redonkable
/
alistair23-linux
1
0
Fork 0
Code Releases Activity
alistair23-linux/sound/hda/hdac_regmap.c

606 lines
15 KiB
C
Raw Normal View History

treewide: Add SPDX license identifier for missed files Add SPDX license identifiers to all files which: - Have no license information of any form - Have EXPORT_.*_SYMBOL_GPL inside which was used in the initial scan/conversion to ignore the file These files fall under the project license, GPL v2 only. The resulting SPDX license identifier is: GPL-2.0-only Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2019-05-19 06:08:55 -06:00
// SPDX-License-Identifier: GPL-2.0-only
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
/*
* Regmap support for HD-audio verbs
*
* A virtual register is translated to one or more hda verbs for write,
* vice versa for read.
*
* A few limitations:
* - Provided for not all verbs but only subset standard non-volatile verbs.
* - For reading, only AC_VERB_GET_* variants can be used.
* - For writing, mapped to the *corresponding* AC_VERB_SET_* variants,
* so can't handle asymmetric verbs for read and write
*/
#include <linux/slab.h>
#include <linux/device.h>
#include <linux/regmap.h>
#include <linux/export.h>
#include <linux/pm.h>
#include <linux/pm_runtime.h>
#include <sound/core.h>
#include <sound/hdaudio.h>
#include <sound/hda_regmap.h>
ALSA: hda: Unexport a few more stuff Drop EXPORT_SYMBOL*() from a few more stuff in HD-audio core that aren't used outside. Particular the unsol event handler can be staticized now because the recent change removed all external callers. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2019-08-14 11:59:44 -06:00
#include "local.h"
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
static int codec_pm_lock(struct hdac_device *codec)
{
return snd_hdac_keep_power_up(codec);
}
static void codec_pm_unlock(struct hdac_device *codec, int lock)
{
if (lock == 1)
snd_hdac_power_down_pm(codec);
}
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
#define get_verb(reg) (((reg) >> 8) & 0xfff)
static bool hda_volatile_reg(struct device *dev, unsigned int reg)
{
ALSA: hda - Add cache support for COEF read/write The 16bit COEF read/write is pretty standard for many codecs, and they can be cached in most cases -- more importantly, they need to be restored at resume. For making this easier, add the cache support to regmap. If the codec driver wants to cache the COEF access, set codec->cache_coef flag and issue AC_VERB_GET_PROC_COEF with the coef index in LSB 8 bits. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-13 08:56:25 -06:00
struct hdac_device *codec = dev_to_hdac_dev(dev);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
unsigned int verb = get_verb(reg);
switch (verb) {
case AC_VERB_GET_PROC_COEF:
ALSA: hda - Add cache support for COEF read/write The 16bit COEF read/write is pretty standard for many codecs, and they can be cached in most cases -- more importantly, they need to be restored at resume. For making this easier, add the cache support to regmap. If the codec driver wants to cache the COEF access, set codec->cache_coef flag and issue AC_VERB_GET_PROC_COEF with the coef index in LSB 8 bits. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-13 08:56:25 -06:00
return !codec->cache_coef;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
case AC_VERB_GET_COEF_INDEX:
case AC_VERB_GET_PROC_STATE:
case AC_VERB_GET_POWER_STATE:
case AC_VERB_GET_PIN_SENSE:
case AC_VERB_GET_HDMI_DIP_SIZE:
case AC_VERB_GET_HDMI_ELDD:
case AC_VERB_GET_HDMI_DIP_INDEX:
case AC_VERB_GET_HDMI_DIP_DATA:
case AC_VERB_GET_HDMI_DIP_XMIT:
case AC_VERB_GET_HDMI_CP_CTRL:
case AC_VERB_GET_HDMI_CHAN_SLOT:
case AC_VERB_GET_DEVICE_SEL:
case AC_VERB_GET_DEVICE_LIST: /* read-only volatile */
return true;
}
return false;
}
static bool hda_writeable_reg(struct device *dev, unsigned int reg)
{
ALSA: hda - Use regmap for parameter caches, too The amp hash table was used for recording the cached reads of some capability values like pin caps or amp caps. Now all these are moved to regmap as well. One addition to the regmap helper is codec->caps_overwriting flag. This is set in snd_hdac_override_parm(), and the regmap helper accepts any register while this flag is set, so that it can overwrite even the read-only verb like AC_VERB_PARAMETERS. The flag is cleared immediately in snd_hdac_override_parm(), as it's a once-off flag. Along with these changes, the no longer needed amp hash and relevant fields are removed from hda_codec struct now. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 00:54:56 -07:00
struct hdac_device *codec = dev_to_hdac_dev(dev);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
unsigned int verb = get_verb(reg);
ALSA: hda - Use a macro for snd_array iteration loops Introduce a new helper macro, snd_array_for_each(), to iterate for each snd_array element. It slightly improves the readability than lengthy open codes at each place. Along with it, add const prefix to some obvious places. There should be no functional changes by this. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2018-04-23 09:24:56 -06:00
const unsigned int *v;
ALSA: hda - Allow driver to add vendor-specific verbs for regmap Codecs may have own vendor-specific verbs, and we need to allow each driver to give such verbs for cached accesses. Here a verb can be put into a single array and looked through it at readable and writeable callbacks. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 04:29:03 -07:00
int i;
ALSA: hda - Use a macro for snd_array iteration loops Introduce a new helper macro, snd_array_for_each(), to iterate for each snd_array element. It slightly improves the readability than lengthy open codes at each place. Along with it, add const prefix to some obvious places. There should be no functional changes by this. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2018-04-23 09:24:56 -06:00
snd_array_for_each(&codec->vendor_verbs, i, v) {
ALSA: hda - Allow driver to add vendor-specific verbs for regmap Codecs may have own vendor-specific verbs, and we need to allow each driver to give such verbs for cached accesses. Here a verb can be put into a single array and looked through it at readable and writeable callbacks. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 04:29:03 -07:00
if (verb == *v)
return true;
}
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
ALSA: hda - Use regmap for parameter caches, too The amp hash table was used for recording the cached reads of some capability values like pin caps or amp caps. Now all these are moved to regmap as well. One addition to the regmap helper is codec->caps_overwriting flag. This is set in snd_hdac_override_parm(), and the regmap helper accepts any register while this flag is set, so that it can overwrite even the read-only verb like AC_VERB_PARAMETERS. The flag is cleared immediately in snd_hdac_override_parm(), as it's a once-off flag. Along with these changes, the no longer needed amp hash and relevant fields are removed from hda_codec struct now. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 00:54:56 -07:00
if (codec->caps_overwriting)
return true;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
switch (verb & 0xf00) {
case AC_VERB_GET_STREAM_FORMAT:
case AC_VERB_GET_AMP_GAIN_MUTE:
return true;
ALSA: hda - Add cache support for COEF read/write The 16bit COEF read/write is pretty standard for many codecs, and they can be cached in most cases -- more importantly, they need to be restored at resume. For making this easier, add the cache support to regmap. If the codec driver wants to cache the COEF access, set codec->cache_coef flag and issue AC_VERB_GET_PROC_COEF with the coef index in LSB 8 bits. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-13 08:56:25 -06:00
case AC_VERB_GET_PROC_COEF:
return codec->cache_coef;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
case 0xf00:
break;
default:
return false;
}
switch (verb) {
case AC_VERB_GET_CONNECT_SEL:
case AC_VERB_GET_SDI_SELECT:
case AC_VERB_GET_PIN_WIDGET_CONTROL:
case AC_VERB_GET_UNSOLICITED_RESPONSE: /* only as SET_UNSOLICITED_ENABLE */
case AC_VERB_GET_BEEP_CONTROL:
case AC_VERB_GET_EAPD_BTLENABLE:
case AC_VERB_GET_DIGI_CONVERT_1:
case AC_VERB_GET_DIGI_CONVERT_2: /* only for beep control */
case AC_VERB_GET_VOLUME_KNOB_CONTROL:
case AC_VERB_GET_GPIO_MASK:
case AC_VERB_GET_GPIO_DIRECTION:
case AC_VERB_GET_GPIO_DATA: /* not for volatile read */
case AC_VERB_GET_GPIO_WAKE_MASK:
case AC_VERB_GET_GPIO_UNSOLICITED_RSP_MASK:
case AC_VERB_GET_GPIO_STICKY_MASK:
return true;
}
return false;
}
static bool hda_readable_reg(struct device *dev, unsigned int reg)
{
ALSA: hda - Use regmap for parameter caches, too The amp hash table was used for recording the cached reads of some capability values like pin caps or amp caps. Now all these are moved to regmap as well. One addition to the regmap helper is codec->caps_overwriting flag. This is set in snd_hdac_override_parm(), and the regmap helper accepts any register while this flag is set, so that it can overwrite even the read-only verb like AC_VERB_PARAMETERS. The flag is cleared immediately in snd_hdac_override_parm(), as it's a once-off flag. Along with these changes, the no longer needed amp hash and relevant fields are removed from hda_codec struct now. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 00:54:56 -07:00
struct hdac_device *codec = dev_to_hdac_dev(dev);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
unsigned int verb = get_verb(reg);
ALSA: hda - Use regmap for parameter caches, too The amp hash table was used for recording the cached reads of some capability values like pin caps or amp caps. Now all these are moved to regmap as well. One addition to the regmap helper is codec->caps_overwriting flag. This is set in snd_hdac_override_parm(), and the regmap helper accepts any register while this flag is set, so that it can overwrite even the read-only verb like AC_VERB_PARAMETERS. The flag is cleared immediately in snd_hdac_override_parm(), as it's a once-off flag. Along with these changes, the no longer needed amp hash and relevant fields are removed from hda_codec struct now. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 00:54:56 -07:00
if (codec->caps_overwriting)
return true;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
switch (verb) {
case AC_VERB_PARAMETERS:
case AC_VERB_GET_CONNECT_LIST:
case AC_VERB_GET_SUBSYSTEM_ID:
return true;
ALSA: hda - Handle a few verbs as read-only Although they can be written, handle a few verbs as read-only in regmap interface: CONFIG_DEFAULT, CONV and CVT_CHAN_COUNT. These are either updated in PCM or HDMI management code in a volatile manner, or just needed only as parameter, thus they don't need to be written at resume sync. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-26 07:18:34 -06:00
/* below are basically writable, but disabled for reducing unnecessary
* writes at sync
*/
case AC_VERB_GET_CONFIG_DEFAULT: /* usually just read */
case AC_VERB_GET_CONV: /* managed in PCM code */
case AC_VERB_GET_CVT_CHAN_COUNT: /* managed in HDMI CA code */
return true;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
return hda_writeable_reg(dev, reg);
}
ALSA: hda - Add a fake stereo amp register support HD-audio spec is inconvenient regarding the handling of stereo volume controls. It can set and get only single channel at once (although there is a special option to set the same value to both channels). This patch provides a fake pseudo-register via the regmap access so that the stereo channels can be read and written by a single call. It'd be useful, for example, for implementing DAPM widgets. A stereo amp pseudo register consists of the encoding like the normal amp verbs but it has both SET_LEFT (bit 13) and SET_RIGHT (bit 12) bits set. The regmap reads and writes a 16bit value for this pseudo register where the upper 8bit is for the right chanel and the lower 8bit for the left channel. Note that the driver doesn't recognize conflicts when both stereo and mono channel registers are mixed. Mixing them would certainly confuse the operation. So, use carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-04 12:43:20 -07:00
/*
* Stereo amp pseudo register:
* for making easier to handle the stereo volume control, we provide a
* fake register to deal both left and right channels by a single
* (pseudo) register access. A verb consisting of SET_AMP_GAIN with
* *both* SET_LEFT and SET_RIGHT bits takes a 16bit value, the lower 8bit
* for the left and the upper 8bit for the right channel.
*/
static bool is_stereo_amp_verb(unsigned int reg)
{
if (((reg >> 8) & 0x700) != AC_VERB_SET_AMP_GAIN_MUTE)
return false;
return (reg & (AC_AMP_SET_LEFT | AC_AMP_SET_RIGHT)) ==
(AC_AMP_SET_LEFT | AC_AMP_SET_RIGHT);
}
/* read a pseudo stereo amp register (16bit left+right) */
static int hda_reg_read_stereo_amp(struct hdac_device *codec,
unsigned int reg, unsigned int *val)
{
unsigned int left, right;
int err;
reg &= ~(AC_AMP_SET_LEFT | AC_AMP_SET_RIGHT);
err = snd_hdac_exec_verb(codec, reg | AC_AMP_GET_LEFT, 0, &left);
if (err < 0)
return err;
err = snd_hdac_exec_verb(codec, reg | AC_AMP_GET_RIGHT, 0, &right);
if (err < 0)
return err;
*val = left | (right << 8);
return 0;
}
/* write a pseudo stereo amp register (16bit left+right) */
static int hda_reg_write_stereo_amp(struct hdac_device *codec,
unsigned int reg, unsigned int val)
{
int err;
unsigned int verb, left, right;
verb = AC_VERB_SET_AMP_GAIN_MUTE << 8;
if (reg & AC_AMP_GET_OUTPUT)
verb |= AC_AMP_SET_OUTPUT;
else
verb |= AC_AMP_SET_INPUT | ((reg & 0xf) << 8);
reg = (reg & ~0xfffff) | verb;
left = val & 0xff;
right = (val >> 8) & 0xff;
if (left == right) {
reg |= AC_AMP_SET_LEFT | AC_AMP_SET_RIGHT;
return snd_hdac_exec_verb(codec, reg | left, 0, NULL);
}
err = snd_hdac_exec_verb(codec, reg | AC_AMP_SET_LEFT | left, 0, NULL);
if (err < 0)
return err;
err = snd_hdac_exec_verb(codec, reg | AC_AMP_SET_RIGHT | right, 0, NULL);
if (err < 0)
return err;
return 0;
}
ALSA: hda - Add cache support for COEF read/write The 16bit COEF read/write is pretty standard for many codecs, and they can be cached in most cases -- more importantly, they need to be restored at resume. For making this easier, add the cache support to regmap. If the codec driver wants to cache the COEF access, set codec->cache_coef flag and issue AC_VERB_GET_PROC_COEF with the coef index in LSB 8 bits. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-13 08:56:25 -06:00
/* read a pseudo coef register (16bit) */
static int hda_reg_read_coef(struct hdac_device *codec, unsigned int reg,
unsigned int *val)
{
unsigned int verb;
int err;
if (!codec->cache_coef)
return -EINVAL;
/* LSB 8bit = coef index */
verb = (reg & ~0xfff00) | (AC_VERB_SET_COEF_INDEX << 8);
err = snd_hdac_exec_verb(codec, verb, 0, NULL);
if (err < 0)
return err;
verb = (reg & ~0xfffff) | (AC_VERB_GET_COEF_INDEX << 8);
return snd_hdac_exec_verb(codec, verb, 0, val);
}
/* write a pseudo coef register (16bit) */
static int hda_reg_write_coef(struct hdac_device *codec, unsigned int reg,
unsigned int val)
{
unsigned int verb;
int err;
if (!codec->cache_coef)
return -EINVAL;
/* LSB 8bit = coef index */
verb = (reg & ~0xfff00) | (AC_VERB_SET_COEF_INDEX << 8);
err = snd_hdac_exec_verb(codec, verb, 0, NULL);
if (err < 0)
return err;
verb = (reg & ~0xfffff) | (AC_VERB_GET_COEF_INDEX << 8) |
(val & 0xffff);
return snd_hdac_exec_verb(codec, verb, 0, NULL);
}
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
static int hda_reg_read(void *context, unsigned int reg, unsigned int *val)
{
struct hdac_device *codec = context;
ALSA: hda - Add cache support for COEF read/write The 16bit COEF read/write is pretty standard for many codecs, and they can be cached in most cases -- more importantly, they need to be restored at resume. For making this easier, add the cache support to regmap. If the codec driver wants to cache the COEF access, set codec->cache_coef flag and issue AC_VERB_GET_PROC_COEF with the coef index in LSB 8 bits. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-13 08:56:25 -06:00
int verb = get_verb(reg);
ALSA: hda - Handle get/set power verb symmetrically via regmap HD-audio has quite a few asymmetrical ways of accessing verbs, and one of typical ones is GET/SET_POWER_STATE verbs. While it takes only the power state for setting, it returns a combination of states for getting. For making the state handling simpler, this patch adds a code to translate the value returned from GET_POWER_STATE to return only the actual state or -1 for error. In that way, the driver can simplify the power state management. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-09 15:19:47 -06:00
int err;
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
int pm_lock = 0;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
if (verb != AC_VERB_GET_POWER_STATE) {
pm_lock = codec_pm_lock(codec);
if (pm_lock < 0)
return -EAGAIN;
}
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
reg |= (codec->addr << 28);
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
if (is_stereo_amp_verb(reg)) {
err = hda_reg_read_stereo_amp(codec, reg, val);
goto out;
}
if (verb == AC_VERB_GET_PROC_COEF) {
err = hda_reg_read_coef(codec, reg, val);
goto out;
}
ALSA: hda - Re-add the lost fake mute support Yet another regression by the transition to regmap cache; for better usability, we had the fake mute control using the zero amp value for Conexant codecs, and this was forgotten in the new hda core code. Since the bits 4-7 are unused for the amp registers (as we follow the syntax of AMP_GET verb), the bit 4 is now used to indicate the fake mute. For setting this flag, snd_hda_codec_amp_update() becomes a function from a simple macro. The bonus is that it gained a proper function description. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-06-11 02:51:28 -06:00
if ((verb & 0x700) == AC_VERB_SET_AMP_GAIN_MUTE)
reg &= ~AC_AMP_FAKE_MUTE;
ALSA: hda - Handle get/set power verb symmetrically via regmap HD-audio has quite a few asymmetrical ways of accessing verbs, and one of typical ones is GET/SET_POWER_STATE verbs. While it takes only the power state for setting, it returns a combination of states for getting. For making the state handling simpler, this patch adds a code to translate the value returned from GET_POWER_STATE to return only the actual state or -1 for error. In that way, the driver can simplify the power state management. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-09 15:19:47 -06:00
err = snd_hdac_exec_verb(codec, reg, 0, val);
if (err < 0)
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
goto out;
ALSA: hda - Handle get/set power verb symmetrically via regmap HD-audio has quite a few asymmetrical ways of accessing verbs, and one of typical ones is GET/SET_POWER_STATE verbs. While it takes only the power state for setting, it returns a combination of states for getting. For making the state handling simpler, this patch adds a code to translate the value returned from GET_POWER_STATE to return only the actual state or -1 for error. In that way, the driver can simplify the power state management. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-09 15:19:47 -06:00
/* special handling for asymmetric reads */
ALSA: hda - Add cache support for COEF read/write The 16bit COEF read/write is pretty standard for many codecs, and they can be cached in most cases -- more importantly, they need to be restored at resume. For making this easier, add the cache support to regmap. If the codec driver wants to cache the COEF access, set codec->cache_coef flag and issue AC_VERB_GET_PROC_COEF with the coef index in LSB 8 bits. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-13 08:56:25 -06:00
if (verb == AC_VERB_GET_POWER_STATE) {
ALSA: hda - Handle get/set power verb symmetrically via regmap HD-audio has quite a few asymmetrical ways of accessing verbs, and one of typical ones is GET/SET_POWER_STATE verbs. While it takes only the power state for setting, it returns a combination of states for getting. For making the state handling simpler, this patch adds a code to translate the value returned from GET_POWER_STATE to return only the actual state or -1 for error. In that way, the driver can simplify the power state management. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-09 15:19:47 -06:00
if (*val & AC_PWRST_ERROR)
*val = -1;
else /* take only the actual state */
*val = (*val >> 4) & 0x0f;
}
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
out:
codec_pm_unlock(codec, pm_lock);
return err;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
static int hda_reg_write(void *context, unsigned int reg, unsigned int val)
{
struct hdac_device *codec = context;
unsigned int verb;
int i, bytes, err;
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
int pm_lock = 0;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
ALSA: hda - Don't actually write registers for caps overwrites Along with the transition to regmap for managing the cached parameter reads, the caps overwrite was also moved to regmap cache. The cache change itself works, but it still tries to write the non-existing verb (the HDA parameter is read-only) wrongly. It's harmless in most cases, but some chips are picky and may result in the codec communication stall. This patch avoids it just by adding the missing flag check in reg_write ops. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-06-10 02:27:00 -06:00
if (codec->caps_overwriting)
return 0;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
reg &= ~0x00080000U; /* drop GET bit */
reg |= (codec->addr << 28);
ALSA: hda - Always allow access for POWER_STATE verbs via regmap The hdac regmap code checks whether the codec is powered on while accessing, but there must be an exception -- the verbs to control the power state. Currently HD-audio driver doesn't access them via regmap, so this patch doesn't fix any current behavior, but it's just for future. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-04-08 23:58:46 -06:00
verb = get_verb(reg);
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
if (verb != AC_VERB_SET_POWER_STATE) {
pm_lock = codec_pm_lock(codec);
if (pm_lock < 0)
return codec->lazy_cache ? 0 : -EAGAIN;
}
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
if (is_stereo_amp_verb(reg)) {
err = hda_reg_write_stereo_amp(codec, reg, val);
goto out;
}
ALSA: hda - Add a fake stereo amp register support HD-audio spec is inconvenient regarding the handling of stereo volume controls. It can set and get only single channel at once (although there is a special option to set the same value to both channels). This patch provides a fake pseudo-register via the regmap access so that the stereo channels can be read and written by a single call. It'd be useful, for example, for implementing DAPM widgets. A stereo amp pseudo register consists of the encoding like the normal amp verbs but it has both SET_LEFT (bit 13) and SET_RIGHT (bit 12) bits set. The regmap reads and writes a 16bit value for this pseudo register where the upper 8bit is for the right chanel and the lower 8bit for the left channel. Note that the driver doesn't recognize conflicts when both stereo and mono channel registers are mixed. Mixing them would certainly confuse the operation. So, use carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-04 12:43:20 -07:00
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
if (verb == AC_VERB_SET_PROC_COEF) {
err = hda_reg_write_coef(codec, reg, val);
goto out;
}
ALSA: hda - Add cache support for COEF read/write The 16bit COEF read/write is pretty standard for many codecs, and they can be cached in most cases -- more importantly, they need to be restored at resume. For making this easier, add the cache support to regmap. If the codec driver wants to cache the COEF access, set codec->cache_coef flag and issue AC_VERB_GET_PROC_COEF with the coef index in LSB 8 bits. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-03-13 08:56:25 -06:00
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
switch (verb & 0xf00) {
case AC_VERB_SET_AMP_GAIN_MUTE:
ALSA: hda - Re-add the lost fake mute support Yet another regression by the transition to regmap cache; for better usability, we had the fake mute control using the zero amp value for Conexant codecs, and this was forgotten in the new hda core code. Since the bits 4-7 are unused for the amp registers (as we follow the syntax of AMP_GET verb), the bit 4 is now used to indicate the fake mute. For setting this flag, snd_hda_codec_amp_update() becomes a function from a simple macro. The bonus is that it gained a proper function description. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-06-11 02:51:28 -06:00
if ((reg & AC_AMP_FAKE_MUTE) && (val & AC_AMP_MUTE))
val = 0;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
verb = AC_VERB_SET_AMP_GAIN_MUTE;
if (reg & AC_AMP_GET_LEFT)
verb |= AC_AMP_SET_LEFT >> 8;
else
verb |= AC_AMP_SET_RIGHT >> 8;
if (reg & AC_AMP_GET_OUTPUT) {
verb |= AC_AMP_SET_OUTPUT >> 8;
} else {
verb |= AC_AMP_SET_INPUT >> 8;
verb |= reg & 0xf;
}
break;
}
switch (verb) {
case AC_VERB_SET_DIGI_CONVERT_1:
bytes = 2;
break;
case AC_VERB_SET_CONFIG_DEFAULT_BYTES_0:
bytes = 4;
break;
default:
bytes = 1;
break;
}
for (i = 0; i < bytes; i++) {
reg &= ~0xfffff;
reg |= (verb + i) << 8 | ((val >> (8 * i)) & 0xff);
err = snd_hdac_exec_verb(codec, reg, 0, NULL);
if (err < 0)
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
goto out;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
ALSA: hda - Fix unexpected resume through regmap code path HD-audio driver has a mechanism to trigger the runtime resume automatically at accessing the verbs. This auto-resume, however, causes the mutex deadlock when invoked from the regmap handler since the regmap keeps the mutex while auto-resuming. For avoiding that, there is some tricky check in the HDA regmap handler to return -EAGAIN error to back-off when the codec is powered down. Then the caller of regmap r/w will retry after properly turning on the codec power. This works in most cases, but there seems a slight race between the codec power check and the actual on-demand auto-resume trigger. This resulted in the lockdep splat, eventually leading to a real deadlock. This patch tries to address the race window by getting the runtime PM refcount at the check time using pm_runtime_get_if_in_use(). With this call, we can keep the power on only when the codec has been already turned on, and back off if not. For keeping the code consistency, the code touching the runtime PM is stored in hdac_device.c although it's used only locally in hdac_regmap.c. Reported-by: Jiri Slaby <jslaby@suse.cz> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-03-04 03:34:18 -07:00
out:
codec_pm_unlock(codec, pm_lock);
return err;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
static const struct regmap_config hda_regmap_cfg = {
.name = "hdaudio",
.reg_bits = 32,
.val_bits = 32,
.max_register = 0xfffffff,
.writeable_reg = hda_writeable_reg,
.readable_reg = hda_readable_reg,
.volatile_reg = hda_volatile_reg,
.cache_type = REGCACHE_RBTREE,
.reg_read = hda_reg_read,
.reg_write = hda_reg_write,
regmap: split up regmap_config.use_single_rw Split regmap_config.use_single_rw into use_single_read and use_single_write. This change enables drivers of devices which only support bulk operations in one direction to use the regmap_bulk_*() functions for both directions and have their bulk operation split into single operations only when necessary. Update all struct regmap_config instances where use_single_rw==true to instead set both use_single_read and use_single_write. No attempt was made to evaluate whether it is possible to set only one of use_single_read or use_single_write. Signed-off-by: David Frey <dpfrey@gmail.com> Signed-off-by: Mark Brown <broonie@kernel.org>
2018-09-01 10:50:41 -06:00
.use_single_read = true,
.use_single_write = true,
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
};
ALSA: hda - Add / fix kernel doc comments Give some readable comment in kernel doc style for each exported function, as I promised in the previous meetings. While we're at it, fix the wrong comments, too. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-10-28 05:26:48 -06:00
/**
* snd_hdac_regmap_init - Initialize regmap for HDA register accesses
* @codec: the codec object
*
* Returns zero for success or a negative error code.
*/
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
int snd_hdac_regmap_init(struct hdac_device *codec)
{
struct regmap *regmap;
regmap = regmap_init(&codec->dev, NULL, codec, &hda_regmap_cfg);
if (IS_ERR(regmap))
return PTR_ERR(regmap);
codec->regmap = regmap;
ALSA: hda - Allow driver to add vendor-specific verbs for regmap Codecs may have own vendor-specific verbs, and we need to allow each driver to give such verbs for cached accesses. Here a verb can be put into a single array and looked through it at readable and writeable callbacks. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 04:29:03 -07:00
snd_array_init(&codec->vendor_verbs, sizeof(unsigned int), 8);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
return 0;
}
EXPORT_SYMBOL_GPL(snd_hdac_regmap_init);
ALSA: hda - Add / fix kernel doc comments Give some readable comment in kernel doc style for each exported function, as I promised in the previous meetings. While we're at it, fix the wrong comments, too. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-10-28 05:26:48 -06:00
/**
* snd_hdac_regmap_init - Release the regmap from HDA codec
* @codec: the codec object
*/
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
void snd_hdac_regmap_exit(struct hdac_device *codec)
{
if (codec->regmap) {
regmap_exit(codec->regmap);
codec->regmap = NULL;
ALSA: hda - Allow driver to add vendor-specific verbs for regmap Codecs may have own vendor-specific verbs, and we need to allow each driver to give such verbs for cached accesses. Here a verb can be put into a single array and looked through it at readable and writeable callbacks. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 04:29:03 -07:00
snd_array_free(&codec->vendor_verbs);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
}
EXPORT_SYMBOL_GPL(snd_hdac_regmap_exit);
ALSA: hda - Allow driver to add vendor-specific verbs for regmap Codecs may have own vendor-specific verbs, and we need to allow each driver to give such verbs for cached accesses. Here a verb can be put into a single array and looked through it at readable and writeable callbacks. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 04:29:03 -07:00
/**
* snd_hdac_regmap_add_vendor_verb - add a vendor-specific verb to regmap
* @codec: the codec object
* @verb: verb to allow accessing via regmap
*
* Returns zero for success or a negative error code.
*/
int snd_hdac_regmap_add_vendor_verb(struct hdac_device *codec,
unsigned int verb)
{
unsigned int *p = snd_array_new(&codec->vendor_verbs);
if (!p)
return -ENOMEM;
ALSA: hda - set GET bit when adding a vendor verb to the codec regmap Some HD-A codecs may add their own vendor 'set' verb to the regmap, thru func snd_hdac_add_vendor_verb(). This patch sets the GET bit (bit 11) when adding the verb so that its peer vendor 'get' verb is actually added. This can avoid I/O error when writing the 'set' verb thru remap, since HD-A regmap internally looks up a writable vendor verb with GET bit set at first. Signed-off-by: Mengdong Lin <mengdong.lin@intel.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-04-13 21:25:36 -06:00
*p = verb | 0x800; /* set GET bit */
ALSA: hda - Allow driver to add vendor-specific verbs for regmap Codecs may have own vendor-specific verbs, and we need to allow each driver to give such verbs for cached accesses. Here a verb can be put into a single array and looked through it at readable and writeable callbacks. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-26 04:29:03 -07:00
return 0;
}
EXPORT_SYMBOL_GPL(snd_hdac_regmap_add_vendor_verb);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
/*
* helper functions
*/
/* write a pseudo-register value (w/o power sequence) */
static int reg_raw_write(struct hdac_device *codec, unsigned int reg,
unsigned int val)
{
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
int err;
mutex_lock(&codec->regmap_lock);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
if (!codec->regmap)
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
err = hda_reg_write(codec, reg, val);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
else
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
err = regmap_write(codec->regmap, reg, val);
mutex_unlock(&codec->regmap_lock);
return err;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
/* a helper macro to call @func_call; retry with power-up if failed */
#define CALL_RAW_FUNC(codec, func_call) \
({ \
int _err = func_call; \
if (_err == -EAGAIN) { \
_err = snd_hdac_power_up_pm(codec); \
if (_err >= 0) \
_err = func_call; \
snd_hdac_power_down_pm(codec); \
} \
_err;})
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
/**
* snd_hdac_regmap_write_raw - write a pseudo register with power mgmt
* @codec: the codec object
* @reg: pseudo register
* @val: value to write
*
* Returns zero if successful or a negative error code.
*/
int snd_hdac_regmap_write_raw(struct hdac_device *codec, unsigned int reg,
unsigned int val)
{
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
return CALL_RAW_FUNC(codec, reg_raw_write(codec, reg, val));
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
EXPORT_SYMBOL_GPL(snd_hdac_regmap_write_raw);
static int reg_raw_read(struct hdac_device *codec, unsigned int reg,
ALSA: hda - Fix possible race on regmap bypass flip HD-audio driver uses regmap cache bypass feature for reading a raw value without the cache. But this is racy since both the cached and the uncached reads may occur concurrently. The former is done via the normal control API access while the latter comes from the proc file read. Even though the regmap itself has the protection against the concurrent accesses, the flag set/reset is done without the protection, so it may lead to inconsistent state of bypass flag that doesn't match with the current read and occasionally result in a kernel WARNING like: WARNING: CPU: 3 PID: 2731 at drivers/base/regmap/regcache.c:499 regcache_cache_only+0x78/0x93 One way to work around such a problem is to wrap with a mutex. But in this case, the solution is simpler: for the uncached read, we just skip the regmap and directly calls its accessor. The verb execution there is protected by itself, so basically it's safe to call individually. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=116171 Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-04-21 09:49:11 -06:00
unsigned int *val, bool uncached)
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
{
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
int err;
mutex_lock(&codec->regmap_lock);
ALSA: hda - Fix possible race on regmap bypass flip HD-audio driver uses regmap cache bypass feature for reading a raw value without the cache. But this is racy since both the cached and the uncached reads may occur concurrently. The former is done via the normal control API access while the latter comes from the proc file read. Even though the regmap itself has the protection against the concurrent accesses, the flag set/reset is done without the protection, so it may lead to inconsistent state of bypass flag that doesn't match with the current read and occasionally result in a kernel WARNING like: WARNING: CPU: 3 PID: 2731 at drivers/base/regmap/regcache.c:499 regcache_cache_only+0x78/0x93 One way to work around such a problem is to wrap with a mutex. But in this case, the solution is simpler: for the uncached read, we just skip the regmap and directly calls its accessor. The verb execution there is protected by itself, so basically it's safe to call individually. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=116171 Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-04-21 09:49:11 -06:00
if (uncached || !codec->regmap)
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
err = hda_reg_read(codec, reg, val);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
else
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
err = regmap_read(codec->regmap, reg, val);
mutex_unlock(&codec->regmap_lock);
return err;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
ALSA: hda - Fix possible race on regmap bypass flip HD-audio driver uses regmap cache bypass feature for reading a raw value without the cache. But this is racy since both the cached and the uncached reads may occur concurrently. The former is done via the normal control API access while the latter comes from the proc file read. Even though the regmap itself has the protection against the concurrent accesses, the flag set/reset is done without the protection, so it may lead to inconsistent state of bypass flag that doesn't match with the current read and occasionally result in a kernel WARNING like: WARNING: CPU: 3 PID: 2731 at drivers/base/regmap/regcache.c:499 regcache_cache_only+0x78/0x93 One way to work around such a problem is to wrap with a mutex. But in this case, the solution is simpler: for the uncached read, we just skip the regmap and directly calls its accessor. The verb execution there is protected by itself, so basically it's safe to call individually. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=116171 Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-04-21 09:49:11 -06:00
static int __snd_hdac_regmap_read_raw(struct hdac_device *codec,
unsigned int reg, unsigned int *val,
bool uncached)
{
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
return CALL_RAW_FUNC(codec, reg_raw_read(codec, reg, val, uncached));
ALSA: hda - Fix possible race on regmap bypass flip HD-audio driver uses regmap cache bypass feature for reading a raw value without the cache. But this is racy since both the cached and the uncached reads may occur concurrently. The former is done via the normal control API access while the latter comes from the proc file read. Even though the regmap itself has the protection against the concurrent accesses, the flag set/reset is done without the protection, so it may lead to inconsistent state of bypass flag that doesn't match with the current read and occasionally result in a kernel WARNING like: WARNING: CPU: 3 PID: 2731 at drivers/base/regmap/regcache.c:499 regcache_cache_only+0x78/0x93 One way to work around such a problem is to wrap with a mutex. But in this case, the solution is simpler: for the uncached read, we just skip the regmap and directly calls its accessor. The verb execution there is protected by itself, so basically it's safe to call individually. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=116171 Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-04-21 09:49:11 -06:00
}
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
/**
* snd_hdac_regmap_read_raw - read a pseudo register with power mgmt
* @codec: the codec object
* @reg: pseudo register
* @val: pointer to store the read value
*
* Returns zero if successful or a negative error code.
*/
int snd_hdac_regmap_read_raw(struct hdac_device *codec, unsigned int reg,
unsigned int *val)
{
ALSA: hda - Fix possible race on regmap bypass flip HD-audio driver uses regmap cache bypass feature for reading a raw value without the cache. But this is racy since both the cached and the uncached reads may occur concurrently. The former is done via the normal control API access while the latter comes from the proc file read. Even though the regmap itself has the protection against the concurrent accesses, the flag set/reset is done without the protection, so it may lead to inconsistent state of bypass flag that doesn't match with the current read and occasionally result in a kernel WARNING like: WARNING: CPU: 3 PID: 2731 at drivers/base/regmap/regcache.c:499 regcache_cache_only+0x78/0x93 One way to work around such a problem is to wrap with a mutex. But in this case, the solution is simpler: for the uncached read, we just skip the regmap and directly calls its accessor. The verb execution there is protected by itself, so basically it's safe to call individually. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=116171 Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-04-21 09:49:11 -06:00
return __snd_hdac_regmap_read_raw(codec, reg, val, false);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
EXPORT_SYMBOL_GPL(snd_hdac_regmap_read_raw);
ALSA: hda - Fix possible race on regmap bypass flip HD-audio driver uses regmap cache bypass feature for reading a raw value without the cache. But this is racy since both the cached and the uncached reads may occur concurrently. The former is done via the normal control API access while the latter comes from the proc file read. Even though the regmap itself has the protection against the concurrent accesses, the flag set/reset is done without the protection, so it may lead to inconsistent state of bypass flag that doesn't match with the current read and occasionally result in a kernel WARNING like: WARNING: CPU: 3 PID: 2731 at drivers/base/regmap/regcache.c:499 regcache_cache_only+0x78/0x93 One way to work around such a problem is to wrap with a mutex. But in this case, the solution is simpler: for the uncached read, we just skip the regmap and directly calls its accessor. The verb execution there is protected by itself, so basically it's safe to call individually. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=116171 Signed-off-by: Takashi Iwai <tiwai@suse.de>
2016-04-21 09:49:11 -06:00
/* Works like snd_hdac_regmap_read_raw(), but this doesn't read from the
* cache but always via hda verbs.
*/
int snd_hdac_regmap_read_raw_uncached(struct hdac_device *codec,
unsigned int reg, unsigned int *val)
{
return __snd_hdac_regmap_read_raw(codec, reg, val, true);
}
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
static int reg_raw_update(struct hdac_device *codec, unsigned int reg,
unsigned int mask, unsigned int val)
{
unsigned int orig;
bool change;
int err;
mutex_lock(&codec->regmap_lock);
if (codec->regmap) {
err = regmap_update_bits_check(codec->regmap, reg, mask, val,
&change);
if (!err)
err = change ? 1 : 0;
} else {
err = hda_reg_read(codec, reg, &orig);
if (!err) {
val &= mask;
val |= orig & ~mask;
if (val != orig) {
err = hda_reg_write(codec, reg, val);
if (!err)
err = 1;
}
}
}
mutex_unlock(&codec->regmap_lock);
return err;
}
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
/**
* snd_hdac_regmap_update_raw - update a pseudo register with power mgmt
* @codec: the codec object
* @reg: pseudo register
* @mask: bit mask to udpate
* @val: value to update
*
* Returns zero if successful or a negative error code.
*/
int snd_hdac_regmap_update_raw(struct hdac_device *codec, unsigned int reg,
unsigned int mask, unsigned int val)
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
{
return CALL_RAW_FUNC(codec, reg_raw_update(codec, reg, mask, val));
}
EXPORT_SYMBOL_GPL(snd_hdac_regmap_update_raw);
static int reg_raw_update_once(struct hdac_device *codec, unsigned int reg,
unsigned int mask, unsigned int val)
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
{
unsigned int orig;
int err;
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
if (!codec->regmap)
return reg_raw_update(codec, reg, mask, val);
mutex_lock(&codec->regmap_lock);
regcache_cache_only(codec->regmap, true);
err = regmap_read(codec->regmap, reg, &orig);
regcache_cache_only(codec->regmap, false);
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
if (err < 0)
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
err = regmap_update_bits(codec->regmap, reg, mask, val);
mutex_unlock(&codec->regmap_lock);
return err;
ALSA: hda - Add regmap support This patch adds an infrastructure to support regmap-based verb accesses. Because o the asymmetric nature of HD-audio verbs, especially the amp verbs, we need to translate the verbs as a sort of pseudo registers to be mapped uniquely in regmap. In this patch, a pseudo register is built from the NID, the AC_VERB_GET_* and 8bit parameters, i.e. almost in the form to be sent to HD-audio bus but without codec address field. OTOH, for writing, the same pseudo register is translated to AC_VERB_SET_* automatically. The AC_VERB_SET_AMP_* verb is re-encoded from the corresponding AC_VERB_GET_AMP_* verb and parameter at writing. Some verbs has a single command for read but multiple for writes. A write for such a verb is split automatically to multiple verbs. The patch provides also a few handy helper functions. They are designed to be accessible even without regmap. When no regmap is set up (e.g. before the codec device instantiation), the direct hardware access is used. Also, it tries to avoid the unnecessary power-up. The power up/down sequence is performed only on demand. The codec driver needs to call snd_hdac_regmap_exit() and snd_hdac_regmap_exit() at probe and remove if it wants the regmap access. There is one flag added to hdac_device. When the flag lazy_cache is set, regmap helper ignores a write for a suspended device and returns as if it was actually written. It reduces the hardware access pretty much, e.g. when adjusting the mixer volume while in idle. This assumes that the driver will sync the cache later at resume properly, so use it carefully. Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-02-25 06:42:38 -07:00
}
ALSA: hda: Manage concurrent reg access more properly [ Upstream commit 1a462be52f4505a2719631fb5aa7bfdbd37bfd8d ] In the commit 8e85def5723e ("ALSA: hda: enable regmap internal locking"), we re-enabled the regmap lock due to the reported regression that showed the possible concurrent accesses. It was a temporary workaround, and there are still a few opened races even after the revert. In this patch, we cover those still opened windows with a proper mutex lock and disable the regmap internal lock again. First off, the patch introduces a new snd_hdac_device.regmap_lock mutex that is applied for each snd_hdac_regmap_*() call, including read, write and update helpers. The mutex is applied carefully so that it won't block the self-power-up procedure in the helper function. Also, this assures the protection for the accesses without regmap, too. The snd_hdac_regmap_update_raw() is refactored to use the standard regmap_update_bits_check() function instead of the open-code. The non-regmap case is still open-coded but it's an easy part. The all read and write operations are in the single mutex protection, so it's now race-free. In addition, a couple of new helper functions are added: snd_hdac_regmap_update_raw_once() and snd_hdac_regmap_sync(). Both are called from HD-audio legacy driver. The former is to initialize the given verb bits but only once when it's not initialized yet. Due to this condition, the function invokes regcache_cache_only(), and it's now performed inside the regmap_lock (formerly it was racy) too. The latter function is for simply invoking regcache_sync() inside the regmap_lock, which is called from the codec resume call path. Along with that, the HD-audio codec driver code is slightly modified / simplified to adapt those new functions. And finally, snd_hdac_regmap_read_raw(), *_write_raw(), etc are rewritten with the helper macro. It's just for simplification because the code logic is identical among all those functions. Tested-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200109090104.26073-1-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
2020-01-09 02:01:04 -07:00
/**
* snd_hdac_regmap_update_raw_once - initialize the register value only once
* @codec: the codec object
* @reg: pseudo register
* @mask: bit mask to update
* @val: value to update
*
* Performs the update of the register bits only once when the register
* hasn't been initialized yet. Used in HD-audio legacy driver.
* Returns zero if successful or a negative error code
*/
int snd_hdac_regmap_update_raw_once(struct hdac_device *codec, unsigned int reg,
unsigned int mask, unsigned int val)
{
return CALL_RAW_FUNC(codec, reg_raw_update_once(codec, reg, mask, val));
}
EXPORT_SYMBOL_GPL(snd_hdac_regmap_update_raw_once);
/**
* snd_hdac_regmap_sync - sync out the cached values for PM resume
* @codec: the codec object
*/
void snd_hdac_regmap_sync(struct hdac_device *codec)
{
if (codec->regmap) {
mutex_lock(&codec->regmap_lock);
regcache_sync(codec->regmap);
mutex_unlock(&codec->regmap_lock);
}
}
EXPORT_SYMBOL_GPL(snd_hdac_regmap_sync);