2006-01-09 10:32:44 -07:00
|
|
|
/*
|
|
|
|
|
*
|
|
|
|
|
* Support for audio capture
|
|
|
|
|
* PCI function #1 of the cx2388x.
|
|
|
|
|
*
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
* (c) 2007 Trent Piepho <xyzzy@speakeasy.org>
|
2006-01-09 10:32:44 -07:00
|
|
|
* (c) 2005,2006 Ricardo Cerqueira <v4l@cerqueira.org>
|
2006-04-03 04:53:40 -06:00
|
|
|
* (c) 2005 Mauro Carvalho Chehab <mchehab@infradead.org>
|
2006-01-09 10:32:44 -07:00
|
|
|
* Based on a dummy cx88 module by Gerd Knorr <kraxel@bytesex.org>
|
2007-10-15 01:50:19 -06:00
|
|
|
* Based on dummy.c by Jaroslav Kysela <perex@perex.cz>
|
2006-01-09 10:32:44 -07:00
|
|
|
*
|
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
|
* (at your option) any later version.
|
|
|
|
|
*
|
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
|
*
|
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
|
* along with this program; if not, write to the Free Software
|
|
|
|
|
* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include <linux/module.h>
|
|
|
|
|
#include <linux/init.h>
|
|
|
|
|
#include <linux/device.h>
|
|
|
|
|
#include <linux/interrupt.h>
|
2007-08-23 22:06:36 -06:00
|
|
|
#include <linux/vmalloc.h>
|
2007-05-06 15:51:55 -06:00
|
|
|
#include <linux/dma-mapping.h>
|
2007-08-18 19:54:49 -06:00
|
|
|
#include <linux/pci.h>
|
2007-05-06 15:51:55 -06:00
|
|
|
|
2006-01-09 10:32:44 -07:00
|
|
|
#include <asm/delay.h>
|
|
|
|
|
#include <sound/driver.h>
|
|
|
|
|
#include <sound/core.h>
|
|
|
|
|
#include <sound/pcm.h>
|
|
|
|
|
#include <sound/pcm_params.h>
|
|
|
|
|
#include <sound/control.h>
|
|
|
|
|
#include <sound/initval.h>
|
|
|
|
|
|
|
|
|
|
#include "cx88.h"
|
|
|
|
|
#include "cx88-reg.h"
|
|
|
|
|
|
|
|
|
|
#define dprintk(level,fmt, arg...) if (debug >= level) \
|
|
|
|
|
printk(KERN_INFO "%s/1: " fmt, chip->core->name , ## arg)
|
|
|
|
|
|
|
|
|
|
#define dprintk_core(level,fmt, arg...) if (debug >= level) \
|
|
|
|
|
printk(KERN_DEBUG "%s/1: " fmt, chip->core->name , ## arg)
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
Data type declarations - Can be moded to a header file later
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
struct cx88_audio_dev {
|
|
|
|
|
struct cx88_core *core;
|
|
|
|
|
struct cx88_dmaqueue q;
|
|
|
|
|
|
|
|
|
|
/* pci i/o */
|
|
|
|
|
struct pci_dev *pci;
|
|
|
|
|
|
|
|
|
|
/* audio controls */
|
|
|
|
|
int irq;
|
|
|
|
|
|
2006-01-13 10:48:06 -07:00
|
|
|
struct snd_card *card;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
spinlock_t reg_lock;
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
atomic_t count;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
unsigned int dma_size;
|
|
|
|
|
unsigned int period_size;
|
|
|
|
|
unsigned int num_periods;
|
|
|
|
|
|
2007-08-23 13:37:49 -06:00
|
|
|
struct videobuf_dmabuf *dma_risc;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
struct cx88_buffer *buf;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
struct snd_pcm_substream *substream;
|
2006-01-09 10:32:44 -07:00
|
|
|
};
|
|
|
|
|
typedef struct cx88_audio_dev snd_cx88_card_t;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
Module global static vars
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
static int index[SNDRV_CARDS] = SNDRV_DEFAULT_IDX; /* Index 0-MAX */
|
|
|
|
|
static char *id[SNDRV_CARDS] = SNDRV_DEFAULT_STR; /* ID for this card */
|
|
|
|
|
static int enable[SNDRV_CARDS] = {1, [1 ... (SNDRV_CARDS - 1)] = 1};
|
|
|
|
|
|
|
|
|
|
module_param_array(enable, bool, NULL, 0444);
|
|
|
|
|
MODULE_PARM_DESC(enable, "Enable cx88x soundcard. default enabled.");
|
|
|
|
|
|
|
|
|
|
module_param_array(index, int, NULL, 0444);
|
|
|
|
|
MODULE_PARM_DESC(index, "Index value for cx88x capture interface(s).");
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
Module macros
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
MODULE_DESCRIPTION("ALSA driver module for cx2388x based TV cards");
|
|
|
|
|
MODULE_AUTHOR("Ricardo Cerqueira");
|
2006-04-03 04:53:40 -06:00
|
|
|
MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@infradead.org>");
|
2006-01-09 10:32:44 -07:00
|
|
|
MODULE_LICENSE("GPL");
|
|
|
|
|
MODULE_SUPPORTED_DEVICE("{{Conexant,23881},"
|
|
|
|
|
"{{Conexant,23882},"
|
|
|
|
|
"{{Conexant,23883}");
|
2006-01-13 09:10:19 -07:00
|
|
|
static unsigned int debug;
|
2006-01-09 10:32:44 -07:00
|
|
|
module_param(debug,int,0644);
|
|
|
|
|
MODULE_PARM_DESC(debug,"enable debug messages");
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
Module specific funtions
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* BOARD Specific: Sets audio DMA
|
|
|
|
|
*/
|
|
|
|
|
|
2006-02-07 01:49:14 -07:00
|
|
|
static int _cx88_start_audio_dma(snd_cx88_card_t *chip)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
struct cx88_buffer *buf = chip->buf;
|
|
|
|
|
struct cx88_core *core=chip->core;
|
|
|
|
|
struct sram_channel *audio_ch = &cx88_sram_channels[SRAM_CH25];
|
|
|
|
|
|
2007-08-18 19:09:42 -06:00
|
|
|
/* Make sure RISC/FIFO are off before changing FIFO/RISC settings */
|
|
|
|
|
cx_clear(MO_AUD_DMACNTRL, 0x11);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
/* setup fifo + format - out channel */
|
2007-08-18 19:09:42 -06:00
|
|
|
cx88_sram_channel_setup(chip->core, audio_ch, buf->bpl, buf->risc.dma);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
/* sets bpl size */
|
|
|
|
|
cx_write(MO_AUDD_LNGTH, buf->bpl);
|
|
|
|
|
|
|
|
|
|
/* reset counter */
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
cx_write(MO_AUDD_GPCNTRL, GP_COUNT_CONTROL_RESET);
|
|
|
|
|
atomic_set(&chip->count, 0);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
dprintk(1, "Start audio DMA, %d B/line, %d lines/FIFO, %d periods, %d "
|
|
|
|
|
"byte buffer\n", buf->bpl, cx_read(audio_ch->cmds_start + 8)>>1,
|
2007-08-18 19:09:42 -06:00
|
|
|
chip->num_periods, buf->bpl * chip->num_periods);
|
|
|
|
|
|
2006-01-09 10:32:44 -07:00
|
|
|
/* Enables corresponding bits at AUD_INT_STAT */
|
2007-08-18 19:09:42 -06:00
|
|
|
cx_write(MO_AUD_INTMSK, AUD_INT_OPC_ERR | AUD_INT_DN_SYNC |
|
|
|
|
|
AUD_INT_DN_RISCI2 | AUD_INT_DN_RISCI1);
|
|
|
|
|
|
|
|
|
|
/* Clean any pending interrupt bits already set */
|
|
|
|
|
cx_write(MO_AUD_INTSTAT, ~0);
|
|
|
|
|
|
|
|
|
|
/* enable audio irqs */
|
|
|
|
|
cx_set(MO_PCI_INTMSK, chip->core->pci_irqmask | PCI_INT_AUDINT);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
/* start dma */
|
|
|
|
|
cx_set(MO_DEV_CNTRL2, (1<<5)); /* Enables Risc Processor */
|
|
|
|
|
cx_set(MO_AUD_DMACNTRL, 0x11); /* audio downstream FIFO and RISC enable */
|
|
|
|
|
|
|
|
|
|
if (debug)
|
2007-08-18 19:09:42 -06:00
|
|
|
cx88_sram_channel_dump(chip->core, audio_ch);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* BOARD Specific: Resets audio DMA
|
|
|
|
|
*/
|
2006-02-07 01:49:14 -07:00
|
|
|
static int _cx88_stop_audio_dma(snd_cx88_card_t *chip)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
struct cx88_core *core=chip->core;
|
|
|
|
|
dprintk(1, "Stopping audio DMA\n");
|
|
|
|
|
|
|
|
|
|
/* stop dma */
|
|
|
|
|
cx_clear(MO_AUD_DMACNTRL, 0x11);
|
|
|
|
|
|
|
|
|
|
/* disable irqs */
|
2007-08-18 03:57:55 -06:00
|
|
|
cx_clear(MO_PCI_INTMSK, PCI_INT_AUDINT);
|
2007-08-18 19:09:42 -06:00
|
|
|
cx_clear(MO_AUD_INTMSK, AUD_INT_OPC_ERR | AUD_INT_DN_SYNC |
|
|
|
|
|
AUD_INT_DN_RISCI2 | AUD_INT_DN_RISCI1);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
if (debug)
|
|
|
|
|
cx88_sram_channel_dump(chip->core, &cx88_sram_channels[SRAM_CH25]);
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
#define MAX_IRQ_LOOP 50
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* BOARD Specific: IRQ dma bits
|
|
|
|
|
*/
|
|
|
|
|
static char *cx88_aud_irqs[32] = {
|
|
|
|
|
"dn_risci1", "up_risci1", "rds_dn_risc1", /* 0-2 */
|
|
|
|
|
NULL, /* reserved */
|
|
|
|
|
"dn_risci2", "up_risci2", "rds_dn_risc2", /* 4-6 */
|
|
|
|
|
NULL, /* reserved */
|
|
|
|
|
"dnf_of", "upf_uf", "rds_dnf_uf", /* 8-10 */
|
|
|
|
|
NULL, /* reserved */
|
|
|
|
|
"dn_sync", "up_sync", "rds_dn_sync", /* 12-14 */
|
|
|
|
|
NULL, /* reserved */
|
|
|
|
|
"opc_err", "par_err", "rip_err", /* 16-18 */
|
|
|
|
|
"pci_abort", "ber_irq", "mchg_irq" /* 19-21 */
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* BOARD Specific: Threats IRQ audio specific calls
|
|
|
|
|
*/
|
|
|
|
|
static void cx8801_aud_irq(snd_cx88_card_t *chip)
|
|
|
|
|
{
|
|
|
|
|
struct cx88_core *core = chip->core;
|
|
|
|
|
u32 status, mask;
|
|
|
|
|
|
|
|
|
|
status = cx_read(MO_AUD_INTSTAT);
|
|
|
|
|
mask = cx_read(MO_AUD_INTMSK);
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
if (0 == (status & mask))
|
2006-01-09 10:32:44 -07:00
|
|
|
return;
|
|
|
|
|
cx_write(MO_AUD_INTSTAT, status);
|
|
|
|
|
if (debug > 1 || (status & mask & ~0xff))
|
|
|
|
|
cx88_print_irqbits(core->name, "irq aud",
|
2007-03-29 05:47:04 -06:00
|
|
|
cx88_aud_irqs, ARRAY_SIZE(cx88_aud_irqs),
|
|
|
|
|
status, mask);
|
2006-01-09 10:32:44 -07:00
|
|
|
/* risc op code error */
|
2007-08-18 19:09:42 -06:00
|
|
|
if (status & AUD_INT_OPC_ERR) {
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
printk(KERN_WARNING "%s/1: Audio risc op code error\n",core->name);
|
2006-01-09 10:32:44 -07:00
|
|
|
cx_clear(MO_AUD_DMACNTRL, 0x11);
|
|
|
|
|
cx88_sram_channel_dump(core, &cx88_sram_channels[SRAM_CH25]);
|
|
|
|
|
}
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
if (status & AUD_INT_DN_SYNC) {
|
|
|
|
|
dprintk(1, "Downstream sync error\n");
|
|
|
|
|
cx_write(MO_AUDD_GPCNTRL, GP_COUNT_CONTROL_RESET);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2006-01-09 10:32:44 -07:00
|
|
|
/* risc1 downstream */
|
2007-08-18 19:09:42 -06:00
|
|
|
if (status & AUD_INT_DN_RISCI1) {
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
atomic_set(&chip->count, cx_read(MO_AUDD_GPCNT));
|
2006-01-09 10:32:44 -07:00
|
|
|
snd_pcm_period_elapsed(chip->substream);
|
|
|
|
|
}
|
|
|
|
|
/* FIXME: Any other status should deserve a special handling? */
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* BOARD Specific: Handles IRQ calls
|
|
|
|
|
*/
|
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers
Maintain a per-CPU global "struct pt_regs *" variable which can be used instead
of passing regs around manually through all ~1800 interrupt handlers in the
Linux kernel.
The regs pointer is used in few places, but it potentially costs both stack
space and code to pass it around. On the FRV arch, removing the regs parameter
from all the genirq function results in a 20% speed up of the IRQ exit path
(ie: from leaving timer_interrupt() to leaving do_IRQ()).
Where appropriate, an arch may override the generic storage facility and do
something different with the variable. On FRV, for instance, the address is
maintained in GR28 at all times inside the kernel as part of general exception
handling.
Having looked over the code, it appears that the parameter may be handed down
through up to twenty or so layers of functions. Consider a USB character
device attached to a USB hub, attached to a USB controller that posts its
interrupts through a cascaded auxiliary interrupt controller. A character
device driver may want to pass regs to the sysrq handler through the input
layer which adds another few layers of parameter passing.
I've build this code with allyesconfig for x86_64 and i386. I've runtested the
main part of the code on FRV and i386, though I can't test most of the drivers.
I've also done partial conversion for powerpc and MIPS - these at least compile
with minimal configurations.
This will affect all archs. Mostly the changes should be relatively easy.
Take do_IRQ(), store the regs pointer at the beginning, saving the old one:
struct pt_regs *old_regs = set_irq_regs(regs);
And put the old one back at the end:
set_irq_regs(old_regs);
Don't pass regs through to generic_handle_irq() or __do_IRQ().
In timer_interrupt(), this sort of change will be necessary:
- update_process_times(user_mode(regs));
- profile_tick(CPU_PROFILING, regs);
+ update_process_times(user_mode(get_irq_regs()));
+ profile_tick(CPU_PROFILING);
I'd like to move update_process_times()'s use of get_irq_regs() into itself,
except that i386, alone of the archs, uses something other than user_mode().
Some notes on the interrupt handling in the drivers:
(*) input_dev() is now gone entirely. The regs pointer is no longer stored in
the input_dev struct.
(*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does
something different depending on whether it's been supplied with a regs
pointer or not.
(*) Various IRQ handler function pointers have been moved to type
irq_handler_t.
Signed-Off-By: David Howells <dhowells@redhat.com>
(cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 07:55:46 -06:00
|
|
|
static irqreturn_t cx8801_irq(int irq, void *dev_id)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip = dev_id;
|
|
|
|
|
struct cx88_core *core = chip->core;
|
|
|
|
|
u32 status;
|
|
|
|
|
int loop, handled = 0;
|
|
|
|
|
|
|
|
|
|
for (loop = 0; loop < MAX_IRQ_LOOP; loop++) {
|
2007-08-18 03:57:55 -06:00
|
|
|
status = cx_read(MO_PCI_INTSTAT) &
|
|
|
|
|
(core->pci_irqmask | PCI_INT_AUDINT);
|
2006-01-09 10:32:44 -07:00
|
|
|
if (0 == status)
|
|
|
|
|
goto out;
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
dprintk(3, "cx8801_irq loop %d/%d, status %x\n",
|
|
|
|
|
loop, MAX_IRQ_LOOP, status);
|
2006-01-09 10:32:44 -07:00
|
|
|
handled = 1;
|
|
|
|
|
cx_write(MO_PCI_INTSTAT, status);
|
|
|
|
|
|
2007-08-18 04:02:26 -06:00
|
|
|
if (status & core->pci_irqmask)
|
|
|
|
|
cx88_core_irq(core, status);
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
if (status & PCI_INT_AUDINT)
|
2006-01-09 10:32:44 -07:00
|
|
|
cx8801_aud_irq(chip);
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
}
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
if (MAX_IRQ_LOOP == loop) {
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
printk(KERN_ERR
|
|
|
|
|
"%s/1: IRQ loop detected, disabling interrupts\n",
|
2006-01-09 10:32:44 -07:00
|
|
|
core->name);
|
2007-08-18 03:57:55 -06:00
|
|
|
cx_clear(MO_PCI_INTMSK, PCI_INT_AUDINT);
|
2006-01-09 10:32:44 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
out:
|
|
|
|
|
return IRQ_RETVAL(handled);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
static int dsp_buffer_free(snd_cx88_card_t *chip)
|
|
|
|
|
{
|
|
|
|
|
BUG_ON(!chip->dma_size);
|
|
|
|
|
|
|
|
|
|
dprintk(2,"Freeing buffer\n");
|
2007-08-23 13:37:49 -06:00
|
|
|
videobuf_pci_dma_unmap(chip->pci, chip->dma_risc);
|
|
|
|
|
videobuf_dma_free(chip->dma_risc);
|
2006-01-09 10:32:44 -07:00
|
|
|
btcx_riscmem_free(chip->pci,&chip->buf->risc);
|
|
|
|
|
kfree(chip->buf);
|
|
|
|
|
|
2007-08-23 13:37:49 -06:00
|
|
|
chip->dma_risc = NULL;
|
2006-01-09 10:32:44 -07:00
|
|
|
chip->dma_size = 0;
|
|
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
return 0;
|
2006-01-09 10:32:44 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
ALSA PCM Interface
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Digital hardware definition
|
|
|
|
|
*/
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
#define DEFAULT_FIFO_SIZE 4096
|
2006-01-13 10:48:06 -07:00
|
|
|
static struct snd_pcm_hardware snd_cx88_digital_hw = {
|
2006-01-09 10:32:44 -07:00
|
|
|
.info = SNDRV_PCM_INFO_MMAP |
|
|
|
|
|
SNDRV_PCM_INFO_INTERLEAVED |
|
|
|
|
|
SNDRV_PCM_INFO_BLOCK_TRANSFER |
|
|
|
|
|
SNDRV_PCM_INFO_MMAP_VALID,
|
|
|
|
|
.formats = SNDRV_PCM_FMTBIT_S16_LE,
|
|
|
|
|
|
|
|
|
|
.rates = SNDRV_PCM_RATE_48000,
|
|
|
|
|
.rate_min = 48000,
|
|
|
|
|
.rate_max = 48000,
|
2007-08-18 18:01:40 -06:00
|
|
|
.channels_min = 2,
|
2006-01-09 10:32:44 -07:00
|
|
|
.channels_max = 2,
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
/* Analog audio output will be full of clicks and pops if there
|
|
|
|
|
are not exactly four lines in the SRAM FIFO buffer. */
|
|
|
|
|
.period_bytes_min = DEFAULT_FIFO_SIZE/4,
|
|
|
|
|
.period_bytes_max = DEFAULT_FIFO_SIZE/4,
|
|
|
|
|
.periods_min = 1,
|
|
|
|
|
.periods_max = 1024,
|
|
|
|
|
.buffer_bytes_max = (1024*1024),
|
2006-01-09 10:32:44 -07:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* audio pcm capture open callback
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_pcm_open(struct snd_pcm_substream *substream)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip = snd_pcm_substream_chip(substream);
|
2006-01-13 10:48:06 -07:00
|
|
|
struct snd_pcm_runtime *runtime = substream->runtime;
|
2006-01-09 10:32:44 -07:00
|
|
|
int err;
|
|
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
err = snd_pcm_hw_constraint_pow2(runtime, 0, SNDRV_PCM_HW_PARAM_PERIODS);
|
2006-01-09 10:32:44 -07:00
|
|
|
if (err < 0)
|
|
|
|
|
goto _error;
|
|
|
|
|
|
|
|
|
|
chip->substream = substream;
|
|
|
|
|
|
|
|
|
|
runtime->hw = snd_cx88_digital_hw;
|
|
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
if (cx88_sram_channels[SRAM_CH25].fifo_size != DEFAULT_FIFO_SIZE) {
|
|
|
|
|
unsigned int bpl = cx88_sram_channels[SRAM_CH25].fifo_size / 4;
|
|
|
|
|
bpl &= ~7; /* must be multiple of 8 */
|
|
|
|
|
runtime->hw.period_bytes_min = bpl;
|
|
|
|
|
runtime->hw.period_bytes_max = bpl;
|
|
|
|
|
}
|
|
|
|
|
|
2006-01-09 10:32:44 -07:00
|
|
|
return 0;
|
|
|
|
|
_error:
|
|
|
|
|
dprintk(1,"Error opening PCM!\n");
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* audio close callback
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_close(struct snd_pcm_substream *substream)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* hw_params callback
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_hw_params(struct snd_pcm_substream * substream,
|
|
|
|
|
struct snd_pcm_hw_params * hw_params)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip = snd_pcm_substream_chip(substream);
|
2007-08-23 13:37:49 -06:00
|
|
|
struct videobuf_dmabuf *dma;
|
|
|
|
|
|
2006-01-09 10:32:44 -07:00
|
|
|
struct cx88_buffer *buf;
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
int ret;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
if (substream->runtime->dma_area) {
|
|
|
|
|
dsp_buffer_free(chip);
|
|
|
|
|
substream->runtime->dma_area = NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
chip->period_size = params_period_bytes(hw_params);
|
|
|
|
|
chip->num_periods = params_periods(hw_params);
|
|
|
|
|
chip->dma_size = chip->period_size * params_periods(hw_params);
|
|
|
|
|
|
|
|
|
|
BUG_ON(!chip->dma_size);
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
BUG_ON(chip->num_periods & (chip->num_periods-1));
|
2006-01-09 10:32:44 -07:00
|
|
|
|
2007-08-23 13:37:49 -06:00
|
|
|
buf = videobuf_pci_alloc(sizeof(*buf));
|
2006-01-09 10:32:44 -07:00
|
|
|
if (NULL == buf)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
|
|
buf->vb.memory = V4L2_MEMORY_MMAP;
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
buf->vb.field = V4L2_FIELD_NONE;
|
2006-01-09 10:32:44 -07:00
|
|
|
buf->vb.width = chip->period_size;
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
buf->bpl = chip->period_size;
|
2006-01-09 10:32:44 -07:00
|
|
|
buf->vb.height = chip->num_periods;
|
|
|
|
|
buf->vb.size = chip->dma_size;
|
|
|
|
|
|
2007-08-23 13:37:49 -06:00
|
|
|
dma=videobuf_to_dma(&buf->vb);
|
|
|
|
|
videobuf_dma_init(dma);
|
|
|
|
|
ret = videobuf_dma_init_kernel(dma, PCI_DMA_FROMDEVICE,
|
2006-01-09 10:32:44 -07:00
|
|
|
(PAGE_ALIGN(buf->vb.size) >> PAGE_SHIFT));
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
if (ret < 0)
|
|
|
|
|
goto error;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
2007-08-23 13:37:49 -06:00
|
|
|
ret = videobuf_pci_dma_map(chip->pci,dma);
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
if (ret < 0)
|
|
|
|
|
goto error;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
2007-08-23 13:37:49 -06:00
|
|
|
ret = cx88_risc_databuffer(chip->pci, &buf->risc, dma->sglist,
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
buf->vb.width, buf->vb.height, 1);
|
|
|
|
|
if (ret < 0)
|
|
|
|
|
goto error;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
/* Loop back to start of program */
|
|
|
|
|
buf->risc.jmp[0] = cpu_to_le32(RISC_JUMP|RISC_IRQ1|RISC_CNT_INC);
|
2006-01-09 10:32:44 -07:00
|
|
|
buf->risc.jmp[1] = cpu_to_le32(buf->risc.dma);
|
|
|
|
|
|
|
|
|
|
buf->vb.state = STATE_PREPARED;
|
|
|
|
|
|
|
|
|
|
chip->buf = buf;
|
2007-08-23 13:37:49 -06:00
|
|
|
chip->dma_risc = dma;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
2007-08-23 13:37:49 -06:00
|
|
|
substream->runtime->dma_area = chip->dma_risc->vmalloc;
|
2007-08-23 22:06:36 -06:00
|
|
|
substream->runtime->dma_bytes = chip->dma_size;
|
|
|
|
|
substream->runtime->dma_addr = 0;
|
2006-01-09 10:32:44 -07:00
|
|
|
return 0;
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
|
|
|
|
|
error:
|
|
|
|
|
kfree(buf);
|
|
|
|
|
return ret;
|
2006-01-09 10:32:44 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* hw free callback
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_hw_free(struct snd_pcm_substream * substream)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
|
|
|
|
|
snd_cx88_card_t *chip = snd_pcm_substream_chip(substream);
|
|
|
|
|
|
|
|
|
|
if (substream->runtime->dma_area) {
|
|
|
|
|
dsp_buffer_free(chip);
|
|
|
|
|
substream->runtime->dma_area = NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* prepare callback
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_prepare(struct snd_pcm_substream *substream)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* trigger callback
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_card_trigger(struct snd_pcm_substream *substream, int cmd)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip = snd_pcm_substream_chip(substream);
|
|
|
|
|
int err;
|
|
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
/* Local interrupts are already disabled by ALSA */
|
2006-01-09 10:32:44 -07:00
|
|
|
spin_lock(&chip->reg_lock);
|
|
|
|
|
|
|
|
|
|
switch (cmd) {
|
|
|
|
|
case SNDRV_PCM_TRIGGER_START:
|
|
|
|
|
err=_cx88_start_audio_dma(chip);
|
|
|
|
|
break;
|
|
|
|
|
case SNDRV_PCM_TRIGGER_STOP:
|
|
|
|
|
err=_cx88_stop_audio_dma(chip);
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
err=-EINVAL;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
spin_unlock(&chip->reg_lock);
|
|
|
|
|
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* pointer callback
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static snd_pcm_uframes_t snd_cx88_pointer(struct snd_pcm_substream *substream)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip = snd_pcm_substream_chip(substream);
|
2006-01-13 10:48:06 -07:00
|
|
|
struct snd_pcm_runtime *runtime = substream->runtime;
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
u16 count;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
count = atomic_read(&chip->count);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
// dprintk(2, "%s - count %d (+%u), period %d, frame %lu\n", __FUNCTION__,
|
|
|
|
|
// count, new, count & (runtime->periods-1),
|
|
|
|
|
// runtime->period_size * (count & (runtime->periods-1)));
|
|
|
|
|
return runtime->period_size * (count & (runtime->periods-1));
|
2006-01-09 10:32:44 -07:00
|
|
|
}
|
|
|
|
|
|
2007-08-23 22:06:36 -06:00
|
|
|
/*
|
|
|
|
|
* page callback (needed for mmap)
|
|
|
|
|
*/
|
|
|
|
|
static struct page *snd_cx88_page(struct snd_pcm_substream *substream,
|
|
|
|
|
unsigned long offset)
|
|
|
|
|
{
|
|
|
|
|
void *pageptr = substream->runtime->dma_area + offset;
|
|
|
|
|
return vmalloc_to_page(pageptr);
|
|
|
|
|
}
|
|
|
|
|
|
2006-01-09 10:32:44 -07:00
|
|
|
/*
|
|
|
|
|
* operators
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static struct snd_pcm_ops snd_cx88_pcm_ops = {
|
2006-01-09 10:32:44 -07:00
|
|
|
.open = snd_cx88_pcm_open,
|
|
|
|
|
.close = snd_cx88_close,
|
|
|
|
|
.ioctl = snd_pcm_lib_ioctl,
|
|
|
|
|
.hw_params = snd_cx88_hw_params,
|
|
|
|
|
.hw_free = snd_cx88_hw_free,
|
|
|
|
|
.prepare = snd_cx88_prepare,
|
|
|
|
|
.trigger = snd_cx88_card_trigger,
|
|
|
|
|
.pointer = snd_cx88_pointer,
|
2007-08-23 22:06:36 -06:00
|
|
|
.page = snd_cx88_page,
|
2006-01-09 10:32:44 -07:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* create a PCM device
|
|
|
|
|
*/
|
|
|
|
|
static int __devinit snd_cx88_pcm(snd_cx88_card_t *chip, int device, char *name)
|
|
|
|
|
{
|
|
|
|
|
int err;
|
2006-01-13 10:48:06 -07:00
|
|
|
struct snd_pcm *pcm;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
err = snd_pcm_new(chip->card, name, device, 0, 1, &pcm);
|
|
|
|
|
if (err < 0)
|
|
|
|
|
return err;
|
|
|
|
|
pcm->private_data = chip;
|
|
|
|
|
strcpy(pcm->name, name);
|
|
|
|
|
snd_pcm_set_ops(pcm, SNDRV_PCM_STREAM_CAPTURE, &snd_cx88_pcm_ops);
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
CONTROL INTERFACE
|
|
|
|
|
****************************************************************************/
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_capture_volume_info(struct snd_kcontrol *kcontrol,
|
|
|
|
|
struct snd_ctl_elem_info *info)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
info->type = SNDRV_CTL_ELEM_TYPE_INTEGER;
|
2007-09-06 20:02:23 -06:00
|
|
|
info->count = 2;
|
2006-01-09 10:32:44 -07:00
|
|
|
info->value.integer.min = 0;
|
|
|
|
|
info->value.integer.max = 0x3f;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* OK - TODO: test it */
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_capture_volume_get(struct snd_kcontrol *kcontrol,
|
|
|
|
|
struct snd_ctl_elem_value *value)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip = snd_kcontrol_chip(kcontrol);
|
|
|
|
|
struct cx88_core *core=chip->core;
|
2007-09-06 20:02:23 -06:00
|
|
|
int vol = 0x3f - (cx_read(AUD_VOL_CTL) & 0x3f),
|
|
|
|
|
bal = cx_read(AUD_BAL_CTL);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
2007-09-06 20:02:23 -06:00
|
|
|
value->value.integer.value[(bal & 0x40) ? 0 : 1] = vol;
|
|
|
|
|
vol -= (bal & 0x3f);
|
|
|
|
|
value->value.integer.value[(bal & 0x40) ? 1 : 0] = vol < 0 ? 0 : vol;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* OK - TODO: test it */
|
2006-01-13 10:48:06 -07:00
|
|
|
static int snd_cx88_capture_volume_put(struct snd_kcontrol *kcontrol,
|
|
|
|
|
struct snd_ctl_elem_value *value)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip = snd_kcontrol_chip(kcontrol);
|
|
|
|
|
struct cx88_core *core=chip->core;
|
2007-09-06 20:02:23 -06:00
|
|
|
int v, b;
|
|
|
|
|
int changed = 0;
|
|
|
|
|
u32 old;
|
|
|
|
|
|
|
|
|
|
b = value->value.integer.value[1] - value->value.integer.value[0];
|
|
|
|
|
if (b < 0) {
|
|
|
|
|
v = 0x3f - value->value.integer.value[0];
|
|
|
|
|
b = (-b) | 0x40;
|
|
|
|
|
} else {
|
|
|
|
|
v = 0x3f - value->value.integer.value[1];
|
|
|
|
|
}
|
V4L/DVB (6083): cx88-alsa: Rework buffer handling
Rework the way the DMA buffer is handled and IRQs are generated.
ALSA uses a ring-buffer of multiple periods. Each period is supposed to
corrispond to one IRQ.
The existing driver was generating one interrupt per ring-buffer, as opposed
to per period. This meant that as soon as the IRQ was generated, the hardware
was already starting to re-write the beginning of the buffer. Since the DMA
happens on a per-line basis, there was only a narrow window to copy the data
out before the buffer was overwritten.
The cx88 core RISC program generator is modified so that it can set the IRQ
and counter flags to count every X lines of DMA transfer. This way we can
generate an interrupt every period instead of every full ring-buffer. Right
now only period of one line are supported, but it should be possible to
support longer periods. Note that a WRITE instruction generates an IRQ when
it starts, not when the transfer is finished. Thus to generate an IRQ when
line X is done, one must set the IRQ flag on the instruction that starts line
X+1, not the one that ends line X.
Change the line size so that there are four lines in the SRAM FIFO. If there
are not four lines, the analog output from the cx88's internal DACs is full of
clicks and pops.
Try to handle FIFO sync errors. Sometimes the chip generates many of these
errors before audio data starts. Up to 50 sync errors will be ignored and the
counter reset.
Have the IRQ handler save the RISC counter to the chip struct, and then have
the pointer callback use this to calculate the pointer position. We could
read the counter from the pointer callback, but sometimes the sync errors on
start up cause the counter to go crazy. ALSA sees this and thinks there has
been an overrun. The IRQ hander can avoid saving the counter position on
sync errors.
The chip "opened" flag wasn't necessary. ALSA won't try to open the same
substream multiple times. Probably this code was cut&pasted from the bt87x
driver, which has multiple sub-streams for one chip.
Do error checking for the videobuf mapping functions.
snd_card_cx88_runtime_free() is useless and can be deleted.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
2007-08-23 22:06:34 -06:00
|
|
|
/* Do we really know this will always be called with IRQs on? */
|
2006-01-09 10:32:44 -07:00
|
|
|
spin_lock_irq(&chip->reg_lock);
|
2007-09-06 20:02:23 -06:00
|
|
|
old = cx_read(AUD_VOL_CTL);
|
|
|
|
|
if (v != (old & 0x3f)) {
|
|
|
|
|
cx_write(AUD_VOL_CTL, (old & ~0x3f) | v);
|
|
|
|
|
changed = 1;
|
|
|
|
|
}
|
|
|
|
|
if (cx_read(AUD_BAL_CTL) != b) {
|
|
|
|
|
cx_write(AUD_BAL_CTL, b);
|
|
|
|
|
changed = 1;
|
|
|
|
|
}
|
2006-01-09 10:32:44 -07:00
|
|
|
spin_unlock_irq(&chip->reg_lock);
|
|
|
|
|
|
2007-09-06 20:02:23 -06:00
|
|
|
return changed;
|
2006-01-09 10:32:44 -07:00
|
|
|
}
|
|
|
|
|
|
2006-01-13 10:48:06 -07:00
|
|
|
static struct snd_kcontrol_new snd_cx88_capture_volume = {
|
2006-01-09 10:32:44 -07:00
|
|
|
.iface = SNDRV_CTL_ELEM_IFACE_MIXER,
|
|
|
|
|
.name = "Capture Volume",
|
|
|
|
|
.info = snd_cx88_capture_volume_info,
|
|
|
|
|
.get = snd_cx88_capture_volume_get,
|
|
|
|
|
.put = snd_cx88_capture_volume_put,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
Basic Flow for Sound Devices
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* PCI ID Table - 14f1:8801 and 14f1:8811 means function 1: Audio
|
|
|
|
|
* Only boards with eeprom and byte 1 at eeprom=1 have it
|
|
|
|
|
*/
|
|
|
|
|
|
2006-04-24 07:59:04 -06:00
|
|
|
static struct pci_device_id cx88_audio_pci_tbl[] __devinitdata = {
|
2006-01-09 10:32:44 -07:00
|
|
|
{0x14f1,0x8801,PCI_ANY_ID,PCI_ANY_ID,0,0,0},
|
|
|
|
|
{0x14f1,0x8811,PCI_ANY_ID,PCI_ANY_ID,0,0,0},
|
|
|
|
|
{0, }
|
|
|
|
|
};
|
|
|
|
|
MODULE_DEVICE_TABLE(pci, cx88_audio_pci_tbl);
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Chip-specific destructor
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
static int snd_cx88_free(snd_cx88_card_t *chip)
|
|
|
|
|
{
|
|
|
|
|
|
|
|
|
|
if (chip->irq >= 0){
|
|
|
|
|
synchronize_irq(chip->irq);
|
|
|
|
|
free_irq(chip->irq, chip);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
cx88_core_put(chip->core,chip->pci);
|
|
|
|
|
|
|
|
|
|
pci_disable_device(chip->pci);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Component Destructor
|
|
|
|
|
*/
|
2006-01-13 10:48:06 -07:00
|
|
|
static void snd_cx88_dev_free(struct snd_card * card)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip = card->private_data;
|
|
|
|
|
|
|
|
|
|
snd_cx88_free(chip);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Alsa Constructor - Component probe
|
|
|
|
|
*/
|
|
|
|
|
|
2006-01-13 09:10:19 -07:00
|
|
|
static int devno;
|
2006-01-13 10:48:06 -07:00
|
|
|
static int __devinit snd_cx88_create(struct snd_card *card,
|
|
|
|
|
struct pci_dev *pci,
|
|
|
|
|
snd_cx88_card_t **rchip)
|
2006-01-09 10:32:44 -07:00
|
|
|
{
|
|
|
|
|
snd_cx88_card_t *chip;
|
|
|
|
|
struct cx88_core *core;
|
|
|
|
|
int err;
|
2007-08-18 19:54:49 -06:00
|
|
|
unsigned char pci_lat;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
*rchip = NULL;
|
|
|
|
|
|
|
|
|
|
err = pci_enable_device(pci);
|
|
|
|
|
if (err < 0)
|
|
|
|
|
return err;
|
|
|
|
|
|
|
|
|
|
pci_set_master(pci);
|
|
|
|
|
|
|
|
|
|
chip = (snd_cx88_card_t *) card->private_data;
|
|
|
|
|
|
|
|
|
|
core = cx88_core_get(pci);
|
2006-03-14 08:12:39 -07:00
|
|
|
if (NULL == core) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
kfree (chip);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
2006-01-09 10:32:44 -07:00
|
|
|
|
2007-03-30 07:58:01 -06:00
|
|
|
if (!pci_dma_supported(pci,DMA_32BIT_MASK)) {
|
2006-01-09 10:32:44 -07:00
|
|
|
dprintk(0, "%s/1: Oops: no 32bit PCI DMA ???\n",core->name);
|
|
|
|
|
err = -EIO;
|
|
|
|
|
cx88_core_put(core,pci);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/* pci init */
|
|
|
|
|
chip->card = card;
|
|
|
|
|
chip->pci = pci;
|
|
|
|
|
chip->irq = -1;
|
|
|
|
|
spin_lock_init(&chip->reg_lock);
|
|
|
|
|
|
|
|
|
|
chip->core = core;
|
|
|
|
|
|
|
|
|
|
/* get irq */
|
|
|
|
|
err = request_irq(chip->pci->irq, cx8801_irq,
|
2006-07-01 20:29:37 -06:00
|
|
|
IRQF_SHARED | IRQF_DISABLED, chip->core->name, chip);
|
2006-01-09 10:32:44 -07:00
|
|
|
if (err < 0) {
|
|
|
|
|
dprintk(0, "%s: can't get IRQ %d\n",
|
|
|
|
|
chip->core->name, chip->pci->irq);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* print pci info */
|
2007-08-18 19:54:49 -06:00
|
|
|
pci_read_config_byte(pci, PCI_LATENCY_TIMER, &pci_lat);
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
dprintk(1,"ALSA %s/%i: found at %s, rev: %d, irq: %d, "
|
2006-06-12 16:16:52 -06:00
|
|
|
"latency: %d, mmio: 0x%llx\n", core->name, devno,
|
2007-08-18 19:54:49 -06:00
|
|
|
pci_name(pci), pci->revision, pci->irq,
|
|
|
|
|
pci_lat, (unsigned long long)pci_resource_start(pci,0));
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
chip->irq = pci->irq;
|
|
|
|
|
synchronize_irq(chip->irq);
|
|
|
|
|
|
|
|
|
|
snd_card_set_dev(card, &pci->dev);
|
|
|
|
|
|
|
|
|
|
*rchip = chip;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int __devinit cx88_audio_initdev(struct pci_dev *pci,
|
|
|
|
|
const struct pci_device_id *pci_id)
|
|
|
|
|
{
|
2006-01-13 10:48:06 -07:00
|
|
|
struct snd_card *card;
|
2006-01-09 10:32:44 -07:00
|
|
|
snd_cx88_card_t *chip;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
if (devno >= SNDRV_CARDS)
|
|
|
|
|
return (-ENODEV);
|
|
|
|
|
|
|
|
|
|
if (!enable[devno]) {
|
|
|
|
|
++devno;
|
|
|
|
|
return (-ENOENT);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
card = snd_card_new(index[devno], id[devno], THIS_MODULE, sizeof(snd_cx88_card_t));
|
|
|
|
|
if (!card)
|
|
|
|
|
return (-ENOMEM);
|
|
|
|
|
|
|
|
|
|
card->private_free = snd_cx88_dev_free;
|
|
|
|
|
|
|
|
|
|
err = snd_cx88_create(card, pci, &chip);
|
|
|
|
|
if (err < 0)
|
|
|
|
|
return (err);
|
|
|
|
|
|
|
|
|
|
err = snd_cx88_pcm(chip, 0, "CX88 Digital");
|
2007-09-06 20:02:23 -06:00
|
|
|
if (err < 0)
|
|
|
|
|
goto error;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
err = snd_ctl_add(card, snd_ctl_new1(&snd_cx88_capture_volume, chip));
|
2007-09-06 20:02:23 -06:00
|
|
|
if (err < 0)
|
|
|
|
|
goto error;
|
2006-01-09 10:32:44 -07:00
|
|
|
|
|
|
|
|
strcpy (card->driver, "CX88x");
|
|
|
|
|
sprintf(card->shortname, "Conexant CX%x", pci->device);
|
2006-06-12 16:16:52 -06:00
|
|
|
sprintf(card->longname, "%s at %#llx",
|
|
|
|
|
card->shortname,(unsigned long long)pci_resource_start(pci, 0));
|
2006-01-09 10:32:44 -07:00
|
|
|
strcpy (card->mixername, "CX88");
|
|
|
|
|
|
|
|
|
|
dprintk (0, "%s/%i: ALSA support for cx2388x boards\n",
|
|
|
|
|
card->driver,devno);
|
|
|
|
|
|
|
|
|
|
err = snd_card_register(card);
|
2007-09-06 20:02:23 -06:00
|
|
|
if (err < 0)
|
|
|
|
|
goto error;
|
2006-01-09 10:32:44 -07:00
|
|
|
pci_set_drvdata(pci,card);
|
|
|
|
|
|
|
|
|
|
devno++;
|
|
|
|
|
return 0;
|
2007-09-06 20:02:23 -06:00
|
|
|
|
|
|
|
|
error:
|
|
|
|
|
snd_card_free(card);
|
|
|
|
|
return err;
|
2006-01-09 10:32:44 -07:00
|
|
|
}
|
|
|
|
|
/*
|
|
|
|
|
* ALSA destructor
|
|
|
|
|
*/
|
|
|
|
|
static void __devexit cx88_audio_finidev(struct pci_dev *pci)
|
|
|
|
|
{
|
|
|
|
|
struct cx88_audio_dev *card = pci_get_drvdata(pci);
|
|
|
|
|
|
|
|
|
|
snd_card_free((void *)card);
|
|
|
|
|
|
|
|
|
|
pci_set_drvdata(pci, NULL);
|
|
|
|
|
|
|
|
|
|
devno--;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* PCI driver definition
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
static struct pci_driver cx88_audio_pci_driver = {
|
|
|
|
|
.name = "cx88_audio",
|
|
|
|
|
.id_table = cx88_audio_pci_tbl,
|
|
|
|
|
.probe = cx88_audio_initdev,
|
|
|
|
|
.remove = cx88_audio_finidev,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/****************************************************************************
|
|
|
|
|
LINUX MODULE INIT
|
|
|
|
|
****************************************************************************/
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* module init
|
|
|
|
|
*/
|
|
|
|
|
static int cx88_audio_init(void)
|
|
|
|
|
{
|
|
|
|
|
printk(KERN_INFO "cx2388x alsa driver version %d.%d.%d loaded\n",
|
|
|
|
|
(CX88_VERSION_CODE >> 16) & 0xff,
|
|
|
|
|
(CX88_VERSION_CODE >> 8) & 0xff,
|
|
|
|
|
CX88_VERSION_CODE & 0xff);
|
|
|
|
|
#ifdef SNAPSHOT
|
|
|
|
|
printk(KERN_INFO "cx2388x: snapshot date %04d-%02d-%02d\n",
|
|
|
|
|
SNAPSHOT/10000, (SNAPSHOT/100)%100, SNAPSHOT%100);
|
|
|
|
|
#endif
|
|
|
|
|
return pci_register_driver(&cx88_audio_pci_driver);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* module remove
|
|
|
|
|
*/
|
|
|
|
|
static void cx88_audio_fini(void)
|
|
|
|
|
{
|
|
|
|
|
|
|
|
|
|
pci_unregister_driver(&cx88_audio_pci_driver);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
module_init(cx88_audio_init);
|
|
|
|
|
module_exit(cx88_audio_fini);
|
|
|
|
|
|
|
|
|
|
/* ----------------------------------------------------------- */
|
|
|
|
|
/*
|
|
|
|
|
* Local variables:
|
|
|
|
|
* c-basic-offset: 8
|
|
|
|
|
* End:
|
|
|
|
|
*/
|