* [PATCH V3 0/2] ASoC: Move pcm writecombine dma buffer allocation to core
@ 2012-07-02 12:59 ` Laxman Dewangan
0 siblings, 0 replies; 6+ messages in thread
From: Laxman Dewangan @ 2012-07-02 12:59 UTC (permalink / raw)
To: tiwai, lrg, broonie, lars, swarren, perex, clemens
Cc: alsa-devel, linux-kernel, Laxman Dewangan
Some of the ARM based soc allocate the writecombine dma buffer for
pcm substreams. They have the same codes for managing this buffer.
Moving this to the core/pcm memory librray so that they can use that
directly.
Remove the code from Tegra PCM and use these new feature from library
function.
This is enabled only for ARM specific and can be extended to other
architecture if they support the writecombine dma buffer.
This patch is based on detail discussion on patch:
[PATCH] ASoC: snd_dmaengine: add common api for pcm_mmap
And suggestion from Lars and Takashi.
Changes from V1:
Takashi had send the sample code to integrate the writecombine with
dma coherant and this is based on that code.
Also change the tegra-pcm driver accordingly.
Changes from V2:
- Added some more comment for WAR for ARM specific.
- remove mmap initialization from tegra_pcm.
- Put back the dma mask initialisation in tegra_pcm.
Laxman Dewangan (2):
ALSA: Support writecombine DMA buffer memory page alloc/free
ASoC: tegra: use sound pcm library for substream dma buffer
include/sound/memalloc.h | 1 +
sound/core/memalloc.c | 70 ++++++++++++++++++++++++---
sound/core/pcm_native.c | 10 ++++
sound/soc/tegra/tegra_pcm.c | 112 ++++++++++--------------------------------
4 files changed, 100 insertions(+), 93 deletions(-)
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH V3 0/2] ASoC: Move pcm writecombine dma buffer allocation to core
@ 2012-07-02 12:59 ` Laxman Dewangan
0 siblings, 0 replies; 6+ messages in thread
From: Laxman Dewangan @ 2012-07-02 12:59 UTC (permalink / raw)
To: tiwai, lrg, broonie, lars, swarren, perex, clemens
Cc: alsa-devel, linux-kernel, Laxman Dewangan
Some of the ARM based soc allocate the writecombine dma buffer for
pcm substreams. They have the same codes for managing this buffer.
Moving this to the core/pcm memory librray so that they can use that
directly.
Remove the code from Tegra PCM and use these new feature from library
function.
This is enabled only for ARM specific and can be extended to other
architecture if they support the writecombine dma buffer.
This patch is based on detail discussion on patch:
[PATCH] ASoC: snd_dmaengine: add common api for pcm_mmap
And suggestion from Lars and Takashi.
Changes from V1:
Takashi had send the sample code to integrate the writecombine with
dma coherant and this is based on that code.
Also change the tegra-pcm driver accordingly.
Changes from V2:
- Added some more comment for WAR for ARM specific.
- remove mmap initialization from tegra_pcm.
- Put back the dma mask initialisation in tegra_pcm.
Laxman Dewangan (2):
ALSA: Support writecombine DMA buffer memory page alloc/free
ASoC: tegra: use sound pcm library for substream dma buffer
include/sound/memalloc.h | 1 +
sound/core/memalloc.c | 70 ++++++++++++++++++++++++---
sound/core/pcm_native.c | 10 ++++
sound/soc/tegra/tegra_pcm.c | 112 ++++++++++--------------------------------
4 files changed, 100 insertions(+), 93 deletions(-)
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH V3 1/2] ALSA: Support writecombine DMA buffer memory page alloc/free
2012-07-02 12:59 ` Laxman Dewangan
@ 2012-07-02 12:59 ` Laxman Dewangan
-1 siblings, 0 replies; 6+ messages in thread
From: Laxman Dewangan @ 2012-07-02 12:59 UTC (permalink / raw)
To: tiwai, lrg, broonie, lars, swarren, perex, clemens
Cc: alsa-devel, linux-kernel, Laxman Dewangan
Some of architecture like ARM support the writecombine DMA
buffer which is used for pcm substream DMA buffer.
Supporting writecombine DMA buffer allocation/free/mmap
through the sound library memory management.
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
---
Changes from V1:
Integrate the changes from Takashi's sample code for supporting
writecombine.
Changes from V2:
- Added some more comment for WAR for ARM specific.
include/sound/memalloc.h | 1 +
sound/core/memalloc.c | 86 +++++++++++++++++++++++++++++++++++++++++----
sound/core/pcm_native.c | 17 +++++++++
3 files changed, 96 insertions(+), 8 deletions(-)
diff --git a/include/sound/memalloc.h b/include/sound/memalloc.h
index c425062..0328e37 100644
--- a/include/sound/memalloc.h
+++ b/include/sound/memalloc.h
@@ -52,6 +52,7 @@ struct snd_dma_device {
#else
#define SNDRV_DMA_TYPE_DEV_SG SNDRV_DMA_TYPE_DEV /* no SG-buf support */
#endif
+#define SNDRV_DMA_TYPE_DEV_WC 4 /* writecombine page */
/*
* info for buffer allocation
diff --git a/sound/core/memalloc.c b/sound/core/memalloc.c
index 6915692..de99743 100644
--- a/sound/core/memalloc.c
+++ b/sound/core/memalloc.c
@@ -124,8 +124,9 @@ void snd_free_pages(void *ptr, size_t size)
*/
#ifdef CONFIG_HAS_DMA
-/* allocate the coherent DMA pages */
-static void *snd_malloc_dev_pages(struct device *dev, size_t size, dma_addr_t *dma)
+static void *_snd_malloc_dev_pages(struct device *dev, size_t size, dma_addr_t *dma,
+ void *(*alloc_func)(struct device *dev, size_t size,
+ dma_addr_t *dma, gfp_t gfp_flags))
{
int pg;
void *res;
@@ -138,16 +139,18 @@ static void *snd_malloc_dev_pages(struct device *dev, size_t size, dma_addr_t *d
| __GFP_COMP /* compound page lets parts be mapped */
| __GFP_NORETRY /* don't trigger OOM-killer */
| __GFP_NOWARN; /* no stack trace print - this call is non-critical */
- res = dma_alloc_coherent(dev, PAGE_SIZE << pg, dma, gfp_flags);
+ res = alloc_func(dev, PAGE_SIZE << pg, dma, gfp_flags);
if (res != NULL)
inc_snd_pages(pg);
return res;
}
-/* free the coherent DMA pages */
-static void snd_free_dev_pages(struct device *dev, size_t size, void *ptr,
- dma_addr_t dma)
+static void _snd_free_dev_pages(struct device *dev, size_t size, void *ptr,
+ dma_addr_t dma,
+ void (*free_func)(struct device *dev,
+ size_t size, void *vaddr,
+ dma_addr_t dma_handle))
{
int pg;
@@ -155,8 +158,57 @@ static void snd_free_dev_pages(struct device *dev, size_t size, void *ptr,
return;
pg = get_order(size);
dec_snd_pages(pg);
- dma_free_coherent(dev, PAGE_SIZE << pg, ptr, dma);
+ free_func(dev, PAGE_SIZE << pg, ptr, dma);
+}
+
+static void *_dma_alloc_coherent(struct device *dev, size_t size,
+ dma_addr_t *dma_handle, gfp_t flag)
+{
+ return dma_alloc_coherent(dev, size, dma_handle, flag);
+}
+
+static void _dma_free_coherent(struct device *dev, size_t size,
+ void *vaddr, dma_addr_t dma_handle)
+{
+ dma_free_coherent(dev, size, vaddr, dma_handle);
+}
+
+
+/* allocate the coherent DMA pages */
+static void *snd_malloc_dev_pages(struct device *dev, size_t size, dma_addr_t *dma)
+{
+ return _snd_malloc_dev_pages(dev, size, dma, _dma_alloc_coherent);
+}
+
+/* free the coherent DMA pages */
+static void snd_free_dev_pages(struct device *dev, size_t size, void *ptr,
+ dma_addr_t dma)
+{
+ _snd_free_dev_pages(dev, size, ptr, dma, _dma_free_coherent);
}
+
+
+/**
+ * Support writecombine DMA buffer allocation and deallocation.
+ * TODO: Currently all architectures do not support the writecombine
+ * dma buffer. It is supported by ARM and hence enabling for ARM.
+ * Once all architectures support, the restriction for ARM only can
+ * be removed.
+ */
+#ifdef CONFIG_ARM
+/* allocate the writecombine DMA pages */
+static void *snd_malloc_dev_wc_pages(struct device *dev, size_t size, dma_addr_t *dma)
+{
+ return _snd_malloc_dev_pages(dev, size, dma, dma_alloc_writecombine);
+}
+
+/* free the writecombine DMA pages */
+static void snd_free_dev_wc_pages(struct device *dev, size_t size, void *ptr,
+ dma_addr_t dma)
+{
+ _snd_free_dev_pages(dev, size, ptr, dma, dma_free_writecombine);
+}
+#endif
#endif /* CONFIG_HAS_DMA */
/*
@@ -200,6 +252,15 @@ int snd_dma_alloc_pages(int type, struct device *device, size_t size,
case SNDRV_DMA_TYPE_DEV:
dmab->area = snd_malloc_dev_pages(device, size, &dmab->addr);
break;
+/**
+ * TODO: Support writecombine DMA type for ARM only as all archs do not
+ * support this. This restriction can be remove once all archs support it.
+ */
+#ifdef CONFIG_ARM
+ case SNDRV_DMA_TYPE_DEV_WC:
+ dmab->area = snd_malloc_dev_wc_pages(device, size, &dmab->addr);
+ break;
+#endif
#endif
#ifdef CONFIG_SND_DMA_SGBUF
case SNDRV_DMA_TYPE_DEV_SG:
@@ -272,6 +333,15 @@ void snd_dma_free_pages(struct snd_dma_buffer *dmab)
case SNDRV_DMA_TYPE_DEV:
snd_free_dev_pages(dmab->dev.dev, dmab->bytes, dmab->area, dmab->addr);
break;
+/**
+ * TODO: Support writecombine DMA type for ARM only as all archs do not
+ * support this. This restriction can be remove once all archs support it.
+ */
+#ifdef CONFIG_ARM
+ case SNDRV_DMA_TYPE_DEV_WC:
+ snd_free_dev_wc_pages(dmab->dev.dev, dmab->bytes, dmab->area, dmab->addr);
+ break;
+#endif
#endif
#ifdef CONFIG_SND_DMA_SGBUF
case SNDRV_DMA_TYPE_DEV_SG:
@@ -378,7 +448,7 @@ static int snd_mem_proc_read(struct seq_file *seq, void *offset)
long pages = snd_allocated_pages >> (PAGE_SHIFT-12);
struct snd_mem_list *mem;
int devno;
- static char *types[] = { "UNKNOWN", "CONT", "DEV", "DEV-SG" };
+ static char *types[] = { "UNKNOWN", "CONT", "DEV", "DEV-SG" , "WC"};
mutex_lock(&list_mutex);
seq_printf(seq, "pages : %li bytes (%li pages per %likB)\n",
diff --git a/sound/core/pcm_native.c b/sound/core/pcm_native.c
index 53b5ada..827835b 100644
--- a/sound/core/pcm_native.c
+++ b/sound/core/pcm_native.c
@@ -3170,6 +3170,23 @@ int snd_pcm_lib_default_mmap(struct snd_pcm_substream *substream,
struct vm_area_struct *area)
{
area->vm_flags |= VM_RESERVED;
+
+/**
+ * Support writecombine DMA device type for mmap.
+ * TODO: Currently all architectures do not support the writecombine
+ * DMA buffer. It is supported by ARM and hence enabling for ARM.
+ * Once all architectures support, the restriction for ARM only can
+ * be removed.
+ */
+#ifdef CONFIG_ARM
+ if (substream->dma_buffer.dev.type == SNDRV_DMA_TYPE_DEV_WC)
+ return dma_mmap_writecombine(substream->dma_buffer.dev.dev,
+ area,
+ substream->runtime->dma_area,
+ substream->runtime->dma_addr,
+ area->vm_end - area->vm_start);
+#endif
+
#ifdef ARCH_HAS_DMA_MMAP_COHERENT
if (!substream->ops->page &&
substream->dma_buffer.dev.type == SNDRV_DMA_TYPE_DEV)
--
1.7.1.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH V3 1/2] ALSA: Support writecombine DMA buffer memory page alloc/free
@ 2012-07-02 12:59 ` Laxman Dewangan
0 siblings, 0 replies; 6+ messages in thread
From: Laxman Dewangan @ 2012-07-02 12:59 UTC (permalink / raw)
To: tiwai, lrg, broonie, lars, swarren, perex, clemens
Cc: alsa-devel, linux-kernel, Laxman Dewangan
Some of architecture like ARM support the writecombine DMA
buffer which is used for pcm substream DMA buffer.
Supporting writecombine DMA buffer allocation/free/mmap
through the sound library memory management.
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
---
Changes from V1:
Integrate the changes from Takashi's sample code for supporting
writecombine.
Changes from V2:
- Added some more comment for WAR for ARM specific.
include/sound/memalloc.h | 1 +
sound/core/memalloc.c | 86 +++++++++++++++++++++++++++++++++++++++++----
sound/core/pcm_native.c | 17 +++++++++
3 files changed, 96 insertions(+), 8 deletions(-)
diff --git a/include/sound/memalloc.h b/include/sound/memalloc.h
index c425062..0328e37 100644
--- a/include/sound/memalloc.h
+++ b/include/sound/memalloc.h
@@ -52,6 +52,7 @@ struct snd_dma_device {
#else
#define SNDRV_DMA_TYPE_DEV_SG SNDRV_DMA_TYPE_DEV /* no SG-buf support */
#endif
+#define SNDRV_DMA_TYPE_DEV_WC 4 /* writecombine page */
/*
* info for buffer allocation
diff --git a/sound/core/memalloc.c b/sound/core/memalloc.c
index 6915692..de99743 100644
--- a/sound/core/memalloc.c
+++ b/sound/core/memalloc.c
@@ -124,8 +124,9 @@ void snd_free_pages(void *ptr, size_t size)
*/
#ifdef CONFIG_HAS_DMA
-/* allocate the coherent DMA pages */
-static void *snd_malloc_dev_pages(struct device *dev, size_t size, dma_addr_t *dma)
+static void *_snd_malloc_dev_pages(struct device *dev, size_t size, dma_addr_t *dma,
+ void *(*alloc_func)(struct device *dev, size_t size,
+ dma_addr_t *dma, gfp_t gfp_flags))
{
int pg;
void *res;
@@ -138,16 +139,18 @@ static void *snd_malloc_dev_pages(struct device *dev, size_t size, dma_addr_t *d
| __GFP_COMP /* compound page lets parts be mapped */
| __GFP_NORETRY /* don't trigger OOM-killer */
| __GFP_NOWARN; /* no stack trace print - this call is non-critical */
- res = dma_alloc_coherent(dev, PAGE_SIZE << pg, dma, gfp_flags);
+ res = alloc_func(dev, PAGE_SIZE << pg, dma, gfp_flags);
if (res != NULL)
inc_snd_pages(pg);
return res;
}
-/* free the coherent DMA pages */
-static void snd_free_dev_pages(struct device *dev, size_t size, void *ptr,
- dma_addr_t dma)
+static void _snd_free_dev_pages(struct device *dev, size_t size, void *ptr,
+ dma_addr_t dma,
+ void (*free_func)(struct device *dev,
+ size_t size, void *vaddr,
+ dma_addr_t dma_handle))
{
int pg;
@@ -155,8 +158,57 @@ static void snd_free_dev_pages(struct device *dev, size_t size, void *ptr,
return;
pg = get_order(size);
dec_snd_pages(pg);
- dma_free_coherent(dev, PAGE_SIZE << pg, ptr, dma);
+ free_func(dev, PAGE_SIZE << pg, ptr, dma);
+}
+
+static void *_dma_alloc_coherent(struct device *dev, size_t size,
+ dma_addr_t *dma_handle, gfp_t flag)
+{
+ return dma_alloc_coherent(dev, size, dma_handle, flag);
+}
+
+static void _dma_free_coherent(struct device *dev, size_t size,
+ void *vaddr, dma_addr_t dma_handle)
+{
+ dma_free_coherent(dev, size, vaddr, dma_handle);
+}
+
+
+/* allocate the coherent DMA pages */
+static void *snd_malloc_dev_pages(struct device *dev, size_t size, dma_addr_t *dma)
+{
+ return _snd_malloc_dev_pages(dev, size, dma, _dma_alloc_coherent);
+}
+
+/* free the coherent DMA pages */
+static void snd_free_dev_pages(struct device *dev, size_t size, void *ptr,
+ dma_addr_t dma)
+{
+ _snd_free_dev_pages(dev, size, ptr, dma, _dma_free_coherent);
}
+
+
+/**
+ * Support writecombine DMA buffer allocation and deallocation.
+ * TODO: Currently all architectures do not support the writecombine
+ * dma buffer. It is supported by ARM and hence enabling for ARM.
+ * Once all architectures support, the restriction for ARM only can
+ * be removed.
+ */
+#ifdef CONFIG_ARM
+/* allocate the writecombine DMA pages */
+static void *snd_malloc_dev_wc_pages(struct device *dev, size_t size, dma_addr_t *dma)
+{
+ return _snd_malloc_dev_pages(dev, size, dma, dma_alloc_writecombine);
+}
+
+/* free the writecombine DMA pages */
+static void snd_free_dev_wc_pages(struct device *dev, size_t size, void *ptr,
+ dma_addr_t dma)
+{
+ _snd_free_dev_pages(dev, size, ptr, dma, dma_free_writecombine);
+}
+#endif
#endif /* CONFIG_HAS_DMA */
/*
@@ -200,6 +252,15 @@ int snd_dma_alloc_pages(int type, struct device *device, size_t size,
case SNDRV_DMA_TYPE_DEV:
dmab->area = snd_malloc_dev_pages(device, size, &dmab->addr);
break;
+/**
+ * TODO: Support writecombine DMA type for ARM only as all archs do not
+ * support this. This restriction can be remove once all archs support it.
+ */
+#ifdef CONFIG_ARM
+ case SNDRV_DMA_TYPE_DEV_WC:
+ dmab->area = snd_malloc_dev_wc_pages(device, size, &dmab->addr);
+ break;
+#endif
#endif
#ifdef CONFIG_SND_DMA_SGBUF
case SNDRV_DMA_TYPE_DEV_SG:
@@ -272,6 +333,15 @@ void snd_dma_free_pages(struct snd_dma_buffer *dmab)
case SNDRV_DMA_TYPE_DEV:
snd_free_dev_pages(dmab->dev.dev, dmab->bytes, dmab->area, dmab->addr);
break;
+/**
+ * TODO: Support writecombine DMA type for ARM only as all archs do not
+ * support this. This restriction can be remove once all archs support it.
+ */
+#ifdef CONFIG_ARM
+ case SNDRV_DMA_TYPE_DEV_WC:
+ snd_free_dev_wc_pages(dmab->dev.dev, dmab->bytes, dmab->area, dmab->addr);
+ break;
+#endif
#endif
#ifdef CONFIG_SND_DMA_SGBUF
case SNDRV_DMA_TYPE_DEV_SG:
@@ -378,7 +448,7 @@ static int snd_mem_proc_read(struct seq_file *seq, void *offset)
long pages = snd_allocated_pages >> (PAGE_SHIFT-12);
struct snd_mem_list *mem;
int devno;
- static char *types[] = { "UNKNOWN", "CONT", "DEV", "DEV-SG" };
+ static char *types[] = { "UNKNOWN", "CONT", "DEV", "DEV-SG" , "WC"};
mutex_lock(&list_mutex);
seq_printf(seq, "pages : %li bytes (%li pages per %likB)\n",
diff --git a/sound/core/pcm_native.c b/sound/core/pcm_native.c
index 53b5ada..827835b 100644
--- a/sound/core/pcm_native.c
+++ b/sound/core/pcm_native.c
@@ -3170,6 +3170,23 @@ int snd_pcm_lib_default_mmap(struct snd_pcm_substream *substream,
struct vm_area_struct *area)
{
area->vm_flags |= VM_RESERVED;
+
+/**
+ * Support writecombine DMA device type for mmap.
+ * TODO: Currently all architectures do not support the writecombine
+ * DMA buffer. It is supported by ARM and hence enabling for ARM.
+ * Once all architectures support, the restriction for ARM only can
+ * be removed.
+ */
+#ifdef CONFIG_ARM
+ if (substream->dma_buffer.dev.type == SNDRV_DMA_TYPE_DEV_WC)
+ return dma_mmap_writecombine(substream->dma_buffer.dev.dev,
+ area,
+ substream->runtime->dma_area,
+ substream->runtime->dma_addr,
+ area->vm_end - area->vm_start);
+#endif
+
#ifdef ARCH_HAS_DMA_MMAP_COHERENT
if (!substream->ops->page &&
substream->dma_buffer.dev.type == SNDRV_DMA_TYPE_DEV)
--
1.7.1.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH V3 2/2] ASoC: tegra: use sound pcm library for substream dma buffer
2012-07-02 12:59 ` Laxman Dewangan
@ 2012-07-02 12:59 ` Laxman Dewangan
-1 siblings, 0 replies; 6+ messages in thread
From: Laxman Dewangan @ 2012-07-02 12:59 UTC (permalink / raw)
To: tiwai, lrg, broonie, lars, swarren, perex, clemens
Cc: alsa-devel, linux-kernel, Laxman Dewangan
The sound pcm library support the writecombine DMA buffer
allocation/deallocation/mapping and hence using the APIs
form sound pcm library inplace of implementing the same
locally.
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
---
Changes from V1:
Changes done to use the existing APIs for sound pcm memory management
which is now supporting writecombine.
Changes from V2:
- remove mmap initialization from tegra_pcm.
- Put back the dma mask initialisation in tegra_pcm.
sound/soc/tegra/tegra_pcm.c | 98 +++++++++----------------------------------
1 files changed, 20 insertions(+), 78 deletions(-)
diff --git a/sound/soc/tegra/tegra_pcm.c b/sound/soc/tegra/tegra_pcm.c
index 127348d..1ed9df0 100644
--- a/sound/soc/tegra/tegra_pcm.c
+++ b/sound/soc/tegra/tegra_pcm.c
@@ -202,8 +202,14 @@ static int tegra_pcm_hw_params(struct snd_pcm_substream *substream,
{
struct snd_pcm_runtime *runtime = substream->runtime;
struct tegra_runtime_data *prtd = runtime->private_data;
+ int ret;
- snd_pcm_set_runtime_buffer(substream, &substream->dma_buffer);
+ ret = snd_pcm_lib_malloc_pages(substream,
+ params_buffer_bytes(params));
+ if (ret < 0) {
+ pr_err("page allocation failed, err %d\n", ret);
+ return ret;
+ }
prtd->dma_req[0].size = params_period_bytes(params);
prtd->dma_req[1].size = prtd->dma_req[0].size;
@@ -213,9 +219,7 @@ static int tegra_pcm_hw_params(struct snd_pcm_substream *substream,
static int tegra_pcm_hw_free(struct snd_pcm_substream *substream)
{
- snd_pcm_set_runtime_buffer(substream, NULL);
-
- return 0;
+ return snd_pcm_lib_free_pages(substream);
}
static int tegra_pcm_trigger(struct snd_pcm_substream *substream, int cmd)
@@ -263,18 +267,6 @@ static snd_pcm_uframes_t tegra_pcm_pointer(struct snd_pcm_substream *substream)
return prtd->period_index * runtime->period_size;
}
-
-static int tegra_pcm_mmap(struct snd_pcm_substream *substream,
- struct vm_area_struct *vma)
-{
- struct snd_pcm_runtime *runtime = substream->runtime;
-
- return dma_mmap_writecombine(substream->pcm->card->dev, vma,
- runtime->dma_area,
- runtime->dma_addr,
- runtime->dma_bytes);
-}
-
static struct snd_pcm_ops tegra_pcm_ops = {
.open = tegra_pcm_open,
.close = tegra_pcm_close,
@@ -283,87 +275,37 @@ static struct snd_pcm_ops tegra_pcm_ops = {
.hw_free = tegra_pcm_hw_free,
.trigger = tegra_pcm_trigger,
.pointer = tegra_pcm_pointer,
- .mmap = tegra_pcm_mmap,
};
-static int tegra_pcm_preallocate_dma_buffer(struct snd_pcm *pcm, int stream)
-{
- struct snd_pcm_substream *substream = pcm->streams[stream].substream;
- struct snd_dma_buffer *buf = &substream->dma_buffer;
- size_t size = tegra_pcm_hardware.buffer_bytes_max;
-
- buf->area = dma_alloc_writecombine(pcm->card->dev, size,
- &buf->addr, GFP_KERNEL);
- if (!buf->area)
- return -ENOMEM;
-
- buf->dev.type = SNDRV_DMA_TYPE_DEV;
- buf->dev.dev = pcm->card->dev;
- buf->private_data = NULL;
- buf->bytes = size;
-
- return 0;
-}
-
-static void tegra_pcm_deallocate_dma_buffer(struct snd_pcm *pcm, int stream)
-{
- struct snd_pcm_substream *substream;
- struct snd_dma_buffer *buf;
-
- substream = pcm->streams[stream].substream;
- if (!substream)
- return;
-
- buf = &substream->dma_buffer;
- if (!buf->area)
- return;
-
- dma_free_writecombine(pcm->card->dev, buf->bytes,
- buf->area, buf->addr);
- buf->area = NULL;
-}
-
static u64 tegra_dma_mask = DMA_BIT_MASK(32);
-
+#define TEGRA_PCM_PREALLOC_BUFFER (32 * 1024)
+#define TEGRA_PCM_PREALLOC_BUFFER_MAX (32 * 1024)
static int tegra_pcm_new(struct snd_soc_pcm_runtime *rtd)
{
- struct snd_card *card = rtd->card->snd_card;
struct snd_pcm *pcm = rtd->pcm;
- int ret = 0;
+ struct snd_card *card = rtd->card->snd_card;
+ int ret;
if (!card->dev->dma_mask)
card->dev->dma_mask = &tegra_dma_mask;
if (!card->dev->coherent_dma_mask)
card->dev->coherent_dma_mask = DMA_BIT_MASK(32);
- if (pcm->streams[SNDRV_PCM_STREAM_PLAYBACK].substream) {
- ret = tegra_pcm_preallocate_dma_buffer(pcm,
- SNDRV_PCM_STREAM_PLAYBACK);
- if (ret)
- goto err;
- }
-
- if (pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream) {
- ret = tegra_pcm_preallocate_dma_buffer(pcm,
- SNDRV_PCM_STREAM_CAPTURE);
- if (ret)
- goto err_free_play;
- }
-
- return 0;
-
-err_free_play:
- tegra_pcm_deallocate_dma_buffer(pcm, SNDRV_PCM_STREAM_PLAYBACK);
-err:
+ ret = snd_pcm_lib_preallocate_pages_for_all(pcm,
+ SNDRV_DMA_TYPE_DEV_WC, card->dev,
+ TEGRA_PCM_PREALLOC_BUFFER,
+ TEGRA_PCM_PREALLOC_BUFFER_MAX);
+ if (ret < 0)
+ dev_err(card->dev, "dma buffer pre-alloc failed err %d\n", ret);
return ret;
}
static void tegra_pcm_free(struct snd_pcm *pcm)
{
- tegra_pcm_deallocate_dma_buffer(pcm, SNDRV_PCM_STREAM_CAPTURE);
- tegra_pcm_deallocate_dma_buffer(pcm, SNDRV_PCM_STREAM_PLAYBACK);
+ snd_pcm_lib_preallocate_free_for_all(pcm);
}
+
static struct snd_soc_platform_driver tegra_pcm_platform = {
.ops = &tegra_pcm_ops,
.pcm_new = tegra_pcm_new,
--
1.7.1.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH V3 2/2] ASoC: tegra: use sound pcm library for substream dma buffer
@ 2012-07-02 12:59 ` Laxman Dewangan
0 siblings, 0 replies; 6+ messages in thread
From: Laxman Dewangan @ 2012-07-02 12:59 UTC (permalink / raw)
To: tiwai, lrg, broonie, lars, swarren, perex, clemens
Cc: alsa-devel, linux-kernel, Laxman Dewangan
The sound pcm library support the writecombine DMA buffer
allocation/deallocation/mapping and hence using the APIs
form sound pcm library inplace of implementing the same
locally.
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
---
Changes from V1:
Changes done to use the existing APIs for sound pcm memory management
which is now supporting writecombine.
Changes from V2:
- remove mmap initialization from tegra_pcm.
- Put back the dma mask initialisation in tegra_pcm.
sound/soc/tegra/tegra_pcm.c | 98 +++++++++----------------------------------
1 files changed, 20 insertions(+), 78 deletions(-)
diff --git a/sound/soc/tegra/tegra_pcm.c b/sound/soc/tegra/tegra_pcm.c
index 127348d..1ed9df0 100644
--- a/sound/soc/tegra/tegra_pcm.c
+++ b/sound/soc/tegra/tegra_pcm.c
@@ -202,8 +202,14 @@ static int tegra_pcm_hw_params(struct snd_pcm_substream *substream,
{
struct snd_pcm_runtime *runtime = substream->runtime;
struct tegra_runtime_data *prtd = runtime->private_data;
+ int ret;
- snd_pcm_set_runtime_buffer(substream, &substream->dma_buffer);
+ ret = snd_pcm_lib_malloc_pages(substream,
+ params_buffer_bytes(params));
+ if (ret < 0) {
+ pr_err("page allocation failed, err %d\n", ret);
+ return ret;
+ }
prtd->dma_req[0].size = params_period_bytes(params);
prtd->dma_req[1].size = prtd->dma_req[0].size;
@@ -213,9 +219,7 @@ static int tegra_pcm_hw_params(struct snd_pcm_substream *substream,
static int tegra_pcm_hw_free(struct snd_pcm_substream *substream)
{
- snd_pcm_set_runtime_buffer(substream, NULL);
-
- return 0;
+ return snd_pcm_lib_free_pages(substream);
}
static int tegra_pcm_trigger(struct snd_pcm_substream *substream, int cmd)
@@ -263,18 +267,6 @@ static snd_pcm_uframes_t tegra_pcm_pointer(struct snd_pcm_substream *substream)
return prtd->period_index * runtime->period_size;
}
-
-static int tegra_pcm_mmap(struct snd_pcm_substream *substream,
- struct vm_area_struct *vma)
-{
- struct snd_pcm_runtime *runtime = substream->runtime;
-
- return dma_mmap_writecombine(substream->pcm->card->dev, vma,
- runtime->dma_area,
- runtime->dma_addr,
- runtime->dma_bytes);
-}
-
static struct snd_pcm_ops tegra_pcm_ops = {
.open = tegra_pcm_open,
.close = tegra_pcm_close,
@@ -283,87 +275,37 @@ static struct snd_pcm_ops tegra_pcm_ops = {
.hw_free = tegra_pcm_hw_free,
.trigger = tegra_pcm_trigger,
.pointer = tegra_pcm_pointer,
- .mmap = tegra_pcm_mmap,
};
-static int tegra_pcm_preallocate_dma_buffer(struct snd_pcm *pcm, int stream)
-{
- struct snd_pcm_substream *substream = pcm->streams[stream].substream;
- struct snd_dma_buffer *buf = &substream->dma_buffer;
- size_t size = tegra_pcm_hardware.buffer_bytes_max;
-
- buf->area = dma_alloc_writecombine(pcm->card->dev, size,
- &buf->addr, GFP_KERNEL);
- if (!buf->area)
- return -ENOMEM;
-
- buf->dev.type = SNDRV_DMA_TYPE_DEV;
- buf->dev.dev = pcm->card->dev;
- buf->private_data = NULL;
- buf->bytes = size;
-
- return 0;
-}
-
-static void tegra_pcm_deallocate_dma_buffer(struct snd_pcm *pcm, int stream)
-{
- struct snd_pcm_substream *substream;
- struct snd_dma_buffer *buf;
-
- substream = pcm->streams[stream].substream;
- if (!substream)
- return;
-
- buf = &substream->dma_buffer;
- if (!buf->area)
- return;
-
- dma_free_writecombine(pcm->card->dev, buf->bytes,
- buf->area, buf->addr);
- buf->area = NULL;
-}
-
static u64 tegra_dma_mask = DMA_BIT_MASK(32);
-
+#define TEGRA_PCM_PREALLOC_BUFFER (32 * 1024)
+#define TEGRA_PCM_PREALLOC_BUFFER_MAX (32 * 1024)
static int tegra_pcm_new(struct snd_soc_pcm_runtime *rtd)
{
- struct snd_card *card = rtd->card->snd_card;
struct snd_pcm *pcm = rtd->pcm;
- int ret = 0;
+ struct snd_card *card = rtd->card->snd_card;
+ int ret;
if (!card->dev->dma_mask)
card->dev->dma_mask = &tegra_dma_mask;
if (!card->dev->coherent_dma_mask)
card->dev->coherent_dma_mask = DMA_BIT_MASK(32);
- if (pcm->streams[SNDRV_PCM_STREAM_PLAYBACK].substream) {
- ret = tegra_pcm_preallocate_dma_buffer(pcm,
- SNDRV_PCM_STREAM_PLAYBACK);
- if (ret)
- goto err;
- }
-
- if (pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream) {
- ret = tegra_pcm_preallocate_dma_buffer(pcm,
- SNDRV_PCM_STREAM_CAPTURE);
- if (ret)
- goto err_free_play;
- }
-
- return 0;
-
-err_free_play:
- tegra_pcm_deallocate_dma_buffer(pcm, SNDRV_PCM_STREAM_PLAYBACK);
-err:
+ ret = snd_pcm_lib_preallocate_pages_for_all(pcm,
+ SNDRV_DMA_TYPE_DEV_WC, card->dev,
+ TEGRA_PCM_PREALLOC_BUFFER,
+ TEGRA_PCM_PREALLOC_BUFFER_MAX);
+ if (ret < 0)
+ dev_err(card->dev, "dma buffer pre-alloc failed err %d\n", ret);
return ret;
}
static void tegra_pcm_free(struct snd_pcm *pcm)
{
- tegra_pcm_deallocate_dma_buffer(pcm, SNDRV_PCM_STREAM_CAPTURE);
- tegra_pcm_deallocate_dma_buffer(pcm, SNDRV_PCM_STREAM_PLAYBACK);
+ snd_pcm_lib_preallocate_free_for_all(pcm);
}
+
static struct snd_soc_platform_driver tegra_pcm_platform = {
.ops = &tegra_pcm_ops,
.pcm_new = tegra_pcm_new,
--
1.7.1.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2012-07-02 13:04 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-07-02 12:59 [PATCH V3 0/2] ASoC: Move pcm writecombine dma buffer allocation to core Laxman Dewangan
2012-07-02 12:59 ` Laxman Dewangan
2012-07-02 12:59 ` [PATCH V3 1/2] ALSA: Support writecombine DMA buffer memory page alloc/free Laxman Dewangan
2012-07-02 12:59 ` Laxman Dewangan
2012-07-02 12:59 ` [PATCH V3 2/2] ASoC: tegra: use sound pcm library for substream dma buffer Laxman Dewangan
2012-07-02 12:59 ` Laxman Dewangan
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.