SEEK */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, acnt, 0);		/* f */
+		xf_emit(ctx, 3, 0);		/* f/1f */
+	}
+	/* SEEK */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 2, 0);	/* RO */
+	else
+		xf_emit(ctx, 5, 0);	/* RO */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* ffff DMA_VTXBUF */
+	/* SEEK */
+	if (dev_priv->chipset < 0xa0) {
+		xf_emit(ctx, 0x41, 0);	/* RO */
+		/* SEEK */
+		xf_emit(ctx, 0x11, 0);	/* RO */
+	} else if (!IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 0x50, 0);	/* RO */
+	else
+		xf_emit(ctx, 0x58, 0);	/* RO */
+	/* SEEK */
+	xf_emit(ctx, 1, 0xf);		/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, (acnt/8)-1, 0);	/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, 1, 1);		/* 1 UNK0DEC */
+	/* SEEK */
+	xf_emit(ctx, acnt*4, 0);	/* ffffffff VTX_ATTR */
+	xf_emit(ctx, 4, 0);		/* f/1f, 0, 0, 0 */
+	/* SEEK */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 0x1d, 0);	/* RO */
+	else
+		xf_emit(ctx, 0x16, 0);	/* RO */
+	/* SEEK */
+	xf_emit(ctx, 1, 0xf);		/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, (acnt/8)-1, 0);	/* ffffffff VP_ATTR_EN */
+	/* SEEK */
+	if (dev_priv->chipset < 0xa0)
+		xf_emit(ctx, 8, 0);	/* RO */
+	else if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 0xc, 0);	/* RO */
+	else
+		xf_emit(ctx, 7, 0);	/* RO */
+	/* SEEK */
+	xf_emit(ctx, 0xa, 0);		/* RO */
+	if (dev_priv->chipset == 0xa0)
+		rep = 0xc;
+	else
+		rep = 4;
+	for (i = 0; i < rep; i++) {
+		/* SEEK */
+		if (IS_NVA3F(dev_priv->chipset))
+			xf_emit(ctx, 0x20, 0);	/* ffffffff */
+		xf_emit(ctx, 0x200, 0);	/* ffffffff */
+		xf_emit(ctx, 4, 0);	/* 7f/ff, 0, 0, 0 */
+		xf_emit(ctx, 4, 0);	/* ffffffff */
+	}
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 113/111 */
+	xf_emit(ctx, 1, 0xf);		/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, (acnt/8)-1, 0);	/* ffffffff VP_ATTR_EN */
+	xf_emit(ctx, acnt/8, 0);	/* ffffffff VTX_ATTR_MASK_UNK0DD0 */
+	xf_emit(ctx, 1, 0);		/* 0000000f VP_GP_BUILTIN_ATTR_EN */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	/* SEEK */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 7, 0);	/* weird... */
+	else
+		xf_emit(ctx, 5, 0);	/* weird... */
+}
+
+static void
+nv50_graph_construct_gene_eng2d(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	/* middle of strand 1 on pre-NVA0 [after vfetch], middle of strand 6 on NVAx */
+	/* SEEK */
+	xf_emit(ctx, 2, 0);		/* 0001ffff CLIP_X, CLIP_Y */
+	xf_emit(ctx, 2, 0);		/* 0000ffff CLIP_W, CLIP_H */
+	xf_emit(ctx, 1, 0);		/* 00000001 CLIP_ENABLE */
+	if (dev_priv->chipset < 0xa0) {
+		/* this is useless on everything but the original NV50,
+		 * guess they forgot to nuke it. Or just didn't bother. */
+		xf_emit(ctx, 2, 0);	/* 0000ffff IFC_CLIP_X, Y */
+		xf_emit(ctx, 2, 1);	/* 0000ffff IFC_CLIP_W, H */
+		xf_emit(ctx, 1, 0);	/* 00000001 IFC_CLIP_ENABLE */
+	}
+	xf_emit(ctx, 1, 1);		/* 00000001 DST_LINEAR */
+	xf_emit(ctx, 1, 0x100);		/* 0001ffff DST_WIDTH */
+	xf_emit(ctx, 1, 0x100);		/* 0001ffff DST_HEIGHT */
+	xf_emit(ctx, 1, 0x11);		/* 3f[NV50]/7f[NV84+] DST_FORMAT */
+	xf_emit(ctx, 1, 0);		/* 0001ffff DRAW_POINT_X */
+	xf_emit(ctx, 1, 8);		/* 0000000f DRAW_UNK58C */
+	xf_emit(ctx, 1, 0);		/* 000fffff SIFC_DST_X_FRACT */
+	xf_emit(ctx, 1, 0);		/* 0001ffff SIFC_DST_X_INT */
+	xf_emit(ctx, 1, 0);		/* 000fffff SIFC_DST_Y_FRACT */
+	xf_emit(ctx, 1, 0);		/* 0001ffff SIFC_DST_Y_INT */
+	xf_emit(ctx, 1, 0);		/* 000fffff SIFC_DX_DU_FRACT */
+	xf_emit(ctx, 1, 1);		/* 0001ffff SIFC_DX_DU_INT */
+	xf_emit(ctx, 1, 0);		/* 000fffff SIFC_DY_DV_FRACT */
+	xf_emit(ctx, 1, 1);		/* 0001ffff SIFC_DY_DV_INT */
+	xf_emit(ctx, 1, 1);		/* 0000ffff SIFC_WIDTH */
+	xf_emit(ctx, 1, 1);		/* 0000ffff SIFC_HEIGHT */
+	xf_emit(ctx, 1, 0xcf);		/* 000000ff SIFC_FORMAT */
+	xf_emit(ctx, 1, 2);		/* 00000003 SIFC_BITMAP_UNK808 */
+	xf_emit(ctx, 1, 0);		/* 00000003 SIFC_BITMAP_LINE_PACK_MODE */
+	xf_emit(ctx, 1, 0);		/* 00000001 SIFC_BITMAP_LSB_FIRST */
+	xf_emit(ctx, 1, 0);		/* 00000001 SIFC_BITMAP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 0000ffff BLIT_DST_X */
+	xf_emit(ctx, 1, 0);		/* 0000ffff BLIT_DST_Y */
+	xf_emit(ctx, 1, 0);		/* 000fffff BLIT_DU_DX_FRACT */
+	xf_emit(ctx, 1, 1);		/* 0001ffff BLIT_DU_DX_INT */
+	xf_emit(ctx, 1, 0);		/* 000fffff BLIT_DV_DY_FRACT */
+	xf_emit(ctx, 1, 1);		/* 0001ffff BLIT_DV_DY_INT */
+	xf_emit(ctx, 1, 1);		/* 0000ffff BLIT_DST_W */
+	xf_emit(ctx, 1, 1);		/* 0000ffff BLIT_DST_H */
+	xf_emit(ctx, 1, 0);		/* 000fffff BLIT_SRC_X_FRACT */
+	xf_emit(ctx, 1, 0);		/* 0001ffff BLIT_SRC_X_INT */
+	xf_emit(ctx, 1, 0);		/* 000fffff BLIT_SRC_Y_FRACT */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK888 */
+	xf_emit(ctx, 1, 4);		/* 0000003f UNK884 */
+	xf_emit(ctx, 1, 0);		/* 00000007 UNK880 */
+	xf_emit(ctx, 1, 1);		/* 0000001f tesla UNK0FB8 */
+	xf_emit(ctx, 1, 0x15);		/* 000000ff tesla UNK128C */
+	xf_emit(ctx, 2, 0);		/* 00000007, ffff0ff3 */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK260 */
+	xf_emit(ctx, 1, 0x4444480);	/* 1fffffff UNK870 */
+	/* SEEK */
+	xf_emit(ctx, 0x10, 0);
+	/* SEEK */
+	xf_emit(ctx, 0x27, 0);
+}
+
+static void
+nv50_graph_construct_gene_csched(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	/* middle of strand 1 on pre-NVA0 [after eng2d], middle of strand 0 on NVAx */
+	/* SEEK */
+	xf_emit(ctx, 2, 0);		/* 00007fff WINDOW_OFFSET_XY... what is it doing here??? */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1924 */
+	xf_emit(ctx, 1, 0);		/* 00000003 WINDOW_ORIGIN */
+	xf_emit(ctx, 1, 0x8100c12);	/* 1fffffff FP_INTERPOLANT_CTRL */
+	xf_emit(ctx, 1, 0);		/* 000003ff */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* ffffffff turing UNK364 */
+	xf_emit(ctx, 1, 0);		/* 0000000f turing UNK36C */
+	xf_emit(ctx, 1, 0);		/* 0000ffff USER_PARAM_COUNT */
+	xf_emit(ctx, 1, 0x100);		/* 00ffffff turing UNK384 */
+	xf_emit(ctx, 1, 0);		/* 0000000f turing UNK2A0 */
+	xf_emit(ctx, 1, 0);		/* 0000ffff GRIDID */
+	xf_emit(ctx, 1, 0x10001);	/* ffffffff GRIDDIM_XY */
+	xf_emit(ctx, 1, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0x10001);	/* ffffffff BLOCKDIM_XY */
+	xf_emit(ctx, 1, 1);		/* 0000ffff BLOCKDIM_Z */
+	xf_emit(ctx, 1, 0x10001);	/* 00ffffff BLOCK_ALLOC */
+	xf_emit(ctx, 1, 1);		/* 00000001 LANES32 */
+	xf_emit(ctx, 1, 4);		/* 000000ff FP_REG_ALLOC_TEMP */
+	xf_emit(ctx, 1, 2);		/* 00000003 REG_MODE */
+	/* SEEK */
+	xf_emit(ctx, 0x40, 0);		/* ffffffff USER_PARAM */
+	switch (dev_priv->chipset) {
+	case 0x50:
+	case 0x92:
+		xf_emit(ctx, 8, 0);	/* 7, 0, 0, 0, ... */
+		xf_emit(ctx, 0x80, 0);	/* fff */
+		xf_emit(ctx, 2, 0);	/* ff, fff */
+		xf_emit(ctx, 0x10*2, 0);	/* ffffffff, 1f */
+		break;
+	case 0x84:
+		xf_emit(ctx, 8, 0);	/* 7, 0, 0, 0, ... */
+		xf_emit(ctx, 0x60, 0);	/* fff */
+		xf_emit(ctx, 2, 0);	/* ff, fff */
+		xf_emit(ctx, 0xc*2, 0);	/* ffffffff, 1f */
+		break;
+	case 0x94:
+	case 0x96:
+		xf_emit(ctx, 8, 0);	/* 7, 0, 0, 0, ... */
+		xf_emit(ctx, 0x40, 0);	/* fff */
+		xf_emit(ctx, 2, 0);	/* ff, fff */
+		xf_emit(ctx, 8*2, 0);	/* ffffffff, 1f */
+		break;
+	case 0x86:
+	case 0x98:
+		xf_emit(ctx, 4, 0);	/* f, 0, 0, 0 */
+		xf_emit(ctx, 0x10, 0);	/* fff */
+		xf_emit(ctx, 2, 0);	/* ff, fff */
+		xf_emit(ctx, 2*2, 0);	/* ffffffff, 1f */
+		break;
+	case 0xa0:
+		xf_emit(ctx, 8, 0);	/* 7, 0, 0, 0, ... */
+		xf_emit(ctx, 0xf0, 0);	/* fff */
+		xf_emit(ctx, 2, 0);	/* ff, fff */
+		xf_emit(ctx, 0x1e*2, 0);	/* ffffffff, 1f */
+		break;
+	case 0xa3:
+		xf_emit(ctx, 8, 0);	/* 7, 0, 0, 0, ... */
+		xf_emit(ctx, 0x60, 0);	/* fff */
+		xf_emit(ctx, 2, 0);	/* ff, fff */
+		xf_emit(ctx, 0xc*2, 0);	/* ffffffff, 1f */
+		break;
+	case 0xa5:
+	case 0xaf:
+		xf_emit(ctx, 8, 0);	/* 7, 0, 0, 0, ... */
+		xf_emit(ctx, 0x30, 0);	/* fff */
+		xf_emit(ctx, 2, 0);	/* ff, fff */
+		xf_emit(ctx, 6*2, 0);	/* ffffffff, 1f */
+		break;
+	case 0xaa:
+		xf_emit(ctx, 0x12, 0);
+		break;
+	case 0xa8:
+	case 0xac:
+		xf_emit(ctx, 4, 0);	/* f, 0, 0, 0 */
+		xf_emit(ctx, 0x10, 0);	/* fff */
+		xf_emit(ctx, 2, 0);	/* ff, fff */
+		xf_emit(ctx, 2*2, 0);	/* ffffffff, 1f */
+		break;
+	}
+	xf_emit(ctx, 1, 0);		/* 0000000f */
+	xf_emit(ctx, 1, 0);		/* 00000000 */
+	xf_emit(ctx, 1, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 0000001f */
+	xf_emit(ctx, 4, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 00000003 turing UNK35C */
+	xf_emit(ctx, 1, 0);		/* ffffffff */
+	xf_emit(ctx, 4, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 00000003 turing UNK35C */
+	xf_emit(ctx, 1, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 000000ff */
+}
+
+static void
+nv50_graph_construct_gene_unk1cxx(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	xf_emit(ctx, 2, 0);		/* 00007fff WINDOW_OFFSET_XY */
+	xf_emit(ctx, 1, 0x3f800000);	/* ffffffff LINE_WIDTH */
+	xf_emit(ctx, 1, 0);		/* 00000001 LINE_SMOOTH_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1658 */
+	xf_emit(ctx, 1, 0);		/* 00000001 POLYGON_SMOOTH_ENABLE */
+	xf_emit(ctx, 3, 0);		/* 00000001 POLYGON_OFFSET_*_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 0000000f CULL_MODE */
+	xf_emit(ctx, 1, 0x1a);		/* 0000001f POLYGON_MODE */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 0);		/* 00000001 POINT_SPRITE_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK165C */
+	xf_emit(ctx, 0x10, 0);		/* 00000001 SCISSOR_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 00000001 LINE_STIPPLE_ENABLE */
+	xf_emit(ctx, 1, 0x00ffff00);	/* 00ffffff LINE_STIPPLE_PATTERN */
+	xf_emit(ctx, 1, 0);		/* ffffffff POLYGON_OFFSET_UNITS */
+	xf_emit(ctx, 1, 0);		/* ffffffff POLYGON_OFFSET_FACTOR */
+	xf_emit(ctx, 1, 0);		/* 00000003 tesla UNK1668 */
+	xf_emit(ctx, 2, 0);		/* 07ffffff SCREEN_SCISSOR */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1900 */
+	xf_emit(ctx, 1, 0xf);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 7, 0);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 1, 0x11);		/* 0000007f RT_FORMAT */
+	xf_emit(ctx, 7, 0);		/* 0000007f RT_FORMAT */
+	xf_emit(ctx, 8, 0);		/* 00000001 RT_HORIZ_LINEAR */
+	xf_emit(ctx, 1, 4);		/* 00000007 FP_CONTROL */
+	xf_emit(ctx, 1, 0);		/* 00000001 ALPHA_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000007 ALPHA_TEST_FUNC */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 3);	/* 00000003 UNK16B4 */
+	else if (dev_priv->chipset >= 0xa0)
+		xf_emit(ctx, 1, 1);	/* 00000001 UNK16B4 */
+	xf_emit(ctx, 1, 0);		/* 00000003 MULTISAMPLE_CTRL */
+	xf_emit(ctx, 1, 0);		/* 00000003 tesla UNK0F90 */
+	xf_emit(ctx, 1, 2);		/* 00000003 tesla UNK143C */
+	xf_emit(ctx, 2, 0x04000000);	/* 07ffffff tesla UNK0D6C */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_MASK */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 SAMPLECNT_ENABLE */
+	xf_emit(ctx, 1, 5);		/* 0000000f UNK1408 */
+	xf_emit(ctx, 1, 0x52);		/* 000001ff SEMANTIC_PTSZ */
+	xf_emit(ctx, 1, 0);		/* ffffffff POINT_SIZE */
+	xf_emit(ctx, 1, 0);		/* 00000001 */
+	xf_emit(ctx, 1, 0);		/* 00000007 tesla UNK0FB4 */
+	if (dev_priv->chipset != 0x50) {
+		xf_emit(ctx, 1, 0);	/* 3ff */
+		xf_emit(ctx, 1, 1);	/* 00000001 tesla UNK1110 */
+	}
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 0);	/* 00000003 tesla UNK1928 */
+	xf_emit(ctx, 0x10, 0);		/* ffffffff DEPTH_RANGE_NEAR */
+	xf_emit(ctx, 0x10, 0x3f800000);	/* ffffffff DEPTH_RANGE_FAR */
+	xf_emit(ctx, 1, 0x10);		/* 000000ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 0x20, 0);		/* 07ffffff VIEWPORT_HORIZ, then VIEWPORT_VERT. (W&0x3fff)<<13 | (X&0x1fff). */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK187C */
+	xf_emit(ctx, 1, 0);		/* 00000003 WINDOW_ORIGIN */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_BACK_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_MASK */
+	xf_emit(ctx, 1, 0x8100c12);	/* 1fffffff FP_INTERPOLANT_CTRL */
+	xf_emit(ctx, 1, 5);		/* 0000000f tesla UNK1220 */
+	xf_emit(ctx, 1, 0);		/* 00000007 MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 000000ff tesla UNK1A20 */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 VERTEX_TWO_SIDE_ENABLE */
+	xf_emit(ctx, 4, 0xffff);	/* 0000ffff MSAA_MASK */
+	if (dev_priv->chipset != 0x50)
+		xf_emit(ctx, 1, 3);	/* 00000003 tesla UNK1100 */
+	if (dev_priv->chipset < 0xa0)
+		xf_emit(ctx, 0x1c, 0);	/* RO */
+	else if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 0x9, 0);
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 00000001 LINE_SMOOTH_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 LINE_STIPPLE_ENABLE */
+	xf_emit(ctx, 1, 0x00ffff00);	/* 00ffffff LINE_STIPPLE_PATTERN */
+	xf_emit(ctx, 1, 0x1a);		/* 0000001f POLYGON_MODE */
+	xf_emit(ctx, 1, 0);		/* 00000003 WINDOW_ORIGIN */
+	if (dev_priv->chipset != 0x50) {
+		xf_emit(ctx, 1, 3);	/* 00000003 tesla UNK1100 */
+		xf_emit(ctx, 1, 0);	/* 3ff */
+	}
+	/* XXX: the following block could belong either to unk1cxx, or
+	 * to STRMOUT. Rather hard to tell. */
+	if (dev_priv->chipset < 0xa0)
+		xf_emit(ctx, 0x25, 0);
+	else
+		xf_emit(ctx, 0x3b, 0);
+}
+
+static void
+nv50_graph_construct_gene_strmout(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	xf_emit(ctx, 1, 0x102);		/* 0000ffff STRMOUT_BUFFER_CTRL */
+	xf_emit(ctx, 1, 0);		/* ffffffff STRMOUT_PRIMITIVE_COUNT */
+	xf_emit(ctx, 4, 4);		/* 000000ff STRMOUT_NUM_ATTRIBS */
+	if (dev_priv->chipset >= 0xa0) {
+		xf_emit(ctx, 4, 0);	/* ffffffff UNK1A8C */
+		xf_emit(ctx, 4, 0);	/* ffffffff UNK1780 */
+	}
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 4);		/* 0000007f VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	if (dev_priv->chipset == 0x50)
+		xf_emit(ctx, 1, 0x3ff);	/* 000003ff tesla UNK0D68 */
+	else
+		xf_emit(ctx, 1, 0x7ff);	/* 000007ff tesla UNK0D68 */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	/* SEEK */
+	xf_emit(ctx, 1, 0x102);		/* 0000ffff STRMOUT_BUFFER_CTRL */
+	xf_emit(ctx, 1, 0);		/* ffffffff STRMOUT_PRIMITIVE_COUNT */
+	xf_emit(ctx, 4, 0);		/* 000000ff STRMOUT_ADDRESS_HIGH */
+	xf_emit(ctx, 4, 0);		/* ffffffff STRMOUT_ADDRESS_LOW */
+	xf_emit(ctx, 4, 4);		/* 000000ff STRMOUT_NUM_ATTRIBS */
+	if (dev_priv->chipset >= 0xa0) {
+		xf_emit(ctx, 4, 0);	/* ffffffff UNK1A8C */
+		xf_emit(ctx, 4, 0);	/* ffffffff UNK1780 */
+	}
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_STRMOUT */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_QUERY */
+	xf_emit(ctx, 1, 0);		/* 000000ff QUERY_ADDRESS_HIGH */
+	xf_emit(ctx, 2, 0);		/* ffffffff QUERY_ADDRESS_LOW QUERY_COUNTER */
+	xf_emit(ctx, 2, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	/* SEEK */
+	xf_emit(ctx, 0x20, 0);		/* ffffffff STRMOUT_MAP */
+	xf_emit(ctx, 1, 0);		/* 0000000f */
+	xf_emit(ctx, 1, 0);		/* 00000000? */
+	xf_emit(ctx, 2, 0);		/* ffffffff */
+}
+
+static void
+nv50_graph_construct_gene_ropm1(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	xf_emit(ctx, 1, 0x4e3bfdf);	/* ffffffff UNK0D64 */
+	xf_emit(ctx, 1, 0x4e3bfdf);	/* ffffffff UNK0DF4 */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	xf_emit(ctx, 1, 0);		/* 000003ff */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 0x11);	/* 000000ff tesla UNK1968 */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
+}
+
+static void
+nv50_graph_construct_gene_ropm2(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_QUERY */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 2, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 000000ff QUERY_ADDRESS_HIGH */
+	xf_emit(ctx, 2, 0);		/* ffffffff QUERY_ADDRESS_LOW, COUNTER */
+	xf_emit(ctx, 1, 0);		/* 00000001 SAMPLECNT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 7 */
+	/* SEEK */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_QUERY */
+	xf_emit(ctx, 1, 0);		/* 000000ff QUERY_ADDRESS_HIGH */
+	xf_emit(ctx, 2, 0);		/* ffffffff QUERY_ADDRESS_LOW, COUNTER */
+	xf_emit(ctx, 1, 0x4e3bfdf);	/* ffffffff UNK0D64 */
+	xf_emit(ctx, 1, 0x4e3bfdf);	/* ffffffff UNK0DF4 */
+	xf_emit(ctx, 1, 0);		/* 00000001 eng2d UNK260 */
+	xf_emit(ctx, 1, 0);		/* ff/3ff */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 0x11);	/* 000000ff tesla UNK1968 */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
 }
 
 static void
@@ -1749,443 +2392,709 @@ nv50_graph_construct_gene_ropc(struct nouveau_grctx *ctx)
 	int magic2;
 	if (dev_priv->chipset == 0x50) {
 		magic2 = 0x00003e60;
-	} else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) {
+	} else if (!IS_NVA3F(dev_priv->chipset)) {
 		magic2 = 0x001ffe67;
 	} else {
 		magic2 = 0x00087e67;
 	}
-	xf_emit(ctx, 8, 0);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, magic2);
-	xf_emit(ctx, 4, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 7, 0);
-	if (dev_priv->chipset >= 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 0x15);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x10);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 4, 0);
+	xf_emit(ctx, 1, 0);		/* f/7 MUTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_BACK_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_MASK */
+	xf_emit(ctx, 3, 0);		/* 00000007 STENCIL_BACK_OP_FAIL, ZFAIL, ZPASS */
+	xf_emit(ctx, 1, 2);		/* 00000003 tesla UNK143C */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, magic2);	/* 001fffff tesla UNK0F78 */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_BOUNDS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000007 DEPTH_TEST_FUNC */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_FRONT_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_MASK */
+	xf_emit(ctx, 3, 0);		/* 00000007 STENCIL_FRONT_OP_FAIL, ZFAIL, ZPASS */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	if (dev_priv->chipset >= 0xa0 && !IS_NVAAF(dev_priv->chipset))
+		xf_emit(ctx, 1, 0x15);	/* 000000ff */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_BACK_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK15B4 */
+	xf_emit(ctx, 1, 0x10);		/* 3ff/ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 1, 0);		/* ffffffff CLEAR_DEPTH */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
 	if (dev_priv->chipset == 0x86 || dev_priv->chipset == 0x92 || dev_priv->chipset == 0x98 || dev_priv->chipset >= 0xa0) {
-		xf_emit(ctx, 1, 4);
-		xf_emit(ctx, 1, 0x400);
-		xf_emit(ctx, 1, 0x300);
-		xf_emit(ctx, 1, 0x1001);
+		xf_emit(ctx, 3, 0);	/* ff, ffffffff, ffffffff */
+		xf_emit(ctx, 1, 4);	/* 7 */
+		xf_emit(ctx, 1, 0x400);	/* fffffff */
+		xf_emit(ctx, 1, 0x300);	/* ffff */
+		xf_emit(ctx, 1, 0x1001);	/* 1fff */
 		if (dev_priv->chipset != 0xa0) {
-			if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-				xf_emit(ctx, 1, 0);
+			if (IS_NVA3F(dev_priv->chipset))
+				xf_emit(ctx, 1, 0);	/* 0000000f UNK15C8 */
 			else
-				xf_emit(ctx, 1, 0x15);
+				xf_emit(ctx, 1, 0x15);	/* ff */
 		}
-		xf_emit(ctx, 3, 0);
 	}
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 8, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x10);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0x13, 0);
-	xf_emit(ctx, 1, 0x10);
-	xf_emit(ctx, 0x10, 0);
-	xf_emit(ctx, 0x10, 0x3f800000);
-	xf_emit(ctx, 0x19, 0);
-	xf_emit(ctx, 1, 0x10);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x3f);
-	xf_emit(ctx, 6, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
+	xf_emit(ctx, 1, 0);		/* 00000007 MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_BACK_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, 2);		/* 00000003 tesla UNK143C */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_BOUNDS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000007 DEPTH_TEST_FUNC */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_FRONT_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_BACK_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK15B4 */
+	xf_emit(ctx, 1, 0x10);		/* 7f/ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1900 */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_BACK_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_FUNC_REF */
+	xf_emit(ctx, 2, 0);		/* ffffffff DEPTH_BOUNDS */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_BOUNDS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000007 DEPTH_TEST_FUNC */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 0000000f */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK0FB0 */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_FRONT_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_FUNC_REF */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_BACK_ENABLE */
+	xf_emit(ctx, 1, 0x10);		/* 7f/ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 0x10, 0);		/* ffffffff DEPTH_RANGE_NEAR */
+	xf_emit(ctx, 0x10, 0x3f800000);	/* ffffffff DEPTH_RANGE_FAR */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 0);		/* 00000007 MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_BACK_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_FUNC_REF */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_MASK */
+	xf_emit(ctx, 3, 0);		/* 00000007 STENCIL_BACK_OP_FAIL, ZFAIL, ZPASS */
+	xf_emit(ctx, 2, 0);		/* ffffffff DEPTH_BOUNDS */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_BOUNDS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000007 DEPTH_TEST_FUNC */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 000000ff CLEAR_STENCIL */
+	xf_emit(ctx, 1, 0);		/* 00000007 STENCIL_FRONT_FUNC_FUNC */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_FUNC_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_FUNC_REF */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_MASK */
+	xf_emit(ctx, 3, 0);		/* 00000007 STENCIL_FRONT_OP_FAIL, ZFAIL, ZPASS */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_BACK_ENABLE */
+	xf_emit(ctx, 1, 0x10);		/* 7f/ff VIEW_VOLUME_CLIP_CTRL */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 0x3f);		/* 0000003f UNK1590 */
+	xf_emit(ctx, 1, 0);		/* 00000007 MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 2, 0);		/* ffff0ff3, ffff */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK0FB0 */
+	xf_emit(ctx, 1, 0);		/* 0001ffff GP_BUILTIN_RESULT_EN */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK15B4 */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff CLEAR_DEPTH */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK19CC */
 	if (dev_priv->chipset >= 0xa0) {
 		xf_emit(ctx, 2, 0);
 		xf_emit(ctx, 1, 0x1001);
 		xf_emit(ctx, 0xb, 0);
 	} else {
-		xf_emit(ctx, 0xc, 0);
+		xf_emit(ctx, 1, 0);	/* 00000007 */
+		xf_emit(ctx, 1, 0);	/* 00000001 tesla UNK1534 */
+		xf_emit(ctx, 1, 0);	/* 00000007 MULTISAMPLE_SAMPLES_LOG2 */
+		xf_emit(ctx, 8, 0);	/* 00000001 BLEND_ENABLE */
+		xf_emit(ctx, 1, 0);	/* ffff0ff3 */
 	}
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0xf);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x11);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 4, 0);
-	else
-		xf_emit(ctx, 6, 0);
-	xf_emit(ctx, 3, 1);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, magic2);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x0fac6881);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 0x18, 1);
-		xf_emit(ctx, 8, 2);
-		xf_emit(ctx, 8, 1);
-		xf_emit(ctx, 8, 2);
-		xf_emit(ctx, 8, 1);
-		xf_emit(ctx, 3, 0);
-		xf_emit(ctx, 1, 1);
-		xf_emit(ctx, 5, 0);
-		xf_emit(ctx, 1, 1);
-		xf_emit(ctx, 0x16, 0);
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 7, 0);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 1, 0xf);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 7, 0);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f */
+	xf_emit(ctx, 1, 0);		/* 00000001 LOGIC_OP_ENABLE */
+	if (dev_priv->chipset != 0x50) {
+		xf_emit(ctx, 1, 0);	/* 0000000f LOGIC_OP */
+		xf_emit(ctx, 1, 0);	/* 000000ff */
+	}
+	xf_emit(ctx, 1, 0);		/* 00000007 OPERATION */
+	xf_emit(ctx, 1, 0);		/* ff/3ff */
+	xf_emit(ctx, 1, 0);		/* 00000003 UNK0F90 */
+	xf_emit(ctx, 2, 1);		/* 00000007 BLEND_EQUATION_RGB, ALPHA */
+	xf_emit(ctx, 1, 1);		/* 00000001 UNK133C */
+	xf_emit(ctx, 1, 2);		/* 0000001f BLEND_FUNC_SRC_RGB */
+	xf_emit(ctx, 1, 1);		/* 0000001f BLEND_FUNC_DST_RGB */
+	xf_emit(ctx, 1, 2);		/* 0000001f BLEND_FUNC_SRC_ALPHA */
+	xf_emit(ctx, 1, 1);		/* 0000001f BLEND_FUNC_DST_ALPHA */
+	xf_emit(ctx, 1, 0);		/* 00000001 */
+	xf_emit(ctx, 1, magic2);	/* 001fffff tesla UNK0F78 */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 1, 0);	/* 00000001 tesla UNK12E4 */
+		xf_emit(ctx, 8, 1);	/* 00000007 IBLEND_EQUATION_RGB */
+		xf_emit(ctx, 8, 1);	/* 00000007 IBLEND_EQUATION_ALPHA */
+		xf_emit(ctx, 8, 1);	/* 00000001 IBLEND_UNK00 */
+		xf_emit(ctx, 8, 2);	/* 0000001f IBLEND_FUNC_SRC_RGB */
+		xf_emit(ctx, 8, 1);	/* 0000001f IBLEND_FUNC_DST_RGB */
+		xf_emit(ctx, 8, 2);	/* 0000001f IBLEND_FUNC_SRC_ALPHA */
+		xf_emit(ctx, 8, 1);	/* 0000001f IBLEND_FUNC_DST_ALPHA */
+		xf_emit(ctx, 1, 0);	/* 00000001 tesla UNK1140 */
+		xf_emit(ctx, 2, 0);	/* 00000001 */
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+		xf_emit(ctx, 1, 0);	/* 0000000f */
+		xf_emit(ctx, 1, 0);	/* 00000003 */
+		xf_emit(ctx, 1, 0);	/* ffffffff */
+		xf_emit(ctx, 2, 0);	/* 00000001 */
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+		xf_emit(ctx, 1, 0);	/* 00000001 */
+		xf_emit(ctx, 1, 0);	/* 000003ff */
+	} else if (dev_priv->chipset >= 0xa0) {
+		xf_emit(ctx, 2, 0);	/* 00000001 */
+		xf_emit(ctx, 1, 0);	/* 00000007 */
+		xf_emit(ctx, 1, 0);	/* 00000003 */
+		xf_emit(ctx, 1, 0);	/* ffffffff */
+		xf_emit(ctx, 2, 0);	/* 00000001 */
 	} else {
-		if (dev_priv->chipset >= 0xa0)
-			xf_emit(ctx, 0x1b, 0);
-		else
-			xf_emit(ctx, 0x15, 0);
+		xf_emit(ctx, 1, 0);	/* 00000007 MULTISAMPLE_SAMPLES_LOG2 */
+		xf_emit(ctx, 1, 0);	/* 00000003 tesla UNK1430 */
+		xf_emit(ctx, 1, 0);	/* ffffffff tesla UNK1A3C */
 	}
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 2, 1);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 2, 1);
+	xf_emit(ctx, 4, 0);		/* ffffffff CLEAR_COLOR */
+	xf_emit(ctx, 4, 0);		/* ffffffff BLEND_COLOR A R G B */
+	xf_emit(ctx, 1, 0);		/* 00000fff eng2d UNK2B0 */
 	if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 4, 0);
-	else
-		xf_emit(ctx, 3, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		xf_emit(ctx, 0x10, 1);
-		xf_emit(ctx, 8, 2);
-		xf_emit(ctx, 0x10, 1);
-		xf_emit(ctx, 8, 2);
-		xf_emit(ctx, 8, 1);
-		xf_emit(ctx, 3, 0);
+		xf_emit(ctx, 2, 0);	/* 00000001 */
+	xf_emit(ctx, 1, 0);		/* 000003ff */
+	xf_emit(ctx, 8, 0);		/* 00000001 BLEND_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 00000001 UNK133C */
+	xf_emit(ctx, 1, 2);		/* 0000001f BLEND_FUNC_SRC_RGB */
+	xf_emit(ctx, 1, 1);		/* 0000001f BLEND_FUNC_DST_RGB */
+	xf_emit(ctx, 1, 1);		/* 00000007 BLEND_EQUATION_RGB */
+	xf_emit(ctx, 1, 2);		/* 0000001f BLEND_FUNC_SRC_ALPHA */
+	xf_emit(ctx, 1, 1);		/* 0000001f BLEND_FUNC_DST_ALPHA */
+	xf_emit(ctx, 1, 1);		/* 00000007 BLEND_EQUATION_ALPHA */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK19C0 */
+	xf_emit(ctx, 1, 0);		/* 00000001 LOGIC_OP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 0000000f LOGIC_OP */
+	if (dev_priv->chipset >= 0xa0)
+		xf_emit(ctx, 1, 0);	/* 00000001 UNK12E4? NVA3+ only? */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 8, 1);	/* 00000001 IBLEND_UNK00 */
+		xf_emit(ctx, 8, 1);	/* 00000007 IBLEND_EQUATION_RGB */
+		xf_emit(ctx, 8, 2);	/* 0000001f IBLEND_FUNC_SRC_RGB */
+		xf_emit(ctx, 8, 1);	/* 0000001f IBLEND_FUNC_DST_RGB */
+		xf_emit(ctx, 8, 1);	/* 00000007 IBLEND_EQUATION_ALPHA */
+		xf_emit(ctx, 8, 2);	/* 0000001f IBLEND_FUNC_SRC_ALPHA */
+		xf_emit(ctx, 8, 1);	/* 0000001f IBLEND_FUNC_DST_ALPHA */
+		xf_emit(ctx, 1, 0);	/* 00000001 tesla UNK15C4 */
+		xf_emit(ctx, 1, 0);	/* 00000001 */
+		xf_emit(ctx, 1, 0);	/* 00000001 tesla UNK1140 */
 	}
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0x5b, 0);
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f DST_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 DST_LINEAR */
+	xf_emit(ctx, 1, 0);		/* 00000007 PATTERN_COLOR_FORMAT */
+	xf_emit(ctx, 2, 0);		/* ffffffff PATTERN_MONO_COLOR */
+	xf_emit(ctx, 1, 0);		/* 00000001 PATTERN_MONO_FORMAT */
+	xf_emit(ctx, 2, 0);		/* ffffffff PATTERN_MONO_BITMAP */
+	xf_emit(ctx, 1, 0);		/* 00000003 PATTERN_SELECT */
+	xf_emit(ctx, 1, 0);		/* 000000ff ROP */
+	xf_emit(ctx, 1, 0);		/* ffffffff BETA1 */
+	xf_emit(ctx, 1, 0);		/* ffffffff BETA4 */
+	xf_emit(ctx, 1, 0);		/* 00000007 OPERATION */
+	xf_emit(ctx, 0x50, 0);		/* 10x ffffff, ffffff, ffffff, ffffff, 3 PATTERN */
 }
 
 static void
-nv50_graph_construct_xfer_tp_x1(struct nouveau_grctx *ctx)
+nv50_graph_construct_xfer_unk84xx(struct nouveau_grctx *ctx)
 {
 	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
 	int magic3;
-	if (dev_priv->chipset == 0x50)
+	switch (dev_priv->chipset) {
+	case 0x50:
 		magic3 = 0x1000;
-	else if (dev_priv->chipset == 0x86 || dev_priv->chipset == 0x98 || dev_priv->chipset >= 0xa8)
+		break;
+	case 0x86:
+	case 0x98:
+	case 0xa8:
+	case 0xaa:
+	case 0xac:
+	case 0xaf:
 		magic3 = 0x1e00;
-	else
+		break;
+	default:
 		magic3 = 0;
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 4);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 0x24, 0);
+	}
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 7f/ff[NVA0+] VP_REG_ALLOC_RESULT */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 0);		/* 111/113[NVA0+] */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 0x1f, 0);	/* ffffffff */
 	else if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 0x14, 0);
+		xf_emit(ctx, 0x0f, 0);	/* ffffffff */
 	else
-		xf_emit(ctx, 0x15, 0);
-	xf_emit(ctx, 2, 4);
+		xf_emit(ctx, 0x10, 0);	/* fffffff VP_RESULT_MAP_1 up */
+	xf_emit(ctx, 2, 0);		/* f/1f[NVA3], fffffff/ffffffff[NVA0+] */
+	xf_emit(ctx, 1, 4);		/* 7f/ff VP_REG_ALLOC_RESULT */
+	xf_emit(ctx, 1, 4);		/* 7f/ff VP_RESULT_MAP_SIZE */
 	if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 1, 0x03020100);
+		xf_emit(ctx, 1, 0x03020100);	/* ffffffff */
 	else
-		xf_emit(ctx, 1, 0x00608080);
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 2, 4);
-	xf_emit(ctx, 1, 0x80);
+		xf_emit(ctx, 1, 0x00608080);	/* fffffff VP_RESULT_MAP_0 */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 2, 0);		/* 111/113, 7f/ff */
+	xf_emit(ctx, 1, 4);		/* 7f/ff VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_REG_ALLOC_RESULT */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0x80);		/* 0000ffff GP_VERTEX_OUTPUT_COUNT */
 	if (magic3)
-		xf_emit(ctx, 1, magic3);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 0x24, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0x80);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0x03020100);
-	xf_emit(ctx, 1, 3);
+		xf_emit(ctx, 1, magic3);	/* 00007fff tesla UNK141C */
+	xf_emit(ctx, 1, 4);		/* 7f/ff VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 0);		/* 111/113 */
+	xf_emit(ctx, 0x1f, 0);		/* ffffffff GP_RESULT_MAP_1 up */
+	xf_emit(ctx, 1, 0);		/* 0000001f */
+	xf_emit(ctx, 1, 0);		/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_REG_ALLOC_RESULT */
+	xf_emit(ctx, 1, 0x80);		/* 0000ffff GP_VERTEX_OUTPUT_COUNT */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0x03020100);	/* ffffffff GP_RESULT_MAP_0 */
+	xf_emit(ctx, 1, 3);		/* 00000003 GP_OUTPUT_PRIMITIVE_TYPE */
 	if (magic3)
-		xf_emit(ctx, 1, magic3);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 3);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 4);
+		xf_emit(ctx, 1, magic3);	/* 7fff tesla UNK141C */
+	xf_emit(ctx, 1, 4);		/* 7f/ff VP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 0);		/* 00000001 PROVOKING_VERTEX_LAST */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 0);		/* 111/113 */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 3);		/* 00000003 GP_OUTPUT_PRIMITIVE_TYPE */
+	xf_emit(ctx, 1, 0);		/* 00000001 PROVOKING_VERTEX_LAST */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 0);		/* 00000003 tesla UNK13A0 */
+	xf_emit(ctx, 1, 4);		/* 7f/ff VP_REG_ALLOC_RESULT */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+	xf_emit(ctx, 1, 0);		/* 111/113 */
 	if (dev_priv->chipset == 0x94 || dev_priv->chipset == 0x96)
-		xf_emit(ctx, 0x1024, 0);
+		xf_emit(ctx, 0x1020, 0);	/* 4 x (0x400 x 0xffffffff, ff, 0, 0, 0, 4 x ffffffff) */
 	else if (dev_priv->chipset < 0xa0)
-		xf_emit(ctx, 0xa24, 0);
-	else if (dev_priv->chipset == 0xa0 || dev_priv->chipset >= 0xaa)
-		xf_emit(ctx, 0x214, 0);
+		xf_emit(ctx, 0xa20, 0);	/* 4 x (0x280 x 0xffffffff, ff, 0, 0, 0, 4 x ffffffff) */
+	else if (!IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 0x210, 0);	/* ffffffff */
 	else
-		xf_emit(ctx, 0x414, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 3);
-	xf_emit(ctx, 2, 0);
+		xf_emit(ctx, 0x410, 0);	/* ffffffff */
+	xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+	xf_emit(ctx, 1, 4);		/* 000000ff GP_RESULT_MAP_SIZE */
+	xf_emit(ctx, 1, 3);		/* 00000003 GP_OUTPUT_PRIMITIVE_TYPE */
+	xf_emit(ctx, 1, 0);		/* 00000001 PROVOKING_VERTEX_LAST */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
 }
 
 static void
-nv50_graph_construct_xfer_tp_x2(struct nouveau_grctx *ctx)
+nv50_graph_construct_xfer_tprop(struct nouveau_grctx *ctx)
 {
 	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
 	int magic1, magic2;
 	if (dev_priv->chipset == 0x50) {
 		magic1 = 0x3ff;
 		magic2 = 0x00003e60;
-	} else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) {
+	} else if (!IS_NVA3F(dev_priv->chipset)) {
 		magic1 = 0x7ff;
 		magic2 = 0x001ffe67;
 	} else {
 		magic1 = 0x7ff;
 		magic2 = 0x00087e67;
 	}
-	xf_emit(ctx, 3, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0xc, 0);
-	xf_emit(ctx, 1, 0xf);
-	xf_emit(ctx, 0xb, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 4, 0xffff);
-	xf_emit(ctx, 8, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 5, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 2, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		xf_emit(ctx, 1, 3);
-		xf_emit(ctx, 1, 0);
-	} else if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0xa, 0);
-	xf_emit(ctx, 2, 1);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 2, 1);
-	xf_emit(ctx, 1, 2);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 0x18, 1);
-		xf_emit(ctx, 8, 2);
-		xf_emit(ctx, 8, 1);
-		xf_emit(ctx, 8, 2);
-		xf_emit(ctx, 8, 1);
-		xf_emit(ctx, 1, 0);
-	}
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x0fac6881);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 3, 0xcf);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0xa, 0);
-	xf_emit(ctx, 2, 1);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 2, 1);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 8, 1);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x0fac6881);
-	xf_emit(ctx, 1, 0xf);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, magic2);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x11);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 2, 1);
-	else
-		xf_emit(ctx, 1, 1);
-	if(dev_priv->chipset == 0x50)
-		xf_emit(ctx, 1, 0);
-	else
-		xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 5, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x0fac6881);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, magic1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 2, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 0x28, 0);
-	xf_emit(ctx, 8, 8);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x0fac6881);
-	xf_emit(ctx, 8, 0x400);
-	xf_emit(ctx, 8, 0x300);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0xf);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x20);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 1, 0x100);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x40);
-	xf_emit(ctx, 1, 0x100);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 3);
-	xf_emit(ctx, 4, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, magic2);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 1, 0x0fac6881);
-	xf_emit(ctx, 9, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x400);
-	xf_emit(ctx, 1, 0x300);
-	xf_emit(ctx, 1, 0x1001);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 4, 0);
-	else
-		xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x0fac6881);
-	xf_emit(ctx, 1, 0xf);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		xf_emit(ctx, 0x15, 0);
-		xf_emit(ctx, 1, 1);
-		xf_emit(ctx, 3, 0);
-	} else
-		xf_emit(ctx, 0x17, 0);
-	if (dev_priv->chipset >= 0xa0)
-		xf_emit(ctx, 1, 0x0fac6881);
-	xf_emit(ctx, 1, magic2);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 2, 1);
-	xf_emit(ctx, 3, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 2, 1);
-	else
-		xf_emit(ctx, 1, 1);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 2, 0);
-	else if (dev_priv->chipset != 0x50)
-		xf_emit(ctx, 1, 0);
-}
-
-static void
-nv50_graph_construct_xfer_tp_x3(struct nouveau_grctx *ctx)
-{
-	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	if (dev_priv->chipset == 0x50)
-		xf_emit(ctx, 2, 0);
-	else
-		xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0x2a712488);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x4085c000);
-	xf_emit(ctx, 1, 0x40);
-	xf_emit(ctx, 1, 0x100);
-	xf_emit(ctx, 1, 0x10100);
-	xf_emit(ctx, 1, 0x02800000);
-}
-
-static void
-nv50_graph_construct_xfer_tp_x4(struct nouveau_grctx *ctx)
-{
-	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	xf_emit(ctx, 2, 0x04e3bfdf);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x00ffff00);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 2, 1);
-	else
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 0x00ffff00);
-	xf_emit(ctx, 8, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0x30201000);
-	xf_emit(ctx, 1, 0x70605040);
-	xf_emit(ctx, 1, 0xb8a89888);
-	xf_emit(ctx, 1, 0xf8e8d8c8);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x1a);
-}
-
-static void
-nv50_graph_construct_xfer_tp_x5(struct nouveau_grctx *ctx)
-{
-	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 0xfac6881);
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 2, 1);
-	xf_emit(ctx, 2, 0);
-	xf_emit(ctx, 1, 1);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 0xb, 0);
-	else
-		xf_emit(ctx, 0xa, 0);
-	xf_emit(ctx, 8, 1);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0xfac6881);
-	xf_emit(ctx, 1, 0xf);
-	xf_emit(ctx, 7, 0);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 1, 1);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		xf_emit(ctx, 6, 0);
-		xf_emit(ctx, 1, 1);
-		xf_emit(ctx, 6, 0);
+	xf_emit(ctx, 1, 0);		/* 00000007 ALPHA_TEST_FUNC */
+	xf_emit(ctx, 1, 0);		/* ffffffff ALPHA_TEST_REF */
+	xf_emit(ctx, 1, 0);		/* 00000001 ALPHA_TEST_ENABLE */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 1);	/* 0000000f UNK16A0 */
+	xf_emit(ctx, 1, 0);		/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_BACK_MASK */
+	xf_emit(ctx, 3, 0);		/* 00000007 STENCIL_BACK_OP_FAIL, ZFAIL, ZPASS */
+	xf_emit(ctx, 4, 0);		/* ffffffff BLEND_COLOR */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK19C0 */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK0FDC */
+	xf_emit(ctx, 1, 0xf);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 7, 0);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 LOGIC_OP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ff[NV50]/3ff[NV84+] */
+	xf_emit(ctx, 1, 4);		/* 00000007 FP_CONTROL */
+	xf_emit(ctx, 4, 0xffff);	/* 0000ffff MSAA_MASK */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_MASK */
+	xf_emit(ctx, 3, 0);		/* 00000007 STENCIL_FRONT_OP_FAIL, ZFAIL, ZPASS */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_BACK_ENABLE */
+	xf_emit(ctx, 2, 0);		/* 00007fff WINDOW_OFFSET_XY */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK19CC */
+	xf_emit(ctx, 1, 0);		/* 7 */
+	xf_emit(ctx, 1, 0);		/* 00000001 SAMPLECNT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff COLOR_KEY */
+	xf_emit(ctx, 1, 0);		/* 00000001 COLOR_KEY_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000007 COLOR_KEY_FORMAT */
+	xf_emit(ctx, 2, 0);		/* ffffffff SIFC_BITMAP_COLOR */
+	xf_emit(ctx, 1, 1);		/* 00000001 SIFC_BITMAP_WRITE_BIT0_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000007 ALPHA_TEST_FUNC */
+	xf_emit(ctx, 1, 0);		/* 00000001 ALPHA_TEST_ENABLE */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 1, 3);	/* 00000003 tesla UNK16B4 */
+		xf_emit(ctx, 1, 0);	/* 00000003 */
+		xf_emit(ctx, 1, 0);	/* 00000003 tesla UNK1298 */
+	} else if (dev_priv->chipset >= 0xa0) {
+		xf_emit(ctx, 1, 1);	/* 00000001 tesla UNK16B4 */
+		xf_emit(ctx, 1, 0);	/* 00000003 */
 	} else {
-		xf_emit(ctx, 0xb, 0);
+		xf_emit(ctx, 1, 0);	/* 00000003 MULTISAMPLE_CTRL */
 	}
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 8, 0);		/* 00000001 BLEND_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 0000001f BLEND_FUNC_DST_ALPHA */
+	xf_emit(ctx, 1, 1);		/* 00000007 BLEND_EQUATION_ALPHA */
+	xf_emit(ctx, 1, 2);		/* 0000001f BLEND_FUNC_SRC_ALPHA */
+	xf_emit(ctx, 1, 1);		/* 0000001f BLEND_FUNC_DST_RGB */
+	xf_emit(ctx, 1, 1);		/* 00000007 BLEND_EQUATION_RGB */
+	xf_emit(ctx, 1, 2);		/* 0000001f BLEND_FUNC_SRC_RGB */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 1, 0);	/* 00000001 UNK12E4 */
+		xf_emit(ctx, 8, 1);	/* 00000007 IBLEND_EQUATION_RGB */
+		xf_emit(ctx, 8, 1);	/* 00000007 IBLEND_EQUATION_ALPHA */
+		xf_emit(ctx, 8, 1);	/* 00000001 IBLEND_UNK00 */
+		xf_emit(ctx, 8, 2);	/* 0000001f IBLEND_SRC_RGB */
+		xf_emit(ctx, 8, 1);	/* 0000001f IBLEND_DST_RGB */
+		xf_emit(ctx, 8, 2);	/* 0000001f IBLEND_SRC_ALPHA */
+		xf_emit(ctx, 8, 1);	/* 0000001f IBLEND_DST_ALPHA */
+		xf_emit(ctx, 1, 0);	/* 00000001 UNK1140 */
+	}
+	xf_emit(ctx, 1, 1);		/* 00000001 UNK133C */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 7, 0);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 1, 0);		/* 00000001 LOGIC_OP_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ff/3ff */
+	xf_emit(ctx, 1, 4);		/* 00000007 FP_CONTROL */
+	xf_emit(ctx, 1, 0);		/* 00000003 UNK0F90 */
+	xf_emit(ctx, 1, 0);		/* 00000001 FRAMEBUFFER_SRGB */
+	xf_emit(ctx, 1, 0);		/* 7 */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f DST_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 DST_LINEAR */
+	xf_emit(ctx, 1, 0);		/* 00000007 OPERATION */
+	xf_emit(ctx, 1, 0xcf);		/* 000000ff SIFC_FORMAT */
+	xf_emit(ctx, 1, 0xcf);		/* 000000ff DRAW_COLOR_FORMAT */
+	xf_emit(ctx, 1, 0xcf);		/* 000000ff SRC_FORMAT */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
+	xf_emit(ctx, 1, 0);		/* 7/f[NVA3] MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 8, 0);		/* 00000001 BLEND_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 0000001f BLEND_FUNC_DST_ALPHA */
+	xf_emit(ctx, 1, 1);		/* 00000007 BLEND_EQUATION_ALPHA */
+	xf_emit(ctx, 1, 2);		/* 0000001f BLEND_FUNC_SRC_ALPHA */
+	xf_emit(ctx, 1, 1);		/* 0000001f BLEND_FUNC_DST_RGB */
+	xf_emit(ctx, 1, 1);		/* 00000007 BLEND_EQUATION_RGB */
+	xf_emit(ctx, 1, 2);		/* 0000001f BLEND_FUNC_SRC_RGB */
+	xf_emit(ctx, 1, 1);		/* 00000001 UNK133C */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 8, 1);		/* 00000001 UNK19E0 */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 7, 0);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 1, 0xf);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 7, 0);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 1, magic2);	/* 001fffff tesla UNK0F78 */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_BOUNDS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f DST_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 DST_LINEAR */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+	if(dev_priv->chipset == 0x50)
+		xf_emit(ctx, 1, 0);	/* ff */
+	else
+		xf_emit(ctx, 3, 0);	/* 1, 7, 3ff */
+	xf_emit(ctx, 1, 4);		/* 00000007 FP_CONTROL */
+	xf_emit(ctx, 1, 0);		/* 00000003 UNK0F90 */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	xf_emit(ctx, 1, 0);		/* 00000001 SAMPLECNT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
+	xf_emit(ctx, 1, 0);		/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 7, 0);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_BOUNDS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f DST_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 DST_LINEAR */
+	xf_emit(ctx, 1, 0);		/* 000fffff BLIT_DU_DX_FRACT */
+	xf_emit(ctx, 1, 1);		/* 0001ffff BLIT_DU_DX_INT */
+	xf_emit(ctx, 1, 0);		/* 000fffff BLIT_DV_DY_FRACT */
+	xf_emit(ctx, 1, 1);		/* 0001ffff BLIT_DV_DY_INT */
+	xf_emit(ctx, 1, 0);		/* ff/3ff */
+	xf_emit(ctx, 1, magic1);	/* 3ff/7ff tesla UNK0D68 */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK15B4 */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+	xf_emit(ctx, 8, 0);		/* 0000ffff DMA_COLOR */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_GLOBAL */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_LOCAL */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_STACK */
+	xf_emit(ctx, 1, 0);		/* ff/3ff */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_DST */
+	xf_emit(ctx, 1, 0);		/* 7 */
+	xf_emit(ctx, 1, 0);		/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 8, 0);		/* 000000ff RT_ADDRESS_HIGH */
+	xf_emit(ctx, 8, 0);		/* ffffffff RT_LAYER_STRIDE */
+	xf_emit(ctx, 8, 0);		/* ffffffff RT_ADDRESS_LOW */
+	xf_emit(ctx, 8, 8);		/* 0000007f RT_TILE_MODE */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 7, 0);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 8, 0x400);		/* 0fffffff RT_HORIZ */
+	xf_emit(ctx, 8, 0x300);		/* 0000ffff RT_VERT */
+	xf_emit(ctx, 1, 1);		/* 00001fff RT_ARRAY_MODE */
+	xf_emit(ctx, 1, 0xf);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 7, 0);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 1, 0x20);		/* 00000fff DST_TILE_MODE */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f DST_FORMAT */
+	xf_emit(ctx, 1, 0x100);		/* 0001ffff DST_HEIGHT */
+	xf_emit(ctx, 1, 0);		/* 000007ff DST_LAYER */
+	xf_emit(ctx, 1, 1);		/* 00000001 DST_LINEAR */
+	xf_emit(ctx, 1, 0);		/* ffffffff DST_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0);		/* 000000ff DST_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0x40);		/* 0007ffff DST_PITCH */
+	xf_emit(ctx, 1, 0x100);		/* 0001ffff DST_WIDTH */
+	xf_emit(ctx, 1, 0);		/* 0000ffff */
+	xf_emit(ctx, 1, 3);		/* 00000003 tesla UNK15AC */
+	xf_emit(ctx, 1, 0);		/* ff/3ff */
+	xf_emit(ctx, 1, 0);		/* 0001ffff GP_BUILTIN_RESULT_EN */
+	xf_emit(ctx, 1, 0);		/* 00000003 UNK0F90 */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+	xf_emit(ctx, 1, magic2);	/* 001fffff tesla UNK0F78 */
+	xf_emit(ctx, 1, 0);		/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, 2);		/* 00000003 tesla UNK143C */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_ZETA */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_BOUNDS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 2, 0);		/* ffff, ff/3ff */
+	xf_emit(ctx, 1, 0);		/* 0001ffff GP_BUILTIN_RESULT_EN */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 000000ff STENCIL_FRONT_MASK */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK15B4 */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	xf_emit(ctx, 1, 0);		/* ffffffff ZETA_LAYER_STRIDE */
+	xf_emit(ctx, 1, 0);		/* 000000ff ZETA_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);		/* ffffffff ZETA_ADDRESS_LOW */
+	xf_emit(ctx, 1, 4);		/* 00000007 ZETA_TILE_MODE */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	xf_emit(ctx, 1, 0x400);		/* 0fffffff ZETA_HORIZ */
+	xf_emit(ctx, 1, 0x300);		/* 0000ffff ZETA_VERT */
+	xf_emit(ctx, 1, 0x1001);	/* 00001fff ZETA_ARRAY_MODE */
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
+	xf_emit(ctx, 1, 0);		/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 0);	/* 00000001 */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 7, 0);		/* 3f/7f RT_FORMAT */
+	xf_emit(ctx, 1, 0x0fac6881);	/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 1, 0xf);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 7, 0);		/* 0000000f COLOR_MASK */
+	xf_emit(ctx, 1, 0);		/* ff/3ff */
+	xf_emit(ctx, 8, 0);		/* 00000001 BLEND_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000003 UNK0F90 */
+	xf_emit(ctx, 1, 0);		/* 00000001 FRAMEBUFFER_SRGB */
+	xf_emit(ctx, 1, 0);		/* 7 */
+	xf_emit(ctx, 1, 0);		/* 00000001 LOGIC_OP_ENABLE */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 1, 0);	/* 00000001 UNK1140 */
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+	}
+	xf_emit(ctx, 1, 0);		/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK1534 */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	if (dev_priv->chipset >= 0xa0)
+		xf_emit(ctx, 1, 0x0fac6881);	/* fffffff */
+	xf_emit(ctx, 1, magic2);	/* 001fffff tesla UNK0F78 */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_BOUNDS_EN */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE_ENABLE */
+	xf_emit(ctx, 1, 0x11);		/* 3f/7f DST_FORMAT */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK0FB0 */
+	xf_emit(ctx, 1, 0);		/* ff/3ff */
+	xf_emit(ctx, 1, 4);		/* 00000007 FP_CONTROL */
+	xf_emit(ctx, 1, 0);		/* 00000001 STENCIL_FRONT_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK15B4 */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK19CC */
+	xf_emit(ctx, 1, 0);		/* 00000007 */
+	xf_emit(ctx, 1, 0);		/* 00000001 SAMPLECNT_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 0000000f ZETA_FORMAT */
+	xf_emit(ctx, 1, 1);		/* 00000001 ZETA_ENABLE */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+		xf_emit(ctx, 1, 0);	/* 0000000f tesla UNK15C8 */
+	}
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A3C */
+	if (dev_priv->chipset >= 0xa0) {
+		xf_emit(ctx, 3, 0);		/* 7/f, 1, ffff0ff3 */
+		xf_emit(ctx, 1, 0xfac6881);	/* fffffff */
+		xf_emit(ctx, 4, 0);		/* 1, 1, 1, 3ff */
+		xf_emit(ctx, 1, 4);		/* 7 */
+		xf_emit(ctx, 1, 0);		/* 1 */
+		xf_emit(ctx, 2, 1);		/* 1 */
+		xf_emit(ctx, 2, 0);		/* 7, f */
+		xf_emit(ctx, 1, 1);		/* 1 */
+		xf_emit(ctx, 1, 0);		/* 7/f */
+		if (IS_NVA3F(dev_priv->chipset))
+			xf_emit(ctx, 0x9, 0);	/* 1 */
+		else
+			xf_emit(ctx, 0x8, 0);	/* 1 */
+		xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+		xf_emit(ctx, 8, 1);		/* 1 */
+		xf_emit(ctx, 1, 0x11);		/* 7f */
+		xf_emit(ctx, 7, 0);		/* 7f */
+		xf_emit(ctx, 1, 0xfac6881);	/* fffffff */
+		xf_emit(ctx, 1, 0xf);		/* f */
+		xf_emit(ctx, 7, 0);		/* f */
+		xf_emit(ctx, 1, 0x11);		/* 7f */
+		xf_emit(ctx, 1, 1);		/* 1 */
+		xf_emit(ctx, 5, 0);		/* 1, 7, 3ff, 3, 7 */
+		if (IS_NVA3F(dev_priv->chipset)) {
+			xf_emit(ctx, 1, 0);	/* 00000001 UNK1140 */
+			xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+		}
+	}
+}
+
+static void
+nv50_graph_construct_xfer_tex(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	xf_emit(ctx, 2, 0);		/* 1 LINKED_TSC. yes, 2. */
+	if (dev_priv->chipset != 0x50)
+		xf_emit(ctx, 1, 0);	/* 3 */
+	xf_emit(ctx, 1, 1);		/* 1ffff BLIT_DU_DX_INT */
+	xf_emit(ctx, 1, 0);		/* fffff BLIT_DU_DX_FRACT */
+	xf_emit(ctx, 1, 1);		/* 1ffff BLIT_DV_DY_INT */
+	xf_emit(ctx, 1, 0);		/* fffff BLIT_DV_DY_FRACT */
+	if (dev_priv->chipset == 0x50)
+		xf_emit(ctx, 1, 0);	/* 3 BLIT_CONTROL */
+	else
+		xf_emit(ctx, 2, 0);	/* 3ff, 1 */
+	xf_emit(ctx, 1, 0x2a712488);	/* ffffffff SRC_TIC_0 */
+	xf_emit(ctx, 1, 0);		/* ffffffff SRC_TIC_1 */
+	xf_emit(ctx, 1, 0x4085c000);	/* ffffffff SRC_TIC_2 */
+	xf_emit(ctx, 1, 0x40);		/* ffffffff SRC_TIC_3 */
+	xf_emit(ctx, 1, 0x100);		/* ffffffff SRC_TIC_4 */
+	xf_emit(ctx, 1, 0x10100);	/* ffffffff SRC_TIC_5 */
+	xf_emit(ctx, 1, 0x02800000);	/* ffffffff SRC_TIC_6 */
+	xf_emit(ctx, 1, 0);		/* ffffffff SRC_TIC_7 */
+	if (dev_priv->chipset == 0x50) {
+		xf_emit(ctx, 1, 0);	/* 00000001 turing UNK358 */
+		xf_emit(ctx, 1, 0);	/* ffffffff tesla UNK1A34? */
+		xf_emit(ctx, 1, 0);	/* 00000003 turing UNK37C tesla UNK1690 */
+		xf_emit(ctx, 1, 0);	/* 00000003 BLIT_CONTROL */
+		xf_emit(ctx, 1, 0);	/* 00000001 turing UNK32C tesla UNK0F94 */
+	} else if (!IS_NVAAF(dev_priv->chipset)) {
+		xf_emit(ctx, 1, 0);	/* ffffffff tesla UNK1A34? */
+		xf_emit(ctx, 1, 0);	/* 00000003 */
+		xf_emit(ctx, 1, 0);	/* 000003ff */
+		xf_emit(ctx, 1, 0);	/* 00000003 */
+		xf_emit(ctx, 1, 0);	/* 000003ff */
+		xf_emit(ctx, 1, 0);	/* 00000003 tesla UNK1664 / turing UNK03E8 */
+		xf_emit(ctx, 1, 0);	/* 00000003 */
+		xf_emit(ctx, 1, 0);	/* 000003ff */
+	} else {
+		xf_emit(ctx, 0x6, 0);
+	}
+	xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A34 */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_TEXTURE */
+	xf_emit(ctx, 1, 0);		/* 0000ffff DMA_SRC */
+}
+
+static void
+nv50_graph_construct_xfer_unk8cxx(struct nouveau_grctx *ctx)
+{
+	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 2, 0);		/* 7, ffff0ff3 */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE */
+	xf_emit(ctx, 1, 0x04e3bfdf);	/* ffffffff UNK0D64 */
+	xf_emit(ctx, 1, 0x04e3bfdf);	/* ffffffff UNK0DF4 */
+	xf_emit(ctx, 1, 1);		/* 00000001 UNK15B4 */
+	xf_emit(ctx, 1, 0);		/* 00000001 LINE_STIPPLE_ENABLE */
+	xf_emit(ctx, 1, 0x00ffff00);	/* 00ffffff LINE_STIPPLE_PATTERN */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK0F98 */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 1);	/* 0000001f tesla UNK169C */
+	xf_emit(ctx, 1, 0);		/* 00000003 tesla UNK1668 */
+	xf_emit(ctx, 1, 0);		/* 00000001 LINE_STIPPLE_ENABLE */
+	xf_emit(ctx, 1, 0x00ffff00);	/* 00ffffff LINE_STIPPLE_PATTERN */
+	xf_emit(ctx, 1, 0);		/* 00000001 POLYGON_SMOOTH_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 UNK1534 */
+	xf_emit(ctx, 1, 0);		/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 1, 0);		/* 00000001 tesla UNK1658 */
+	xf_emit(ctx, 1, 0);		/* 00000001 LINE_SMOOTH_ENABLE */
+	xf_emit(ctx, 1, 0);		/* ffff0ff3 */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);		/* 00000001 DEPTH_WRITE */
+	xf_emit(ctx, 1, 1);		/* 00000001 UNK15B4 */
+	xf_emit(ctx, 1, 0);		/* 00000001 POINT_SPRITE_ENABLE */
+	xf_emit(ctx, 1, 1);		/* 00000001 tesla UNK165C */
+	xf_emit(ctx, 1, 0x30201000);	/* ffffffff tesla UNK1670 */
+	xf_emit(ctx, 1, 0x70605040);	/* ffffffff tesla UNK1670 */
+	xf_emit(ctx, 1, 0xb8a89888);	/* ffffffff tesla UNK1670 */
+	xf_emit(ctx, 1, 0xf8e8d8c8);	/* ffffffff tesla UNK1670 */
+	xf_emit(ctx, 1, 0);		/* 00000001 VERTEX_TWO_SIDE_ENABLE */
+	xf_emit(ctx, 1, 0x1a);		/* 0000001f POLYGON_MODE */
 }
 
 static void
@@ -2193,108 +3102,136 @@ nv50_graph_construct_xfer_tp(struct nouveau_grctx *ctx)
 {
 	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
 	if (dev_priv->chipset < 0xa0) {
-		nv50_graph_construct_xfer_tp_x1(ctx);
-		nv50_graph_construct_xfer_tp_x2(ctx);
-		nv50_graph_construct_xfer_tp_x3(ctx);
-		if (dev_priv->chipset == 0x50)
-			xf_emit(ctx, 0xf, 0);
-		else
-			xf_emit(ctx, 0x12, 0);
-		nv50_graph_construct_xfer_tp_x4(ctx);
+		nv50_graph_construct_xfer_unk84xx(ctx);
+		nv50_graph_construct_xfer_tprop(ctx);
+		nv50_graph_construct_xfer_tex(ctx);
+		nv50_graph_construct_xfer_unk8cxx(ctx);
 	} else {
-		nv50_graph_construct_xfer_tp_x3(ctx);
-		if (dev_priv->chipset < 0xaa)
-			xf_emit(ctx, 0xc, 0);
-		else
-			xf_emit(ctx, 0xa, 0);
-		nv50_graph_construct_xfer_tp_x2(ctx);
-		nv50_graph_construct_xfer_tp_x5(ctx);
-		nv50_graph_construct_xfer_tp_x4(ctx);
-		nv50_graph_construct_xfer_tp_x1(ctx);
+		nv50_graph_construct_xfer_tex(ctx);
+		nv50_graph_construct_xfer_tprop(ctx);
+		nv50_graph_construct_xfer_unk8cxx(ctx);
+		nv50_graph_construct_xfer_unk84xx(ctx);
 	}
 }
 
 static void
-nv50_graph_construct_xfer_tp2(struct nouveau_grctx *ctx)
+nv50_graph_construct_xfer_mpc(struct nouveau_grctx *ctx)
 {
 	struct drm_nouveau_private *dev_priv = ctx->dev->dev_private;
-	int i, mpcnt;
-	if (dev_priv->chipset == 0x98 || dev_priv->chipset == 0xaa)
-		mpcnt = 1;
-	else if (dev_priv->chipset < 0xa0 || dev_priv->chipset >= 0xa8)
-		mpcnt = 2;
-	else
-		mpcnt = 3;
+	int i, mpcnt = 2;
+	switch (dev_priv->chipset) {
+		case 0x98:
+		case 0xaa:
+			mpcnt = 1;
+			break;
+		case 0x50:
+		case 0x84:
+		case 0x86:
+		case 0x92:
+		case 0x94:
+		case 0x96:
+		case 0xa8:
+		case 0xac:
+			mpcnt = 2;
+			break;
+		case 0xa0:
+		case 0xa3:
+		case 0xa5:
+		case 0xaf:
+			mpcnt = 3;
+			break;
+	}
 	for (i = 0; i < mpcnt; i++) {
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 1, 0x80);
-		xf_emit(ctx, 1, 0x80007004);
-		xf_emit(ctx, 1, 0x04000400);
+		xf_emit(ctx, 1, 0);		/* ff */
+		xf_emit(ctx, 1, 0x80);		/* ffffffff tesla UNK1404 */
+		xf_emit(ctx, 1, 0x80007004);	/* ffffffff tesla UNK12B0 */
+		xf_emit(ctx, 1, 0x04000400);	/* ffffffff */
 		if (dev_priv->chipset >= 0xa0)
-			xf_emit(ctx, 1, 0xc0);
-		xf_emit(ctx, 1, 0x1000);
-		xf_emit(ctx, 2, 0);
-		if (dev_priv->chipset == 0x86 || dev_priv->chipset == 0x98 || dev_priv->chipset >= 0xa8) {
-			xf_emit(ctx, 1, 0xe00);
-			xf_emit(ctx, 1, 0x1e00);
+			xf_emit(ctx, 1, 0xc0);	/* 00007fff tesla UNK152C */
+		xf_emit(ctx, 1, 0x1000);	/* 0000ffff tesla UNK0D60 */
+		xf_emit(ctx, 1, 0);		/* ff/3ff */
+		xf_emit(ctx, 1, 0);		/* ffffffff tesla UNK1A30 */
+		if (dev_priv->chipset == 0x86 || dev_priv->chipset == 0x98 || dev_priv->chipset == 0xa8 || IS_NVAAF(dev_priv->chipset)) {
+			xf_emit(ctx, 1, 0xe00);		/* 7fff */
+			xf_emit(ctx, 1, 0x1e00);	/* 7fff */
 		}
-		xf_emit(ctx, 1, 1);
-		xf_emit(ctx, 2, 0);
+		xf_emit(ctx, 1, 1);		/* 000000ff VP_REG_ALLOC_TEMP */
+		xf_emit(ctx, 1, 0);		/* 00000001 LINKED_TSC */
+		xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
 		if (dev_priv->chipset == 0x50)
-			xf_emit(ctx, 2, 0x1000);
-		xf_emit(ctx, 1, 1);
-		xf_emit(ctx, 1, 0);
-		xf_emit(ctx, 1, 4);
-		xf_emit(ctx, 1, 2);
-		if (dev_priv->chipset >= 0xaa)
-			xf_emit(ctx, 0xb, 0);
+			xf_emit(ctx, 2, 0x1000);	/* 7fff tesla UNK141C */
+		xf_emit(ctx, 1, 1);		/* 000000ff GP_REG_ALLOC_TEMP */
+		xf_emit(ctx, 1, 0);		/* 00000001 GP_ENABLE */
+		xf_emit(ctx, 1, 4);		/* 000000ff FP_REG_ALLOC_TEMP */
+		xf_emit(ctx, 1, 2);		/* 00000003 REG_MODE */
+		if (IS_NVAAF(dev_priv->chipset))
+			xf_emit(ctx, 0xb, 0);	/* RO */
 		else if (dev_priv->chipset >= 0xa0)
-			xf_emit(ctx, 0xc, 0);
+			xf_emit(ctx, 0xc, 0);	/* RO */
 		else
-			xf_emit(ctx, 0xa, 0);
+			xf_emit(ctx, 0xa, 0);	/* RO */
 	}
-	xf_emit(ctx, 1, 0x08100c12);
-	xf_emit(ctx, 1, 0);
+	xf_emit(ctx, 1, 0x08100c12);		/* 1fffffff FP_INTERPOLANT_CTRL */
+	xf_emit(ctx, 1, 0);			/* ff/3ff */
 	if (dev_priv->chipset >= 0xa0) {
-		xf_emit(ctx, 1, 0x1fe21);
+		xf_emit(ctx, 1, 0x1fe21);	/* 0003ffff tesla UNK0FAC */
 	}
-	xf_emit(ctx, 5, 0);
-	xf_emit(ctx, 4, 0xffff);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 2, 0x10001);
-	xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 0x1fe21);
-	xf_emit(ctx, 1, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 1);
-	xf_emit(ctx, 4, 0);
-	xf_emit(ctx, 1, 0x08100c12);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 8, 0);
-	xf_emit(ctx, 1, 0xfac6881);
-	xf_emit(ctx, 1, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa)
-		xf_emit(ctx, 1, 3);
-	xf_emit(ctx, 3, 0);
-	xf_emit(ctx, 1, 4);
-	xf_emit(ctx, 9, 0);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 2, 1);
-	xf_emit(ctx, 1, 2);
-	xf_emit(ctx, 3, 1);
-	xf_emit(ctx, 1, 0);
-	if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) {
-		xf_emit(ctx, 8, 2);
-		xf_emit(ctx, 0x10, 1);
-		xf_emit(ctx, 8, 2);
-		xf_emit(ctx, 0x18, 1);
-		xf_emit(ctx, 3, 0);
+	xf_emit(ctx, 3, 0);			/* 7fff, 0, 0 */
+	xf_emit(ctx, 1, 0);			/* 00000001 tesla UNK1534 */
+	xf_emit(ctx, 1, 0);			/* 7/f MULTISAMPLE_SAMPLES_LOG2 */
+	xf_emit(ctx, 4, 0xffff);		/* 0000ffff MSAA_MASK */
+	xf_emit(ctx, 1, 1);			/* 00000001 LANES32 */
+	xf_emit(ctx, 1, 0x10001);		/* 00ffffff BLOCK_ALLOC */
+	xf_emit(ctx, 1, 0x10001);		/* ffffffff BLOCKDIM_XY */
+	xf_emit(ctx, 1, 1);			/* 0000ffff BLOCKDIM_Z */
+	xf_emit(ctx, 1, 0);			/* ffffffff SHARED_SIZE */
+	xf_emit(ctx, 1, 0x1fe21);		/* 1ffff/3ffff[NVA0+] tesla UNk0FAC */
+	xf_emit(ctx, 1, 0);			/* ffffffff tesla UNK1A34 */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 1);		/* 0000001f tesla UNK169C */
+	xf_emit(ctx, 1, 0);			/* ff/3ff */
+	xf_emit(ctx, 1, 0);			/* 1 LINKED_TSC */
+	xf_emit(ctx, 1, 0);			/* ff FP_ADDRESS_HIGH */
+	xf_emit(ctx, 1, 0);			/* ffffffff FP_ADDRESS_LOW */
+	xf_emit(ctx, 1, 0x08100c12);		/* 1fffffff FP_INTERPOLANT_CTRL */
+	xf_emit(ctx, 1, 4);			/* 00000007 FP_CONTROL */
+	xf_emit(ctx, 1, 0);			/* 000000ff FRAG_COLOR_CLAMP_EN */
+	xf_emit(ctx, 1, 2);			/* 00000003 REG_MODE */
+	xf_emit(ctx, 1, 0x11);			/* 0000007f RT_FORMAT */
+	xf_emit(ctx, 7, 0);			/* 0000007f RT_FORMAT */
+	xf_emit(ctx, 1, 0);			/* 00000007 */
+	xf_emit(ctx, 1, 0xfac6881);		/* 0fffffff RT_CONTROL */
+	xf_emit(ctx, 1, 0);			/* 00000003 MULTISAMPLE_CTRL */
+	if (IS_NVA3F(dev_priv->chipset))
+		xf_emit(ctx, 1, 3);		/* 00000003 tesla UNK16B4 */
+	xf_emit(ctx, 1, 0);			/* 00000001 ALPHA_TEST_ENABLE */
+	xf_emit(ctx, 1, 0);			/* 00000007 ALPHA_TEST_FUNC */
+	xf_emit(ctx, 1, 0);			/* 00000001 FRAMEBUFFER_SRGB */
+	xf_emit(ctx, 1, 4);			/* ffffffff tesla UNK1400 */
+	xf_emit(ctx, 8, 0);			/* 00000001 BLEND_ENABLE */
+	xf_emit(ctx, 1, 0);			/* 00000001 LOGIC_OP_ENABLE */
+	xf_emit(ctx, 1, 2);			/* 0000001f BLEND_FUNC_SRC_RGB */
+	xf_emit(ctx, 1, 1);			/* 0000001f BLEND_FUNC_DST_RGB */
+	xf_emit(ctx, 1, 1);			/* 00000007 BLEND_EQUATION_RGB */
+	xf_emit(ctx, 1, 2);			/* 0000001f BLEND_FUNC_SRC_ALPHA */
+	xf_emit(ctx, 1, 1);			/* 0000001f BLEND_FUNC_DST_ALPHA */
+	xf_emit(ctx, 1, 1);			/* 00000007 BLEND_EQUATION_ALPHA */
+	xf_emit(ctx, 1, 1);			/* 00000001 UNK133C */
+	if (IS_NVA3F(dev_priv->chipset)) {
+		xf_emit(ctx, 1, 0);		/* 00000001 UNK12E4 */
+		xf_emit(ctx, 8, 2);		/* 0000001f IBLEND_FUNC_SRC_RGB */
+		xf_emit(ctx, 8, 1);		/* 0000001f IBLEND_FUNC_DST_RGB */
+		xf_emit(ctx, 8, 1);		/* 00000007 IBLEND_EQUATION_RGB */
+		xf_emit(ctx, 8, 2);		/* 0000001f IBLEND_FUNC_SRC_ALPHA */
+		xf_emit(ctx, 8, 1);		/* 0000001f IBLEND_FUNC_DST_ALPHA */
+		xf_emit(ctx, 8, 1);		/* 00000007 IBLEND_EQUATION_ALPHA */
+		xf_emit(ctx, 8, 1);		/* 00000001 IBLEND_UNK00 */
+		xf_emit(ctx, 1, 0);		/* 00000003 tesla UNK1928 */
+		xf_emit(ctx, 1, 0);		/* 00000001 UNK1140 */
 	}
-	xf_emit(ctx, 1, 4);
+	xf_emit(ctx, 1, 0);			/* 00000003 tesla UNK0F90 */
+	xf_emit(ctx, 1, 4);			/* 000000ff FP_RESULT_COUNT */
+	/* XXX: demagic this part some day */
 	if (dev_priv->chipset == 0x50)
 		xf_emit(ctx, 0x3a0, 0);
 	else if (dev_priv->chipset < 0x94)
@@ -2303,9 +3240,9 @@ nv50_graph_construct_xfer_tp2(struct nouveau_grctx *ctx)
 		xf_emit(ctx, 0x39f, 0);
 	else
 		xf_emit(ctx, 0x3a3, 0);
-	xf_emit(ctx, 1, 0x11);
-	xf_emit(ctx, 1, 0);
-	xf_emit(ctx, 1, 1);
+	xf_emit(ctx, 1, 0x11);			/* 3f/7f DST_FORMAT */
+	xf_emit(ctx, 1, 0);			/* 7 OPERATION */
+	xf_emit(ctx, 1, 1);			/* 1 DST_LINEAR */
 	xf_emit(ctx, 0x2d, 0);
 }
 
@@ -2323,52 +3260,56 @@ nv50_graph_construct_xfer2(struct nouveau_grctx *ctx)
 	if (dev_priv->chipset < 0xa0) {
 		for (i = 0; i < 8; i++) {
 			ctx->ctxvals_pos = offset + i;
+			/* that little bugger belongs to csched. No idea
+			 * what it's doing here. */
 			if (i == 0)
-				xf_emit(ctx, 1, 0x08100c12);
+				xf_emit(ctx, 1, 0x08100c12); /* FP_INTERPOLANT_CTRL */
 			if (units & (1 << i))
-				nv50_graph_construct_xfer_tp2(ctx);
+				nv50_graph_construct_xfer_mpc(ctx);
 			if ((ctx->ctxvals_pos-offset)/8 > size)
 				size = (ctx->ctxvals_pos-offset)/8;
 		}
 	} else {
 		/* Strand 0: TPs 0, 1 */
 		ctx->ctxvals_pos = offset;
-		xf_emit(ctx, 1, 0x08100c12);
+		/* that little bugger belongs to csched. No idea
+		 * what it's doing here. */
+		xf_emit(ctx, 1, 0x08100c12); /* FP_INTERPOLANT_CTRL */
 		if (units & (1 << 0))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if (units & (1 << 1))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
-		/* Strand 0: TPs 2, 3 */
+		/* Strand 1: TPs 2, 3 */
 		ctx->ctxvals_pos = offset + 1;
 		if (units & (1 << 2))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if (units & (1 << 3))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
-		/* Strand 0: TPs 4, 5, 6 */
+		/* Strand 2: TPs 4, 5, 6 */
 		ctx->ctxvals_pos = offset + 2;
 		if (units & (1 << 4))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if (units & (1 << 5))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if (units & (1 << 6))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 
-		/* Strand 0: TPs 7, 8, 9 */
+		/* Strand 3: TPs 7, 8, 9 */
 		ctx->ctxvals_pos = offset + 3;
 		if (units & (1 << 7))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if (units & (1 << 8))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if (units & (1 << 9))
-			nv50_graph_construct_xfer_tp2(ctx);
+			nv50_graph_construct_xfer_mpc(ctx);
 		if ((ctx->ctxvals_pos-offset)/8 > size)
 			size = (ctx->ctxvals_pos-offset)/8;
 	}
diff --git a/drivers/gpu/drm/nouveau/nv50_instmem.c b/drivers/gpu/drm/nouveau/nv50_instmem.c
index 91ef93cf1f35..a53fc974332b 100644
--- a/drivers/gpu/drm/nouveau/nv50_instmem.c
+++ b/drivers/gpu/drm/nouveau/nv50_instmem.c
@@ -32,39 +32,87 @@
 struct nv50_instmem_priv {
 	uint32_t save1700[5]; /* 0x1700->0x1710 */
 
-	struct nouveau_gpuobj_ref *pramin_pt;
-	struct nouveau_gpuobj_ref *pramin_bar;
-	struct nouveau_gpuobj_ref *fb_bar;
+	struct nouveau_gpuobj *pramin_pt;
+	struct nouveau_gpuobj *pramin_bar;
+	struct nouveau_gpuobj *fb_bar;
 };
 
-#define NV50_INSTMEM_PAGE_SHIFT 12
-#define NV50_INSTMEM_PAGE_SIZE  (1 << NV50_INSTMEM_PAGE_SHIFT)
-#define NV50_INSTMEM_PT_SIZE(a)	(((a) >> 12) << 3)
+static void
+nv50_channel_del(struct nouveau_channel **pchan)
+{
+	struct nouveau_channel *chan;
 
-/*NOTE: - Assumes 0x1700 already covers the correct MiB of PRAMIN
- */
-#define BAR0_WI32(g, o, v) do {                                   \
-	uint32_t offset;                                          \
-	if ((g)->im_backing) {                                    \
-		offset = (g)->im_backing_start;                   \
-	} else {                                                  \
-		offset  = chan->ramin->gpuobj->im_backing_start;  \
-		offset += (g)->im_pramin->start;                  \
-	}                                                         \
-	offset += (o);                                            \
-	nv_wr32(dev, NV_RAMIN + (offset & 0xfffff), (v));              \
-} while (0)
+	chan = *pchan;
+	*pchan = NULL;
+	if (!chan)
+		return;
+
+	nouveau_gpuobj_ref(NULL, &chan->ramfc);
+	nouveau_gpuobj_ref(NULL, &chan->vm_pd);
+	if (chan->ramin_heap.free_stack.next)
+		drm_mm_takedown(&chan->ramin_heap);
+	nouveau_gpuobj_ref(NULL, &chan->ramin);
+	kfree(chan);
+}
+
+static int
+nv50_channel_new(struct drm_device *dev, u32 size,
+		 struct nouveau_channel **pchan)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	u32 pgd = (dev_priv->chipset == 0x50) ? 0x1400 : 0x0200;
+	u32  fc = (dev_priv->chipset == 0x50) ? 0x0000 : 0x4200;
+	struct nouveau_channel *chan;
+	int ret;
+
+	chan = kzalloc(sizeof(*chan), GFP_KERNEL);
+	if (!chan)
+		return -ENOMEM;
+	chan->dev = dev;
+
+	ret = nouveau_gpuobj_new(dev, NULL, size, 0x1000, 0, &chan->ramin);
+	if (ret) {
+		nv50_channel_del(&chan);
+		return ret;
+	}
+
+	ret = drm_mm_init(&chan->ramin_heap, 0x6000, chan->ramin->size);
+	if (ret) {
+		nv50_channel_del(&chan);
+		return ret;
+	}
+
+	ret = nouveau_gpuobj_new_fake(dev, chan->ramin->pinst == ~0 ? ~0 :
+				      chan->ramin->pinst + pgd,
+				      chan->ramin->vinst + pgd,
+				      0x4000, NVOBJ_FLAG_ZERO_ALLOC,
+				      &chan->vm_pd);
+	if (ret) {
+		nv50_channel_del(&chan);
+		return ret;
+	}
+
+	ret = nouveau_gpuobj_new_fake(dev, chan->ramin->pinst == ~0 ? ~0 :
+				      chan->ramin->pinst + fc,
+				      chan->ramin->vinst + fc, 0x100,
+				      NVOBJ_FLAG_ZERO_ALLOC, &chan->ramfc);
+	if (ret) {
+		nv50_channel_del(&chan);
+		return ret;
+	}
+
+	*pchan = chan;
+	return 0;
+}
 
 int
 nv50_instmem_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_channel *chan;
-	uint32_t c_offset, c_size, c_ramfc, c_vmpd, c_base, pt_size;
-	uint32_t save_nv001700;
-	uint64_t v;
 	struct nv50_instmem_priv *priv;
+	struct nouveau_channel *chan;
 	int ret, i;
+	u32 tmp;
 
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
 	if (!priv)
@@ -75,27 +123,61 @@ nv50_instmem_init(struct drm_device *dev)
 	for (i = 0x1700; i <= 0x1710; i += 4)
 		priv->save1700[(i-0x1700)/4] = nv_rd32(dev, i);
 
-	/* Reserve the last MiB of VRAM, we should probably try to avoid
-	 * setting up the below tables over the top of the VBIOS image at
-	 * some point.
-	 */
-	dev_priv->ramin_rsvd_vram = 1 << 20;
-	c_offset = dev_priv->vram_size - dev_priv->ramin_rsvd_vram;
-	c_size   = 128 << 10;
-	c_vmpd   = ((dev_priv->chipset & 0xf0) == 0x50) ? 0x1400 : 0x200;
-	c_ramfc  = ((dev_priv->chipset & 0xf0) == 0x50) ? 0x0 : 0x20;
-	c_base   = c_vmpd + 0x4000;
-	pt_size  = NV50_INSTMEM_PT_SIZE(dev_priv->ramin_size);
+	/* Global PRAMIN heap */
+	ret = drm_mm_init(&dev_priv->ramin_heap, 0, dev_priv->ramin_size);
+	if (ret) {
+		NV_ERROR(dev, "Failed to init RAMIN heap\n");
+		return -ENOMEM;
+	}
 
-	NV_DEBUG(dev, " Rsvd VRAM base: 0x%08x\n", c_offset);
-	NV_DEBUG(dev, "    VBIOS image: 0x%08x\n",
-				(nv_rd32(dev, 0x619f04) & ~0xff) << 8);
-	NV_DEBUG(dev, "  Aperture size: %d MiB\n", dev_priv->ramin_size >> 20);
-	NV_DEBUG(dev, "        PT size: %d KiB\n", pt_size >> 10);
+	/* we need a channel to plug into the hw to control the BARs */
+	ret = nv50_channel_new(dev, 128*1024, &dev_priv->fifos[0]);
+	if (ret)
+		return ret;
+	chan = dev_priv->fifos[127] = dev_priv->fifos[0];
 
-	/* Determine VM layout, we need to do this first to make sure
-	 * we allocate enough memory for all the page tables.
-	 */
+	/* allocate page table for PRAMIN BAR */
+	ret = nouveau_gpuobj_new(dev, chan, (dev_priv->ramin_size >> 12) * 8,
+				 0x1000, NVOBJ_FLAG_ZERO_ALLOC,
+				 &priv->pramin_pt);
+	if (ret)
+		return ret;
+
+	nv_wo32(chan->vm_pd, 0x0000, priv->pramin_pt->vinst | 0x63);
+	nv_wo32(chan->vm_pd, 0x0004, 0);
+
+	/* DMA object for PRAMIN BAR */
+	ret = nouveau_gpuobj_new(dev, chan, 6*4, 16, 0, &priv->pramin_bar);
+	if (ret)
+		return ret;
+	nv_wo32(priv->pramin_bar, 0x00, 0x7fc00000);
+	nv_wo32(priv->pramin_bar, 0x04, dev_priv->ramin_size - 1);
+	nv_wo32(priv->pramin_bar, 0x08, 0x00000000);
+	nv_wo32(priv->pramin_bar, 0x0c, 0x00000000);
+	nv_wo32(priv->pramin_bar, 0x10, 0x00000000);
+	nv_wo32(priv->pramin_bar, 0x14, 0x00000000);
+
+	/* map channel into PRAMIN, gpuobj didn't do it for us */
+	ret = nv50_instmem_bind(dev, chan->ramin);
+	if (ret)
+		return ret;
+
+	/* poke regs... */
+	nv_wr32(dev, 0x001704, 0x00000000 | (chan->ramin->vinst >> 12));
+	nv_wr32(dev, 0x001704, 0x40000000 | (chan->ramin->vinst >> 12));
+	nv_wr32(dev, 0x00170c, 0x80000000 | (priv->pramin_bar->cinst >> 4));
+
+	tmp = nv_ri32(dev, 0);
+	nv_wi32(dev, 0, ~tmp);
+	if (nv_ri32(dev, 0) != ~tmp) {
+		NV_ERROR(dev, "PRAMIN readback failed\n");
+		return -EIO;
+	}
+	nv_wi32(dev, 0, tmp);
+
+	dev_priv->ramin_available = true;
+
+	/* Determine VM layout */
 	dev_priv->vm_gart_base = roundup(NV50_VM_BLOCK, NV50_VM_BLOCK);
 	dev_priv->vm_gart_size = NV50_VM_BLOCK;
 
@@ -115,172 +197,41 @@ nv50_instmem_init(struct drm_device *dev)
 		 dev_priv->vm_vram_base,
 		 dev_priv->vm_vram_base + dev_priv->vm_vram_size - 1);
 
-	c_size += dev_priv->vm_vram_pt_nr * (NV50_VM_BLOCK / 65536 * 8);
-
-	/* Map BAR0 PRAMIN aperture over the memory we want to use */
-	save_nv001700 = nv_rd32(dev, NV50_PUNK_BAR0_PRAMIN);
-	nv_wr32(dev, NV50_PUNK_BAR0_PRAMIN, (c_offset >> 16));
-
-	/* Create a fake channel, and use it as our "dummy" channels 0/127.
-	 * The main reason for creating a channel is so we can use the gpuobj
-	 * code.  However, it's probably worth noting that NVIDIA also setup
-	 * their channels 0/127 with the same values they configure here.
-	 * So, there may be some other reason for doing this.
-	 *
-	 * Have to create the entire channel manually, as the real channel
-	 * creation code assumes we have PRAMIN access, and we don't until
-	 * we're done here.
-	 */
-	chan = kzalloc(sizeof(*chan), GFP_KERNEL);
-	if (!chan)
-		return -ENOMEM;
-	chan->id = 0;
-	chan->dev = dev;
-	chan->file_priv = (struct drm_file *)-2;
-	dev_priv->fifos[0] = dev_priv->fifos[127] = chan;
-
-	INIT_LIST_HEAD(&chan->ramht_refs);
-
-	/* Channel's PRAMIN object + heap */
-	ret = nouveau_gpuobj_new_fake(dev, 0, c_offset, c_size, 0,
-							NULL, &chan->ramin);
-	if (ret)
-		return ret;
-
-	if (drm_mm_init(&chan->ramin_heap, c_base, c_size - c_base))
-		return -ENOMEM;
-
-	/* RAMFC + zero channel's PRAMIN up to start of VM pagedir */
-	ret = nouveau_gpuobj_new_fake(dev, c_ramfc, c_offset + c_ramfc,
-						0x4000, 0, NULL, &chan->ramfc);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < c_vmpd; i += 4)
-		BAR0_WI32(chan->ramin->gpuobj, i, 0);
-
-	/* VM page directory */
-	ret = nouveau_gpuobj_new_fake(dev, c_vmpd, c_offset + c_vmpd,
-					   0x4000, 0, &chan->vm_pd, NULL);
-	if (ret)
-		return ret;
-	for (i = 0; i < 0x4000; i += 8) {
-		BAR0_WI32(chan->vm_pd, i + 0x00, 0x00000000);
-		BAR0_WI32(chan->vm_pd, i + 0x04, 0x00000000);
-	}
-
-	/* PRAMIN page table, cheat and map into VM at 0x0000000000.
-	 * We map the entire fake channel into the start of the PRAMIN BAR
-	 */
-	ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, pt_size, 0x1000,
-				     0, &priv->pramin_pt);
-	if (ret)
-		return ret;
-
-	v = c_offset | 1;
-	if (dev_priv->vram_sys_base) {
-		v += dev_priv->vram_sys_base;
-		v |= 0x30;
-	}
-
-	i = 0;
-	while (v < dev_priv->vram_sys_base + c_offset + c_size) {
-		BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, lower_32_bits(v));
-		BAR0_WI32(priv->pramin_pt->gpuobj, i + 4, upper_32_bits(v));
-		v += 0x1000;
-		i += 8;
-	}
-
-	while (i < pt_size) {
-		BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, 0x00000000);
-		BAR0_WI32(priv->pramin_pt->gpuobj, i + 4, 0x00000000);
-		i += 8;
-	}
-
-	BAR0_WI32(chan->vm_pd, 0x00, priv->pramin_pt->instance | 0x63);
-	BAR0_WI32(chan->vm_pd, 0x04, 0x00000000);
-
 	/* VRAM page table(s), mapped into VM at +1GiB  */
 	for (i = 0; i < dev_priv->vm_vram_pt_nr; i++) {
-		ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0,
-					     NV50_VM_BLOCK/65536*8, 0, 0,
-					     &chan->vm_vram_pt[i]);
+		ret = nouveau_gpuobj_new(dev, NULL, NV50_VM_BLOCK / 0x10000 * 8,
+					 0, NVOBJ_FLAG_ZERO_ALLOC,
+					 &chan->vm_vram_pt[i]);
 		if (ret) {
-			NV_ERROR(dev, "Error creating VRAM page tables: %d\n",
-									ret);
+			NV_ERROR(dev, "Error creating VRAM PGT: %d\n", ret);
 			dev_priv->vm_vram_pt_nr = i;
 			return ret;
 		}
-		dev_priv->vm_vram_pt[i] = chan->vm_vram_pt[i]->gpuobj;
+		dev_priv->vm_vram_pt[i] = chan->vm_vram_pt[i];
 
-		for (v = 0; v < dev_priv->vm_vram_pt[i]->im_pramin->size;
-								v += 4)
-			BAR0_WI32(dev_priv->vm_vram_pt[i], v, 0);
-
-		BAR0_WI32(chan->vm_pd, 0x10 + (i*8),
-			  chan->vm_vram_pt[i]->instance | 0x61);
-		BAR0_WI32(chan->vm_pd, 0x14 + (i*8), 0);
+		nv_wo32(chan->vm_pd, 0x10 + (i*8),
+			chan->vm_vram_pt[i]->vinst | 0x61);
+		nv_wo32(chan->vm_pd, 0x14 + (i*8), 0);
 	}
 
-	/* DMA object for PRAMIN BAR */
-	ret = nouveau_gpuobj_new_ref(dev, chan, chan, 0, 6*4, 16, 0,
-							&priv->pramin_bar);
-	if (ret)
-		return ret;
-	BAR0_WI32(priv->pramin_bar->gpuobj, 0x00, 0x7fc00000);
-	BAR0_WI32(priv->pramin_bar->gpuobj, 0x04, dev_priv->ramin_size - 1);
-	BAR0_WI32(priv->pramin_bar->gpuobj, 0x08, 0x00000000);
-	BAR0_WI32(priv->pramin_bar->gpuobj, 0x0c, 0x00000000);
-	BAR0_WI32(priv->pramin_bar->gpuobj, 0x10, 0x00000000);
-	BAR0_WI32(priv->pramin_bar->gpuobj, 0x14, 0x00000000);
-
 	/* DMA object for FB BAR */
-	ret = nouveau_gpuobj_new_ref(dev, chan, chan, 0, 6*4, 16, 0,
-							&priv->fb_bar);
+	ret = nouveau_gpuobj_new(dev, chan, 6*4, 16, 0, &priv->fb_bar);
 	if (ret)
 		return ret;
-	BAR0_WI32(priv->fb_bar->gpuobj, 0x00, 0x7fc00000);
-	BAR0_WI32(priv->fb_bar->gpuobj, 0x04, 0x40000000 +
-					      pci_resource_len(dev->pdev, 1) - 1);
-	BAR0_WI32(priv->fb_bar->gpuobj, 0x08, 0x40000000);
-	BAR0_WI32(priv->fb_bar->gpuobj, 0x0c, 0x00000000);
-	BAR0_WI32(priv->fb_bar->gpuobj, 0x10, 0x00000000);
-	BAR0_WI32(priv->fb_bar->gpuobj, 0x14, 0x00000000);
+	nv_wo32(priv->fb_bar, 0x00, 0x7fc00000);
+	nv_wo32(priv->fb_bar, 0x04, 0x40000000 +
+				    pci_resource_len(dev->pdev, 1) - 1);
+	nv_wo32(priv->fb_bar, 0x08, 0x40000000);
+	nv_wo32(priv->fb_bar, 0x0c, 0x00000000);
+	nv_wo32(priv->fb_bar, 0x10, 0x00000000);
+	nv_wo32(priv->fb_bar, 0x14, 0x00000000);
 
-	/* Poke the relevant regs, and pray it works :) */
-	nv_wr32(dev, NV50_PUNK_BAR_CFG_BASE, (chan->ramin->instance >> 12));
-	nv_wr32(dev, NV50_PUNK_UNK1710, 0);
-	nv_wr32(dev, NV50_PUNK_BAR_CFG_BASE, (chan->ramin->instance >> 12) |
-					 NV50_PUNK_BAR_CFG_BASE_VALID);
-	nv_wr32(dev, NV50_PUNK_BAR1_CTXDMA, (priv->fb_bar->instance >> 4) |
-					NV50_PUNK_BAR1_CTXDMA_VALID);
-	nv_wr32(dev, NV50_PUNK_BAR3_CTXDMA, (priv->pramin_bar->instance >> 4) |
-					NV50_PUNK_BAR3_CTXDMA_VALID);
+	dev_priv->engine.instmem.flush(dev);
 
+	nv_wr32(dev, 0x001708, 0x80000000 | (priv->fb_bar->cinst >> 4));
 	for (i = 0; i < 8; i++)
 		nv_wr32(dev, 0x1900 + (i*4), 0);
 
-	/* Assume that praying isn't enough, check that we can re-read the
-	 * entire fake channel back from the PRAMIN BAR */
-	for (i = 0; i < c_size; i += 4) {
-		if (nv_rd32(dev, NV_RAMIN + i) != nv_ri32(dev, i)) {
-			NV_ERROR(dev, "Error reading back PRAMIN at 0x%08x\n",
-									i);
-			return -EINVAL;
-		}
-	}
-
-	nv_wr32(dev, NV50_PUNK_BAR0_PRAMIN, save_nv001700);
-
-	/* Global PRAMIN heap */
-	if (drm_mm_init(&dev_priv->ramin_heap, c_size, dev_priv->ramin_size - c_size)) {
-		NV_ERROR(dev, "Failed to init RAMIN heap\n");
-	}
-
-	/*XXX: incorrect, but needed to make hash func "work" */
-	dev_priv->ramht_offset = 0x10000;
-	dev_priv->ramht_bits   = 9;
-	dev_priv->ramht_size   = (1 << dev_priv->ramht_bits) * 8;
 	return 0;
 }
 
@@ -297,29 +248,24 @@ nv50_instmem_takedown(struct drm_device *dev)
 	if (!priv)
 		return;
 
+	dev_priv->ramin_available = false;
+
 	/* Restore state from before init */
 	for (i = 0x1700; i <= 0x1710; i += 4)
 		nv_wr32(dev, i, priv->save1700[(i - 0x1700) / 4]);
 
-	nouveau_gpuobj_ref_del(dev, &priv->fb_bar);
-	nouveau_gpuobj_ref_del(dev, &priv->pramin_bar);
-	nouveau_gpuobj_ref_del(dev, &priv->pramin_pt);
+	nouveau_gpuobj_ref(NULL, &priv->fb_bar);
+	nouveau_gpuobj_ref(NULL, &priv->pramin_bar);
+	nouveau_gpuobj_ref(NULL, &priv->pramin_pt);
 
 	/* Destroy dummy channel */
 	if (chan) {
-		for (i = 0; i < dev_priv->vm_vram_pt_nr; i++) {
-			nouveau_gpuobj_ref_del(dev, &chan->vm_vram_pt[i]);
-			dev_priv->vm_vram_pt[i] = NULL;
-		}
+		for (i = 0; i < dev_priv->vm_vram_pt_nr; i++)
+			nouveau_gpuobj_ref(NULL, &chan->vm_vram_pt[i]);
 		dev_priv->vm_vram_pt_nr = 0;
 
-		nouveau_gpuobj_del(dev, &chan->vm_pd);
-		nouveau_gpuobj_ref_del(dev, &chan->ramfc);
-		nouveau_gpuobj_ref_del(dev, &chan->ramin);
-		drm_mm_takedown(&chan->ramin_heap);
-
-		dev_priv->fifos[0] = dev_priv->fifos[127] = NULL;
-		kfree(chan);
+		nv50_channel_del(&dev_priv->fifos[0]);
+		dev_priv->fifos[127] = NULL;
 	}
 
 	dev_priv->engine.instmem.priv = NULL;
@@ -331,14 +277,14 @@ nv50_instmem_suspend(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_channel *chan = dev_priv->fifos[0];
-	struct nouveau_gpuobj *ramin = chan->ramin->gpuobj;
+	struct nouveau_gpuobj *ramin = chan->ramin;
 	int i;
 
-	ramin->im_backing_suspend = vmalloc(ramin->im_pramin->size);
+	ramin->im_backing_suspend = vmalloc(ramin->size);
 	if (!ramin->im_backing_suspend)
 		return -ENOMEM;
 
-	for (i = 0; i < ramin->im_pramin->size; i += 4)
+	for (i = 0; i < ramin->size; i += 4)
 		ramin->im_backing_suspend[i/4] = nv_ri32(dev, i);
 	return 0;
 }
@@ -349,23 +295,25 @@ nv50_instmem_resume(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv50_instmem_priv *priv = dev_priv->engine.instmem.priv;
 	struct nouveau_channel *chan = dev_priv->fifos[0];
-	struct nouveau_gpuobj *ramin = chan->ramin->gpuobj;
+	struct nouveau_gpuobj *ramin = chan->ramin;
 	int i;
 
-	nv_wr32(dev, NV50_PUNK_BAR0_PRAMIN, (ramin->im_backing_start >> 16));
-	for (i = 0; i < ramin->im_pramin->size; i += 4)
-		BAR0_WI32(ramin, i, ramin->im_backing_suspend[i/4]);
+	dev_priv->ramin_available = false;
+	dev_priv->ramin_base = ~0;
+	for (i = 0; i < ramin->size; i += 4)
+		nv_wo32(ramin, i, ramin->im_backing_suspend[i/4]);
+	dev_priv->ramin_available = true;
 	vfree(ramin->im_backing_suspend);
 	ramin->im_backing_suspend = NULL;
 
 	/* Poke the relevant regs, and pray it works :) */
-	nv_wr32(dev, NV50_PUNK_BAR_CFG_BASE, (chan->ramin->instance >> 12));
+	nv_wr32(dev, NV50_PUNK_BAR_CFG_BASE, (chan->ramin->vinst >> 12));
 	nv_wr32(dev, NV50_PUNK_UNK1710, 0);
-	nv_wr32(dev, NV50_PUNK_BAR_CFG_BASE, (chan->ramin->instance >> 12) |
+	nv_wr32(dev, NV50_PUNK_BAR_CFG_BASE, (chan->ramin->vinst >> 12) |
 					 NV50_PUNK_BAR_CFG_BASE_VALID);
-	nv_wr32(dev, NV50_PUNK_BAR1_CTXDMA, (priv->fb_bar->instance >> 4) |
+	nv_wr32(dev, NV50_PUNK_BAR1_CTXDMA, (priv->fb_bar->cinst >> 4) |
 					NV50_PUNK_BAR1_CTXDMA_VALID);
-	nv_wr32(dev, NV50_PUNK_BAR3_CTXDMA, (priv->pramin_bar->instance >> 4) |
+	nv_wr32(dev, NV50_PUNK_BAR3_CTXDMA, (priv->pramin_bar->cinst >> 4) |
 					NV50_PUNK_BAR3_CTXDMA_VALID);
 
 	for (i = 0; i < 8; i++)
@@ -381,7 +329,7 @@ nv50_instmem_populate(struct drm_device *dev, struct nouveau_gpuobj *gpuobj,
 	if (gpuobj->im_backing)
 		return -EINVAL;
 
-	*sz = ALIGN(*sz, NV50_INSTMEM_PAGE_SIZE);
+	*sz = ALIGN(*sz, 4096);
 	if (*sz == 0)
 		return -EINVAL;
 
@@ -399,9 +347,7 @@ nv50_instmem_populate(struct drm_device *dev, struct nouveau_gpuobj *gpuobj,
 		return ret;
 	}
 
-	gpuobj->im_backing_start = gpuobj->im_backing->bo.mem.mm_node->start;
-	gpuobj->im_backing_start <<= PAGE_SHIFT;
-
+	gpuobj->vinst = gpuobj->im_backing->bo.mem.start << PAGE_SHIFT;
 	return 0;
 }
 
@@ -424,7 +370,7 @@ nv50_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv50_instmem_priv *priv = dev_priv->engine.instmem.priv;
-	struct nouveau_gpuobj *pramin_pt = priv->pramin_pt->gpuobj;
+	struct nouveau_gpuobj *pramin_pt = priv->pramin_pt;
 	uint32_t pte, pte_end;
 	uint64_t vram;
 
@@ -436,11 +382,11 @@ nv50_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj)
 
 	pte     = (gpuobj->im_pramin->start >> 12) << 1;
 	pte_end = ((gpuobj->im_pramin->size >> 12) << 1) + pte;
-	vram    = gpuobj->im_backing_start;
+	vram    = gpuobj->vinst;
 
 	NV_DEBUG(dev, "pramin=0x%lx, pte=%d, pte_end=%d\n",
 		 gpuobj->im_pramin->start, pte, pte_end);
-	NV_DEBUG(dev, "first vram page: 0x%08x\n", gpuobj->im_backing_start);
+	NV_DEBUG(dev, "first vram page: 0x%010llx\n", gpuobj->vinst);
 
 	vram |= 1;
 	if (dev_priv->vram_sys_base) {
@@ -449,9 +395,10 @@ nv50_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj)
 	}
 
 	while (pte < pte_end) {
-		nv_wo32(dev, pramin_pt, pte++, lower_32_bits(vram));
-		nv_wo32(dev, pramin_pt, pte++, upper_32_bits(vram));
-		vram += NV50_INSTMEM_PAGE_SIZE;
+		nv_wo32(pramin_pt, (pte * 4) + 0, lower_32_bits(vram));
+		nv_wo32(pramin_pt, (pte * 4) + 4, upper_32_bits(vram));
+		vram += 0x1000;
+		pte += 2;
 	}
 	dev_priv->engine.instmem.flush(dev);
 
@@ -472,12 +419,17 @@ nv50_instmem_unbind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj)
 	if (gpuobj->im_bound == 0)
 		return -EINVAL;
 
+	/* can happen during late takedown */
+	if (unlikely(!dev_priv->ramin_available))
+		return 0;
+
 	pte     = (gpuobj->im_pramin->start >> 12) << 1;
 	pte_end = ((gpuobj->im_pramin->size >> 12) << 1) + pte;
 
 	while (pte < pte_end) {
-		nv_wo32(dev, priv->pramin_pt->gpuobj, pte++, 0x00000000);
-		nv_wo32(dev, priv->pramin_pt->gpuobj, pte++, 0x00000000);
+		nv_wo32(priv->pramin_pt, (pte * 4) + 0, 0x00000000);
+		nv_wo32(priv->pramin_pt, (pte * 4) + 4, 0x00000000);
+		pte += 2;
 	}
 	dev_priv->engine.instmem.flush(dev);
 
@@ -489,7 +441,7 @@ void
 nv50_instmem_flush(struct drm_device *dev)
 {
 	nv_wr32(dev, 0x00330c, 0x00000001);
-	if (!nv_wait(0x00330c, 0x00000002, 0x00000000))
+	if (!nv_wait(dev, 0x00330c, 0x00000002, 0x00000000))
 		NV_ERROR(dev, "PRAMIN flush timeout\n");
 }
 
@@ -497,7 +449,7 @@ void
 nv84_instmem_flush(struct drm_device *dev)
 {
 	nv_wr32(dev, 0x070000, 0x00000001);
-	if (!nv_wait(0x070000, 0x00000002, 0x00000000))
+	if (!nv_wait(dev, 0x070000, 0x00000002, 0x00000000))
 		NV_ERROR(dev, "PRAMIN flush timeout\n");
 }
 
@@ -505,7 +457,7 @@ void
 nv50_vm_flush(struct drm_device *dev, int engine)
 {
 	nv_wr32(dev, 0x100c80, (engine << 16) | 1);
-	if (!nv_wait(0x100c80, 0x00000001, 0x00000000))
+	if (!nv_wait(dev, 0x100c80, 0x00000001, 0x00000000))
 		NV_ERROR(dev, "vm flush timeout: engine %d\n", engine);
 }
 
diff --git a/drivers/gpu/drm/nouveau/nv50_pm.c b/drivers/gpu/drm/nouveau/nv50_pm.c
new file mode 100644
index 000000000000..7dbb305d7e63
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nv50_pm.c
@@ -0,0 +1,131 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "drmP.h"
+#include "nouveau_drv.h"
+#include "nouveau_bios.h"
+#include "nouveau_pm.h"
+
+struct nv50_pm_state {
+	struct nouveau_pm_level *perflvl;
+	struct pll_lims pll;
+	enum pll_types type;
+	int N, M, P;
+};
+
+int
+nv50_pm_clock_get(struct drm_device *dev, u32 id)
+{
+	struct pll_lims pll;
+	int P, N, M, ret;
+	u32 reg0, reg1;
+
+	ret = get_pll_limits(dev, id, &pll);
+	if (ret)
+		return ret;
+
+	reg0 = nv_rd32(dev, pll.reg + 0);
+	reg1 = nv_rd32(dev, pll.reg + 4);
+	P = (reg0 & 0x00070000) >> 16;
+	N = (reg1 & 0x0000ff00) >> 8;
+	M = (reg1 & 0x000000ff);
+
+	return ((pll.refclk * N / M) >> P);
+}
+
+void *
+nv50_pm_clock_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+		  u32 id, int khz)
+{
+	struct nv50_pm_state *state;
+	int dummy, ret;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return ERR_PTR(-ENOMEM);
+	state->type = id;
+	state->perflvl = perflvl;
+
+	ret = get_pll_limits(dev, id, &state->pll);
+	if (ret < 0) {
+		kfree(state);
+		return (ret == -ENOENT) ? NULL : ERR_PTR(ret);
+	}
+
+	ret = nv50_calc_pll(dev, &state->pll, khz, &state->N, &state->M,
+			    &dummy, &dummy, &state->P);
+	if (ret < 0) {
+		kfree(state);
+		return ERR_PTR(ret);
+	}
+
+	return state;
+}
+
+void
+nv50_pm_clock_set(struct drm_device *dev, void *pre_state)
+{
+	struct nv50_pm_state *state = pre_state;
+	struct nouveau_pm_level *perflvl = state->perflvl;
+	u32 reg = state->pll.reg, tmp;
+	struct bit_entry BIT_M;
+	u16 script;
+	int N = state->N;
+	int M = state->M;
+	int P = state->P;
+
+	if (state->type == PLL_MEMORY && perflvl->memscript &&
+	    bit_table(dev, 'M', &BIT_M) == 0 &&
+	    BIT_M.version == 1 && BIT_M.length >= 0x0b) {
+		script = ROM16(BIT_M.data[0x05]);
+		if (script)
+			nouveau_bios_run_init_table(dev, script, NULL);
+		script = ROM16(BIT_M.data[0x07]);
+		if (script)
+			nouveau_bios_run_init_table(dev, script, NULL);
+		script = ROM16(BIT_M.data[0x09]);
+		if (script)
+			nouveau_bios_run_init_table(dev, script, NULL);
+
+		nouveau_bios_run_init_table(dev, perflvl->memscript, NULL);
+	}
+
+	if (state->type == PLL_MEMORY) {
+		nv_wr32(dev, 0x100210, 0);
+		nv_wr32(dev, 0x1002dc, 1);
+	}
+
+	tmp  = nv_rd32(dev, reg + 0) & 0xfff8ffff;
+	tmp |= 0x80000000 | (P << 16);
+	nv_wr32(dev, reg + 0, tmp);
+	nv_wr32(dev, reg + 4, (N << 8) | M);
+
+	if (state->type == PLL_MEMORY) {
+		nv_wr32(dev, 0x1002dc, 0);
+		nv_wr32(dev, 0x100210, 0x80000000);
+	}
+
+	kfree(state);
+}
+
diff --git a/drivers/gpu/drm/nouveau/nv50_sor.c b/drivers/gpu/drm/nouveau/nv50_sor.c
index bcd4cf84a7e6..b4a5ecb199f9 100644
--- a/drivers/gpu/drm/nouveau/nv50_sor.c
+++ b/drivers/gpu/drm/nouveau/nv50_sor.c
@@ -92,7 +92,7 @@ nv50_sor_dpms(struct drm_encoder *encoder, int mode)
 	}
 
 	/* wait for it to be done */
-	if (!nv_wait(NV50_PDISPLAY_SOR_DPMS_CTRL(or),
+	if (!nv_wait(dev, NV50_PDISPLAY_SOR_DPMS_CTRL(or),
 		     NV50_PDISPLAY_SOR_DPMS_CTRL_PENDING, 0)) {
 		NV_ERROR(dev, "timeout: SOR_DPMS_CTRL_PENDING(%d) == 0\n", or);
 		NV_ERROR(dev, "SOR_DPMS_CTRL(%d) = 0x%08x\n", or,
@@ -108,7 +108,7 @@ nv50_sor_dpms(struct drm_encoder *encoder, int mode)
 
 	nv_wr32(dev, NV50_PDISPLAY_SOR_DPMS_CTRL(or), val |
 		NV50_PDISPLAY_SOR_DPMS_CTRL_PENDING);
-	if (!nv_wait(NV50_PDISPLAY_SOR_DPMS_STATE(or),
+	if (!nv_wait(dev, NV50_PDISPLAY_SOR_DPMS_STATE(or),
 		     NV50_PDISPLAY_SOR_DPMS_STATE_WAIT, 0)) {
 		NV_ERROR(dev, "timeout: SOR_DPMS_STATE_WAIT(%d) == 0\n", or);
 		NV_ERROR(dev, "SOR_DPMS_STATE(%d) = 0x%08x\n", or,
diff --git a/drivers/gpu/drm/nouveau/nva3_pm.c b/drivers/gpu/drm/nouveau/nva3_pm.c
new file mode 100644
index 000000000000..dbbafed36406
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/nva3_pm.c
@@ -0,0 +1,95 @@
+/*
+ * Copyright 2010 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "drmP.h"
+#include "nouveau_drv.h"
+#include "nouveau_bios.h"
+#include "nouveau_pm.h"
+
+/*XXX: boards using limits 0x40 need fixing, the register layout
+ *     is correct here, but, there's some other funny magic
+ *     that modifies things, so it's not likely we'll set/read
+ *     the correct timings yet..  working on it...
+ */
+
+struct nva3_pm_state {
+	struct pll_lims pll;
+	int N, M, P;
+};
+
+int
+nva3_pm_clock_get(struct drm_device *dev, u32 id)
+{
+	struct pll_lims pll;
+	int P, N, M, ret;
+	u32 reg;
+
+	ret = get_pll_limits(dev, id, &pll);
+	if (ret)
+		return ret;
+
+	reg = nv_rd32(dev, pll.reg + 4);
+	P = (reg & 0x003f0000) >> 16;
+	N = (reg & 0x0000ff00) >> 8;
+	M = (reg & 0x000000ff);
+	return pll.refclk * N / M / P;
+}
+
+void *
+nva3_pm_clock_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+		  u32 id, int khz)
+{
+	struct nva3_pm_state *state;
+	int dummy, ret;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return ERR_PTR(-ENOMEM);
+
+	ret = get_pll_limits(dev, id, &state->pll);
+	if (ret < 0) {
+		kfree(state);
+		return (ret == -ENOENT) ? NULL : ERR_PTR(ret);
+	}
+
+	ret = nv50_calc_pll2(dev, &state->pll, khz, &state->N, &dummy,
+			     &state->M, &state->P);
+	if (ret < 0) {
+		kfree(state);
+		return ERR_PTR(ret);
+	}
+
+	return state;
+}
+
+void
+nva3_pm_clock_set(struct drm_device *dev, void *pre_state)
+{
+	struct nva3_pm_state *state = pre_state;
+	u32 reg = state->pll.reg;
+
+	nv_wr32(dev, reg + 4, (state->P << 16) | (state->N << 8) | state->M);
+	kfree(state);
+}
+
diff --git a/drivers/gpu/drm/nouveau/nvc0_fifo.c b/drivers/gpu/drm/nouveau/nvc0_fifo.c
index d64375871979..890c2b95fbc1 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fifo.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fifo.c
@@ -42,12 +42,6 @@ nvc0_fifo_reassign(struct drm_device *dev, bool enable)
 	return false;
 }
 
-bool
-nvc0_fifo_cache_flush(struct drm_device *dev)
-{
-	return true;
-}
-
 bool
 nvc0_fifo_cache_pull(struct drm_device *dev, bool enable)
 {
diff --git a/drivers/gpu/drm/nouveau/nvc0_instmem.c b/drivers/gpu/drm/nouveau/nvc0_instmem.c
index 6b451f864783..13a0f78a9088 100644
--- a/drivers/gpu/drm/nouveau/nvc0_instmem.c
+++ b/drivers/gpu/drm/nouveau/nvc0_instmem.c
@@ -50,8 +50,7 @@ nvc0_instmem_populate(struct drm_device *dev, struct nouveau_gpuobj *gpuobj,
 		return ret;
 	}
 
-	gpuobj->im_backing_start = gpuobj->im_backing->bo.mem.mm_node->start;
-	gpuobj->im_backing_start <<= PAGE_SHIFT;
+	gpuobj->vinst = gpuobj->im_backing->bo.mem.start << PAGE_SHIFT;
 	return 0;
 }
 
@@ -84,11 +83,11 @@ nvc0_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj)
 
 	pte     = gpuobj->im_pramin->start >> 12;
 	pte_end = (gpuobj->im_pramin->size >> 12) + pte;
-	vram    = gpuobj->im_backing_start;
+	vram    = gpuobj->vinst;
 
 	NV_DEBUG(dev, "pramin=0x%lx, pte=%d, pte_end=%d\n",
 		 gpuobj->im_pramin->start, pte, pte_end);
-	NV_DEBUG(dev, "first vram page: 0x%08x\n", gpuobj->im_backing_start);
+	NV_DEBUG(dev, "first vram page: 0x%010llx\n", gpuobj->vinst);
 
 	while (pte < pte_end) {
 		nv_wr32(dev, 0x702000 + (pte * 8), (vram >> 8) | 1);
@@ -134,7 +133,7 @@ void
 nvc0_instmem_flush(struct drm_device *dev)
 {
 	nv_wr32(dev, 0x070000, 1);
-	if (!nv_wait(0x070000, 0x00000002, 0x00000000))
+	if (!nv_wait(dev, 0x070000, 0x00000002, 0x00000000))
 		NV_ERROR(dev, "PRAMIN flush timeout\n");
 }
 
@@ -221,10 +220,6 @@ nvc0_instmem_init(struct drm_device *dev)
 		return -ENOMEM;
 	}
 
-	/*XXX: incorrect, but needed to make hash func "work" */
-	dev_priv->ramht_offset = 0x10000;
-	dev_priv->ramht_bits   = 9;
-	dev_priv->ramht_size   = (1 << dev_priv->ramht_bits) * 8;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nvreg.h b/drivers/gpu/drm/nouveau/nvreg.h
index ad64673ace1f..881f8a585613 100644
--- a/drivers/gpu/drm/nouveau/nvreg.h
+++ b/drivers/gpu/drm/nouveau/nvreg.h
@@ -263,6 +263,7 @@
 #		define NV_CIO_CRE_HCUR_ADDR1_ADR	7:2
 #	define NV_CIO_CRE_LCD__INDEX		0x33
 #		define NV_CIO_CRE_LCD_LCD_SELECT	0:0
+#		define NV_CIO_CRE_LCD_ROUTE_MASK	0x3b
 #	define NV_CIO_CRE_DDC0_STATUS__INDEX	0x36
 #	define NV_CIO_CRE_DDC0_WR__INDEX	0x37
 #	define NV_CIO_CRE_ILACE__INDEX		0x39	/* interlace */
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 0afd1e62347d..c26106066ec2 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -435,7 +435,7 @@ int radeon_bo_get_surface_reg(struct radeon_bo *bo)
 
 out:
 	radeon_set_surface_reg(rdev, i, bo->tiling_flags, bo->pitch,
-			       bo->tbo.mem.mm_node->start << PAGE_SHIFT,
+			       bo->tbo.mem.start << PAGE_SHIFT,
 			       bo->tbo.num_pages << PAGE_SHIFT);
 	return 0;
 }
@@ -532,7 +532,7 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 	rdev = rbo->rdev;
 	if (bo->mem.mem_type == TTM_PL_VRAM) {
 		size = bo->mem.num_pages << PAGE_SHIFT;
-		offset = bo->mem.mm_node->start << PAGE_SHIFT;
+		offset = bo->mem.start << PAGE_SHIFT;
 		if ((offset + size) > rdev->mc.visible_vram_size) {
 			/* hurrah the memory is not visible ! */
 			radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_VRAM);
@@ -540,7 +540,7 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 			r = ttm_bo_validate(bo, &rbo->placement, false, true, false);
 			if (unlikely(r != 0))
 				return r;
-			offset = bo->mem.mm_node->start << PAGE_SHIFT;
+			offset = bo->mem.start << PAGE_SHIFT;
 			/* this should not happen */
 			if ((offset + size) > rdev->mc.visible_vram_size)
 				return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 84c53e41a88f..0921910698d4 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -152,6 +152,7 @@ static int radeon_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 		man->default_caching = TTM_PL_FLAG_CACHED;
 		break;
 	case TTM_PL_TT:
+		man->func = &ttm_bo_manager_func;
 		man->gpu_offset = rdev->mc.gtt_start;
 		man->available_caching = TTM_PL_MASK_CACHING;
 		man->default_caching = TTM_PL_FLAG_CACHED;
@@ -173,6 +174,7 @@ static int radeon_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 		break;
 	case TTM_PL_VRAM:
 		/* "On-card" video ram */
+		man->func = &ttm_bo_manager_func;
 		man->gpu_offset = rdev->mc.vram_start;
 		man->flags = TTM_MEMTYPE_FLAG_FIXED |
 			     TTM_MEMTYPE_FLAG_MAPPABLE;
@@ -246,8 +248,8 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 	if (unlikely(r)) {
 		return r;
 	}
-	old_start = old_mem->mm_node->start << PAGE_SHIFT;
-	new_start = new_mem->mm_node->start << PAGE_SHIFT;
+	old_start = old_mem->start << PAGE_SHIFT;
+	new_start = new_mem->start << PAGE_SHIFT;
 
 	switch (old_mem->mem_type) {
 	case TTM_PL_VRAM:
@@ -326,14 +328,7 @@ static int radeon_move_vram_ram(struct ttm_buffer_object *bo,
 	}
 	r = ttm_bo_move_ttm(bo, true, no_wait_reserve, no_wait_gpu, new_mem);
 out_cleanup:
-	if (tmp_mem.mm_node) {
-		struct ttm_bo_global *glob = rdev->mman.bdev.glob;
-
-		spin_lock(&glob->lru_lock);
-		drm_mm_put_block(tmp_mem.mm_node);
-		spin_unlock(&glob->lru_lock);
-		return r;
-	}
+	ttm_bo_mem_put(bo, &tmp_mem);
 	return r;
 }
 
@@ -372,14 +367,7 @@ static int radeon_move_ram_vram(struct ttm_buffer_object *bo,
 		goto out_cleanup;
 	}
 out_cleanup:
-	if (tmp_mem.mm_node) {
-		struct ttm_bo_global *glob = rdev->mman.bdev.glob;
-
-		spin_lock(&glob->lru_lock);
-		drm_mm_put_block(tmp_mem.mm_node);
-		spin_unlock(&glob->lru_lock);
-		return r;
-	}
+	ttm_bo_mem_put(bo, &tmp_mem);
 	return r;
 }
 
@@ -449,14 +437,14 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_
 #if __OS_HAS_AGP
 		if (rdev->flags & RADEON_IS_AGP) {
 			/* RADEON_IS_AGP is set only if AGP is active */
-			mem->bus.offset = mem->mm_node->start << PAGE_SHIFT;
+			mem->bus.offset = mem->start << PAGE_SHIFT;
 			mem->bus.base = rdev->mc.agp_base;
 			mem->bus.is_iomem = !rdev->ddev->agp->cant_use_aperture;
 		}
 #endif
 		break;
 	case TTM_PL_VRAM:
-		mem->bus.offset = mem->mm_node->start << PAGE_SHIFT;
+		mem->bus.offset = mem->start << PAGE_SHIFT;
 		/* check if it's visible */
 		if ((mem->bus.offset + mem->bus.size) > rdev->mc.visible_vram_size)
 			return -EINVAL;
@@ -699,7 +687,7 @@ static int radeon_ttm_backend_bind(struct ttm_backend *backend,
 	int r;
 
 	gtt = container_of(backend, struct radeon_ttm_backend, backend);
-	gtt->offset = bo_mem->mm_node->start << PAGE_SHIFT;
+	gtt->offset = bo_mem->start << PAGE_SHIFT;
 	if (!gtt->num_pages) {
 		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n", gtt->num_pages, bo_mem, backend);
 	}
@@ -798,9 +786,9 @@ static int radeon_ttm_debugfs_init(struct radeon_device *rdev)
 		radeon_mem_types_list[i].show = &radeon_mm_dump_table;
 		radeon_mem_types_list[i].driver_features = 0;
 		if (i == 0)
-			radeon_mem_types_list[i].data = &rdev->mman.bdev.man[TTM_PL_VRAM].manager;
+			radeon_mem_types_list[i].data = &rdev->mman.bdev.man[TTM_PL_VRAM].priv;
 		else
-			radeon_mem_types_list[i].data = &rdev->mman.bdev.man[TTM_PL_TT].manager;
+			radeon_mem_types_list[i].data = &rdev->mman.bdev.man[TTM_PL_TT].priv;
 
 	}
 	/* Add ttm page pool to debugfs */
diff --git a/drivers/gpu/drm/ttm/Makefile b/drivers/gpu/drm/ttm/Makefile
index b256d4adfafe..f3cf6f02c997 100644
--- a/drivers/gpu/drm/ttm/Makefile
+++ b/drivers/gpu/drm/ttm/Makefile
@@ -4,6 +4,7 @@
 ccflags-y := -Iinclude/drm
 ttm-y := ttm_agp_backend.o ttm_memory.o ttm_tt.o ttm_bo.o \
 	ttm_bo_util.o ttm_bo_vm.o ttm_module.o \
-	ttm_object.o ttm_lock.o ttm_execbuf_util.o ttm_page_alloc.o
+	ttm_object.o ttm_lock.o ttm_execbuf_util.o ttm_page_alloc.o \
+	ttm_bo_manager.o
 
 obj-$(CONFIG_DRM_TTM) += ttm.o
diff --git a/drivers/gpu/drm/ttm/ttm_agp_backend.c b/drivers/gpu/drm/ttm/ttm_agp_backend.c
index 4bf69c404491..f999e36f30b4 100644
--- a/drivers/gpu/drm/ttm/ttm_agp_backend.c
+++ b/drivers/gpu/drm/ttm/ttm_agp_backend.c
@@ -74,6 +74,7 @@ static int ttm_agp_bind(struct ttm_backend *backend, struct ttm_mem_reg *bo_mem)
 {
 	struct ttm_agp_backend *agp_be =
 	    container_of(backend, struct ttm_agp_backend, backend);
+	struct drm_mm_node *node = bo_mem->mm_node;
 	struct agp_memory *mem = agp_be->mem;
 	int cached = (bo_mem->placement & TTM_PL_FLAG_CACHED);
 	int ret;
@@ -81,7 +82,7 @@ static int ttm_agp_bind(struct ttm_backend *backend, struct ttm_mem_reg *bo_mem)
 	mem->is_flushed = 1;
 	mem->type = (cached) ? AGP_USER_CACHED_MEMORY : AGP_USER_MEMORY;
 
-	ret = agp_bind_memory(mem, bo_mem->mm_node->start);
+	ret = agp_bind_memory(mem, node->start);
 	if (ret)
 		printk(KERN_ERR TTM_PFX "AGP Bind memory failed.\n");
 
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index cb4cf7ef4d1e..af7b57a47fbc 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -84,11 +84,8 @@ static void ttm_mem_type_debug(struct ttm_bo_device *bdev, int mem_type)
 		man->available_caching);
 	printk(KERN_ERR TTM_PFX "    default_caching: 0x%08X\n",
 		man->default_caching);
-	if (mem_type != TTM_PL_SYSTEM) {
-		spin_lock(&bdev->glob->lru_lock);
-		drm_mm_debug_table(&man->manager, TTM_PFX);
-		spin_unlock(&bdev->glob->lru_lock);
-	}
+	if (mem_type != TTM_PL_SYSTEM)
+		(*man->func->debug)(man, TTM_PFX);
 }
 
 static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo,
@@ -421,7 +418,7 @@ moved:
 
 	if (bo->mem.mm_node) {
 		spin_lock(&bo->lock);
-		bo->offset = (bo->mem.mm_node->start << PAGE_SHIFT) +
+		bo->offset = (bo->mem.start << PAGE_SHIFT) +
 		    bdev->man[bo->mem.mem_type].gpu_offset;
 		bo->cur_placement = bo->mem.placement;
 		spin_unlock(&bo->lock);
@@ -475,11 +472,8 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, bool remove_all)
 			list_del_init(&bo->ddestroy);
 			++put_count;
 		}
-		if (bo->mem.mm_node) {
-			drm_mm_put_block(bo->mem.mm_node);
-			bo->mem.mm_node = NULL;
-		}
 		spin_unlock(&glob->lru_lock);
+		ttm_bo_mem_put(bo, &bo->mem);
 
 		atomic_set(&bo->reserved, 0);
 
@@ -621,7 +615,6 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 			bool no_wait_reserve, bool no_wait_gpu)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
-	struct ttm_bo_global *glob = bo->glob;
 	struct ttm_mem_reg evict_mem;
 	struct ttm_placement placement;
 	int ret = 0;
@@ -667,12 +660,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 	if (ret) {
 		if (ret != -ERESTARTSYS)
 			printk(KERN_ERR TTM_PFX "Buffer eviction failed\n");
-		spin_lock(&glob->lru_lock);
-		if (evict_mem.mm_node) {
-			drm_mm_put_block(evict_mem.mm_node);
-			evict_mem.mm_node = NULL;
-		}
-		spin_unlock(&glob->lru_lock);
+		ttm_bo_mem_put(bo, &evict_mem);
 		goto out;
 	}
 	bo->evicted = true;
@@ -733,41 +721,14 @@ retry:
 	return ret;
 }
 
-static int ttm_bo_man_get_node(struct ttm_buffer_object *bo,
-				struct ttm_mem_type_manager *man,
-				struct ttm_placement *placement,
-				struct ttm_mem_reg *mem,
-				struct drm_mm_node **node)
+void ttm_bo_mem_put(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem)
 {
-	struct ttm_bo_global *glob = bo->glob;
-	unsigned long lpfn;
-	int ret;
+	struct ttm_mem_type_manager *man = &bo->bdev->man[mem->mem_type];
 
-	lpfn = placement->lpfn;
-	if (!lpfn)
-		lpfn = man->size;
-	*node = NULL;
-	do {
-		ret = drm_mm_pre_get(&man->manager);
-		if (unlikely(ret))
-			return ret;
-
-		spin_lock(&glob->lru_lock);
-		*node = drm_mm_search_free_in_range(&man->manager,
-					mem->num_pages, mem->page_alignment,
-					placement->fpfn, lpfn, 1);
-		if (unlikely(*node == NULL)) {
-			spin_unlock(&glob->lru_lock);
-			return 0;
-		}
-		*node = drm_mm_get_block_atomic_range(*node, mem->num_pages,
-							mem->page_alignment,
-							placement->fpfn,
-							lpfn);
-		spin_unlock(&glob->lru_lock);
-	} while (*node == NULL);
-	return 0;
+	if (mem->mm_node)
+		(*man->func->put_node)(man, mem);
 }
+EXPORT_SYMBOL(ttm_bo_mem_put);
 
 /**
  * Repeatedly evict memory from the LRU for @mem_type until we create enough
@@ -784,14 +745,13 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_bo_global *glob = bdev->glob;
 	struct ttm_mem_type_manager *man = &bdev->man[mem_type];
-	struct drm_mm_node *node;
 	int ret;
 
 	do {
-		ret = ttm_bo_man_get_node(bo, man, placement, mem, &node);
+		ret = (*man->func->get_node)(man, bo, placement, mem);
 		if (unlikely(ret != 0))
 			return ret;
-		if (node)
+		if (mem->mm_node)
 			break;
 		spin_lock(&glob->lru_lock);
 		if (list_empty(&man->lru)) {
@@ -804,9 +764,8 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
 		if (unlikely(ret != 0))
 			return ret;
 	} while (1);
-	if (node == NULL)
+	if (mem->mm_node == NULL)
 		return -ENOMEM;
-	mem->mm_node = node;
 	mem->mem_type = mem_type;
 	return 0;
 }
@@ -880,7 +839,6 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 	bool type_found = false;
 	bool type_ok = false;
 	bool has_erestartsys = false;
-	struct drm_mm_node *node = NULL;
 	int i, ret;
 
 	mem->mm_node = NULL;
@@ -914,17 +872,15 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 
 		if (man->has_type && man->use_type) {
 			type_found = true;
-			ret = ttm_bo_man_get_node(bo, man, placement, mem,
-							&node);
+			ret = (*man->func->get_node)(man, bo, placement, mem);
 			if (unlikely(ret))
 				return ret;
 		}
-		if (node)
+		if (mem->mm_node)
 			break;
 	}
 
-	if ((type_ok && (mem_type == TTM_PL_SYSTEM)) || node) {
-		mem->mm_node = node;
+	if ((type_ok && (mem_type == TTM_PL_SYSTEM)) || mem->mm_node) {
 		mem->mem_type = mem_type;
 		mem->placement = cur_flags;
 		return 0;
@@ -994,7 +950,6 @@ int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 			bool interruptible, bool no_wait_reserve,
 			bool no_wait_gpu)
 {
-	struct ttm_bo_global *glob = bo->glob;
 	int ret = 0;
 	struct ttm_mem_reg mem;
 
@@ -1022,11 +977,8 @@ int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 		goto out_unlock;
 	ret = ttm_bo_handle_move_mem(bo, &mem, false, interruptible, no_wait_reserve, no_wait_gpu);
 out_unlock:
-	if (ret && mem.mm_node) {
-		spin_lock(&glob->lru_lock);
-		drm_mm_put_block(mem.mm_node);
-		spin_unlock(&glob->lru_lock);
-	}
+	if (ret && mem.mm_node)
+		ttm_bo_mem_put(bo, &mem);
 	return ret;
 }
 
@@ -1034,11 +986,10 @@ static int ttm_bo_mem_compat(struct ttm_placement *placement,
 			     struct ttm_mem_reg *mem)
 {
 	int i;
-	struct drm_mm_node *node = mem->mm_node;
 
-	if (node && placement->lpfn != 0 &&
-	    (node->start < placement->fpfn ||
-	     node->start + node->size > placement->lpfn))
+	if (mem->mm_node && placement->lpfn != 0 &&
+	    (mem->start < placement->fpfn ||
+	     mem->start + mem->num_pages > placement->lpfn))
 		return -1;
 
 	for (i = 0; i < placement->num_placement; i++) {
@@ -1282,7 +1233,6 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
 
 int ttm_bo_clean_mm(struct ttm_bo_device *bdev, unsigned mem_type)
 {
-	struct ttm_bo_global *glob = bdev->glob;
 	struct ttm_mem_type_manager *man;
 	int ret = -EINVAL;
 
@@ -1305,13 +1255,7 @@ int ttm_bo_clean_mm(struct ttm_bo_device *bdev, unsigned mem_type)
 	if (mem_type > 0) {
 		ttm_bo_force_list_clean(bdev, mem_type, false);
 
-		spin_lock(&glob->lru_lock);
-		if (drm_mm_clean(&man->manager))
-			drm_mm_takedown(&man->manager);
-		else
-			ret = -EBUSY;
-
-		spin_unlock(&glob->lru_lock);
+		ret = (*man->func->takedown)(man);
 	}
 
 	return ret;
@@ -1362,6 +1306,7 @@ int ttm_bo_init_mm(struct ttm_bo_device *bdev, unsigned type,
 	ret = bdev->driver->init_mem_type(bdev, type, man);
 	if (ret)
 		return ret;
+	man->bdev = bdev;
 
 	ret = 0;
 	if (type != TTM_PL_SYSTEM) {
@@ -1371,7 +1316,8 @@ int ttm_bo_init_mm(struct ttm_bo_device *bdev, unsigned type,
 			       type);
 			return ret;
 		}
-		ret = drm_mm_init(&man->manager, 0, p_size);
+
+		ret = (*man->func->init)(man, p_size);
 		if (ret)
 			return ret;
 	}
diff --git a/drivers/gpu/drm/ttm/ttm_bo_manager.c b/drivers/gpu/drm/ttm/ttm_bo_manager.c
new file mode 100644
index 000000000000..7410c190c891
--- /dev/null
+++ b/drivers/gpu/drm/ttm/ttm_bo_manager.c
@@ -0,0 +1,148 @@
+/**************************************************************************
+ *
+ * Copyright (c) 2007-2009 VMware, Inc., Palo Alto, CA., USA
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+/*
+ * Authors: Thomas Hellstrom <thellstrom-at-vmware-dot-com>
+ */
+
+#include "ttm/ttm_module.h"
+#include "ttm/ttm_bo_driver.h"
+#include "ttm/ttm_placement.h"
+#include <linux/jiffies.h>
+#include <linux/slab.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/file.h>
+#include <linux/module.h>
+
+static int ttm_bo_man_get_node(struct ttm_mem_type_manager *man,
+			       struct ttm_buffer_object *bo,
+			       struct ttm_placement *placement,
+			       struct ttm_mem_reg *mem)
+{
+	struct ttm_bo_global *glob = man->bdev->glob;
+	struct drm_mm *mm = man->priv;
+	struct drm_mm_node *node = NULL;
+	unsigned long lpfn;
+	int ret;
+
+	lpfn = placement->lpfn;
+	if (!lpfn)
+		lpfn = man->size;
+	do {
+		ret = drm_mm_pre_get(mm);
+		if (unlikely(ret))
+			return ret;
+
+		spin_lock(&glob->lru_lock);
+		node = drm_mm_search_free_in_range(mm,
+					mem->num_pages, mem->page_alignment,
+					placement->fpfn, lpfn, 1);
+		if (unlikely(node == NULL)) {
+			spin_unlock(&glob->lru_lock);
+			return 0;
+		}
+		node = drm_mm_get_block_atomic_range(node, mem->num_pages,
+							mem->page_alignment,
+							placement->fpfn,
+							lpfn);
+		spin_unlock(&glob->lru_lock);
+	} while (node == NULL);
+
+	mem->mm_node = node;
+	mem->start = node->start;
+	return 0;
+}
+
+static void ttm_bo_man_put_node(struct ttm_mem_type_manager *man,
+				struct ttm_mem_reg *mem)
+{
+	struct ttm_bo_global *glob = man->bdev->glob;
+
+	if (mem->mm_node) {
+		spin_lock(&glob->lru_lock);
+		drm_mm_put_block(mem->mm_node);
+		spin_unlock(&glob->lru_lock);
+		mem->mm_node = NULL;
+	}
+}
+
+static int ttm_bo_man_init(struct ttm_mem_type_manager *man,
+			   unsigned long p_size)
+{
+	struct drm_mm *mm;
+	int ret;
+
+	mm = kzalloc(sizeof(*mm), GFP_KERNEL);
+	if (!mm)
+		return -ENOMEM;
+
+	ret = drm_mm_init(mm, 0, p_size);
+	if (ret) {
+		kfree(mm);
+		return ret;
+	}
+
+	man->priv = mm;
+	return 0;
+}
+
+static int ttm_bo_man_takedown(struct ttm_mem_type_manager *man)
+{
+	struct ttm_bo_global *glob = man->bdev->glob;
+	struct drm_mm *mm = man->priv;
+	int ret = 0;
+
+	spin_lock(&glob->lru_lock);
+	if (drm_mm_clean(mm)) {
+		drm_mm_takedown(mm);
+		kfree(mm);
+		man->priv = NULL;
+	} else
+		ret = -EBUSY;
+	spin_unlock(&glob->lru_lock);
+	return ret;
+}
+
+static void ttm_bo_man_debug(struct ttm_mem_type_manager *man,
+			     const char *prefix)
+{
+	struct ttm_bo_global *glob = man->bdev->glob;
+	struct drm_mm *mm = man->priv;
+
+	spin_lock(&glob->lru_lock);
+	drm_mm_debug_table(mm, prefix);
+	spin_unlock(&glob->lru_lock);
+}
+
+const struct ttm_mem_type_manager_func ttm_bo_manager_func = {
+	ttm_bo_man_init,
+	ttm_bo_man_takedown,
+	ttm_bo_man_get_node,
+	ttm_bo_man_put_node,
+	ttm_bo_man_debug
+};
+EXPORT_SYMBOL(ttm_bo_manager_func);
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 3451a82adba7..ff358ad45aa3 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -39,14 +39,7 @@
 
 void ttm_bo_free_old_node(struct ttm_buffer_object *bo)
 {
-	struct ttm_mem_reg *old_mem = &bo->mem;
-
-	if (old_mem->mm_node) {
-		spin_lock(&bo->glob->lru_lock);
-		drm_mm_put_block(old_mem->mm_node);
-		spin_unlock(&bo->glob->lru_lock);
-	}
-	old_mem->mm_node = NULL;
+	ttm_bo_mem_put(bo, &bo->mem);
 }
 
 int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
@@ -263,8 +256,7 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	dir = 1;
 
 	if ((old_mem->mem_type == new_mem->mem_type) &&
-	    (new_mem->mm_node->start <
-	     old_mem->mm_node->start + old_mem->mm_node->size)) {
+	    (new_mem->start < old_mem->start + old_mem->size)) {
 		dir = -1;
 		add = new_mem->num_pages - 1;
 	}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
index c4f5114aee7c..1b3bd8c6c67e 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
@@ -147,6 +147,7 @@ int vmw_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 		break;
 	case TTM_PL_VRAM:
 		/* "On-card" video ram */
+		man->func = &ttm_bo_manager_func;
 		man->gpu_offset = 0;
 		man->flags = TTM_MEMTYPE_FLAG_FIXED | TTM_MEMTYPE_FLAG_MAPPABLE;
 		man->available_caching = TTM_PL_MASK_CACHING;
@@ -203,7 +204,7 @@ static int vmw_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg
 		/* System memory */
 		return 0;
 	case TTM_PL_VRAM:
-		mem->bus.offset = mem->mm_node->start << PAGE_SHIFT;
+		mem->bus.offset = mem->start << PAGE_SHIFT;
 		mem->bus.base = dev_priv->vram_start;
 		mem->bus.is_iomem = true;
 		break;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
index b27a9f2887d2..e7304188a784 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
@@ -612,6 +612,7 @@ int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv,
 {
 	struct ttm_buffer_object *bo = &vmw_bo->base;
 	struct ttm_placement ne_placement = vmw_vram_ne_placement;
+	struct drm_mm_node *mm_node;
 	int ret = 0;
 
 	ne_placement.lpfn = bo->num_pages;
@@ -625,8 +626,9 @@ int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv,
 	if (unlikely(ret != 0))
 		goto err_unlock;
 
+	mm_node = bo->mem.mm_node;
 	if (bo->mem.mem_type == TTM_PL_VRAM &&
-	    bo->mem.mm_node->start < bo->num_pages)
+	    mm_node->start < bo->num_pages)
 		(void) ttm_bo_validate(bo, &vmw_sys_placement, false,
 				       false, false);
 
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index 267a86c74e2e..49b43c23636a 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -102,7 +102,8 @@ struct ttm_bus_placement {
  */
 
 struct ttm_mem_reg {
-	struct drm_mm_node *mm_node;
+	void *mm_node;
+	unsigned long start;
 	unsigned long size;
 	unsigned long num_pages;
 	uint32_t page_alignment;
diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h
index b87504235f18..e3371dbe6a10 100644
--- a/include/drm/ttm/ttm_bo_driver.h
+++ b/include/drm/ttm/ttm_bo_driver.h
@@ -203,7 +203,22 @@ struct ttm_tt {
  * It's set up by the ttm_bo_driver::init_mem_type method.
  */
 
+struct ttm_mem_type_manager;
+
+struct ttm_mem_type_manager_func {
+	int  (*init)(struct ttm_mem_type_manager *man, unsigned long p_size);
+	int  (*takedown)(struct ttm_mem_type_manager *man);
+	int  (*get_node)(struct ttm_mem_type_manager *man,
+			 struct ttm_buffer_object *bo,
+			 struct ttm_placement *placement,
+			 struct ttm_mem_reg *mem);
+	void (*put_node)(struct ttm_mem_type_manager *man,
+			 struct ttm_mem_reg *mem);
+	void (*debug)(struct ttm_mem_type_manager *man, const char *prefix);
+};
+
 struct ttm_mem_type_manager {
+	struct ttm_bo_device *bdev;
 
 	/*
 	 * No protection. Constant from start.
@@ -222,8 +237,8 @@ struct ttm_mem_type_manager {
 	 * TODO: Consider one lru_lock per ttm_mem_type_manager.
 	 * Plays ill with list removal, though.
 	 */
-
-	struct drm_mm manager;
+	const struct ttm_mem_type_manager_func *func;
+	void *priv;
 	struct list_head lru;
 };
 
@@ -649,6 +664,10 @@ extern int ttm_bo_mem_space(struct ttm_buffer_object *bo,
 				struct ttm_mem_reg *mem,
 				bool interruptible,
 				bool no_wait_reserve, bool no_wait_gpu);
+
+extern void ttm_bo_mem_put(struct ttm_buffer_object *bo,
+			   struct ttm_mem_reg *mem);
+
 /**
  * ttm_bo_wait_for_cpu
  *
@@ -891,6 +910,8 @@ extern int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
  */
 extern pgprot_t ttm_io_prot(uint32_t caching_flags, pgprot_t tmp);
 
+extern const struct ttm_mem_type_manager_func ttm_bo_manager_func;
+
 #if (defined(CONFIG_AGP) || (defined(CONFIG_AGP_MODULE) && defined(MODULE)))
 #define TTM_HAS_AGP
 #include <linux/agp_backend.h>