| Message ID | 1345126397-28864-6-git-send-email-ldewangan@nvidia.com (mailing list archive) |
|---|---|
| State | Superseded, archived |
| Headers | show |
On 08/16/2012 08:13 AM, Laxman Dewangan wrote: > Remove the support code which uses the legacy APB DMA driver > for accessing the I2S FIFO. > The driver will use the dmaengine based APB DMA driver for > accessing reqding/writing to I2S FIFO. > > Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com> Mark, Liam, I'd like to take this patch through the Tegra tree, since it relies on the previous Tegra patches to convert to dmaengine. Could I please get an ack if that's OK? Thanks. ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
On 08/16/2012 08:13 AM, Laxman Dewangan wrote: > Remove the support code which uses the legacy APB DMA driver > for accessing the I2S FIFO. > The driver will use the dmaengine based APB DMA driver for > accessing reqding/writing to I2S FIFO. > > Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com> Please note that I've applied this to Tegra's for-3.7/dmaengine branch, even though it hasn't been ack'd by an ASoC maintainer. ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://ad.doubleclick.net/clk;258768047;13503038;j? http://info.appdynamics.com/FreeJavaPerformanceDownload.html
On Thu, Sep 13, 2012 at 05:36:30PM -0600, Stephen Warren wrote: > Please note that I've applied this to Tegra's for-3.7/dmaengine branch, > even though it hasn't been ack'd by an ASoC maintainer. Please don't apply this without review, we've got quite enough problems with the dmaengine stuff as it is unfortunately so I want to keep a handle on what's going on. I didn't look at it since when it was originally posted people found problems in testing so I was expecting a respin of the series. ------------------------------------------------------------------------------ Got visibility? Most devs has no idea what their production app looks like. Find out how fast your code is with AppDynamics Lite. http://ad.doubleclick.net/clk;262219671;13503038;y? http://info.appdynamics.com/FreeJavaPerformanceDownload.html
On 09/06/2012 05:50 PM, Stephen Warren wrote: > On 08/16/2012 08:13 AM, Laxman Dewangan wrote: >> Remove the support code which uses the legacy APB DMA driver >> for accessing the I2S FIFO. >> The driver will use the dmaengine based APB DMA driver for >> accessing reqding/writing to I2S FIFO. >> >> Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com> > > Mark, Liam, > > I'd like to take this patch through the Tegra tree, since it relies on > the previous Tegra patches to convert to dmaengine. Could I please get > an ack if that's OK? Thanks. I'm still looking for an ack. ------------------------------------------------------------------------------ Got visibility? Most devs has no idea what their production app looks like. Find out how fast your code is with AppDynamics Lite. http://ad.doubleclick.net/clk;262219671;13503038;y? http://info.appdynamics.com/FreeJavaPerformanceDownload.html
On 09/14/2012 09:53 AM, Mark Brown wrote: > On Thu, Sep 13, 2012 at 05:36:30PM -0600, Stephen Warren wrote: > >> Please note that I've applied this to Tegra's for-3.7/dmaengine branch, >> even though it hasn't been ack'd by an ASoC maintainer. > > Please don't apply this without review, we've got quite enough problems > with the dmaengine stuff as it is unfortunately so I want to keep a > handle on what's going on. I didn't look at it since when it was > originally posted people found problems in testing so I was expecting a > respin of the series. OK, I'll remove the series from next then. But do note that I pinged 8 days ago for an ack, and received no objections, and you were well aware that the bug was fixed since you applied the fix to the Tegra ASoC PCM driver yourself, and we had discussed this being a dependency for the dmaengine patches, and you'd told me to just base the Tegra branch on top of your tree until a signed tag was available for me to base upon. ------------------------------------------------------------------------------ Got visibility? Most devs has no idea what their production app looks like. Find out how fast your code is with AppDynamics Lite. http://ad.doubleclick.net/clk;262219671;13503038;y? http://info.appdynamics.com/FreeJavaPerformanceDownload.html
On Fri, Sep 14, 2012 at 09:57:35AM -0600, Stephen Warren wrote: > On 09/06/2012 05:50 PM, Stephen Warren wrote: > > I'd like to take this patch through the Tegra tree, since it relies on > > the previous Tegra patches to convert to dmaengine. Could I please get > > an ack if that's OK? Thanks. > I'm still looking for an ack. As with all the others here I need to see this so it can get reviewed, the previous review was that the patch didn't work. In general contentless pings like this aren't terribly helpful... ------------------------------------------------------------------------------ Got visibility? Most devs has no idea what their production app looks like. Find out how fast your code is with AppDynamics Lite. http://ad.doubleclick.net/clk;262219671;13503038;y? http://info.appdynamics.com/FreeJavaPerformanceDownload.html
On Fri, Sep 14, 2012 at 10:00:39AM -0600, Stephen Warren wrote: > But do note that I pinged 8 days ago for an ack, and received no > objections, and you were well aware that the bug was fixed since you I didn't read these mails enough to notice that they were asking for an ack, they were followups to already discarded patches (the were discarded due to the testing problems). One of the things I do to deal with mail overload is to drop followups to patches that I don't remember reviewing (as in this case where the negative test reports arrived before I saw the patches) or where I remember I was already happy. > applied the fix to the Tegra ASoC PCM driver yourself, and we had > discussed this being a dependency for the dmaengine patches, and you'd > told me to just base the Tegra branch on top of your tree until a signed > tag was available for me to base upon. I knew it was a dependency, I didn't know that there were no further changes required to the series - your note when you posted the patch didn't mention anything. ------------------------------------------------------------------------------ Got visibility? Most devs has no idea what their production app looks like. Find out how fast your code is with AppDynamics Lite. http://ad.doubleclick.net/clk;262219671;13503038;y? http://info.appdynamics.com/FreeJavaPerformanceDownload.html
diff --git a/sound/soc/tegra/Kconfig b/sound/soc/tegra/Kconfig index 02bcd30..19e5fe7 100644 --- a/sound/soc/tegra/Kconfig +++ b/sound/soc/tegra/Kconfig @@ -1,6 +1,6 @@ config SND_SOC_TEGRA tristate "SoC Audio for the Tegra System-on-Chip" - depends on ARCH_TEGRA && (TEGRA_SYSTEM_DMA || TEGRA20_APB_DMA) + depends on ARCH_TEGRA && TEGRA20_APB_DMA select REGMAP_MMIO select SND_SOC_DMAENGINE_PCM if TEGRA20_APB_DMA help diff --git a/sound/soc/tegra/tegra_pcm.c b/sound/soc/tegra/tegra_pcm.c index 5658bce..1cab2cd 100644 --- a/sound/soc/tegra/tegra_pcm.c +++ b/sound/soc/tegra/tegra_pcm.c @@ -57,237 +57,6 @@ static const struct snd_pcm_hardware tegra_pcm_hardware = { .fifo_size = 4, }; -#if defined(CONFIG_TEGRA_SYSTEM_DMA) -static void tegra_pcm_queue_dma(struct tegra_runtime_data *prtd) -{ - struct snd_pcm_substream *substream = prtd->substream; - struct snd_dma_buffer *buf = &substream->dma_buffer; - struct tegra_dma_req *dma_req; - unsigned long addr; - - dma_req = &prtd->dma_req[prtd->dma_req_idx]; - prtd->dma_req_idx = 1 - prtd->dma_req_idx; - - addr = buf->addr + prtd->dma_pos; - prtd->dma_pos += dma_req->size; - if (prtd->dma_pos >= prtd->dma_pos_end) - prtd->dma_pos = 0; - - if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) - dma_req->source_addr = addr; - else - dma_req->dest_addr = addr; - - tegra_dma_enqueue_req(prtd->dma_chan, dma_req); -} - -static void dma_complete_callback(struct tegra_dma_req *req) -{ - struct tegra_runtime_data *prtd = (struct tegra_runtime_data *)req->dev; - struct snd_pcm_substream *substream = prtd->substream; - struct snd_pcm_runtime *runtime = substream->runtime; - - spin_lock(&prtd->lock); - - if (!prtd->running) { - spin_unlock(&prtd->lock); - return; - } - - if (++prtd->period_index >= runtime->periods) - prtd->period_index = 0; - - tegra_pcm_queue_dma(prtd); - - spin_unlock(&prtd->lock); - - snd_pcm_period_elapsed(substream); -} - -static void setup_dma_tx_request(struct tegra_dma_req *req, - struct tegra_pcm_dma_params * dmap) -{ - req->complete = dma_complete_callback; - req->to_memory = false; - req->dest_addr = dmap->addr; - req->dest_wrap = dmap->wrap; - req->source_bus_width = 32; - req->source_wrap = 0; - req->dest_bus_width = dmap->width; - req->req_sel = dmap->req_sel; -} - -static void setup_dma_rx_request(struct tegra_dma_req *req, - struct tegra_pcm_dma_params * dmap) -{ - req->complete = dma_complete_callback; - req->to_memory = true; - req->source_addr = dmap->addr; - req->dest_wrap = 0; - req->source_bus_width = dmap->width; - req->source_wrap = dmap->wrap; - req->dest_bus_width = 32; - req->req_sel = dmap->req_sel; -} - -static int tegra_pcm_open(struct snd_pcm_substream *substream) -{ - struct snd_pcm_runtime *runtime = substream->runtime; - struct tegra_runtime_data *prtd; - struct snd_soc_pcm_runtime *rtd = substream->private_data; - struct tegra_pcm_dma_params * dmap; - int ret = 0; - - prtd = kzalloc(sizeof(struct tegra_runtime_data), GFP_KERNEL); - if (prtd == NULL) - return -ENOMEM; - - runtime->private_data = prtd; - prtd->substream = substream; - - spin_lock_init(&prtd->lock); - - if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) { - dmap = snd_soc_dai_get_dma_data(rtd->cpu_dai, substream); - setup_dma_tx_request(&prtd->dma_req[0], dmap); - setup_dma_tx_request(&prtd->dma_req[1], dmap); - } else { - dmap = snd_soc_dai_get_dma_data(rtd->cpu_dai, substream); - setup_dma_rx_request(&prtd->dma_req[0], dmap); - setup_dma_rx_request(&prtd->dma_req[1], dmap); - } - - prtd->dma_req[0].dev = prtd; - prtd->dma_req[1].dev = prtd; - - prtd->dma_chan = tegra_dma_allocate_channel(TEGRA_DMA_MODE_ONESHOT); - if (prtd->dma_chan == NULL) { - ret = -ENOMEM; - goto err; - } - - /* Set HW params now that initialization is complete */ - snd_soc_set_runtime_hwparams(substream, &tegra_pcm_hardware); - - /* Ensure that buffer size is a multiple of period size */ - ret = snd_pcm_hw_constraint_integer(runtime, - SNDRV_PCM_HW_PARAM_PERIODS); - if (ret < 0) - goto err; - - return 0; - -err: - if (prtd->dma_chan) { - tegra_dma_free_channel(prtd->dma_chan); - } - - kfree(prtd); - - return ret; -} - -static int tegra_pcm_close(struct snd_pcm_substream *substream) -{ - struct snd_pcm_runtime *runtime = substream->runtime; - struct tegra_runtime_data *prtd = runtime->private_data; - - tegra_dma_free_channel(prtd->dma_chan); - - kfree(prtd); - - return 0; -} - -static int tegra_pcm_hw_params(struct snd_pcm_substream *substream, - struct snd_pcm_hw_params *params) -{ - struct snd_pcm_runtime *runtime = substream->runtime; - struct tegra_runtime_data *prtd = runtime->private_data; - - snd_pcm_set_runtime_buffer(substream, &substream->dma_buffer); - - prtd->dma_req[0].size = params_period_bytes(params); - prtd->dma_req[1].size = prtd->dma_req[0].size; - - return 0; -} - -static int tegra_pcm_hw_free(struct snd_pcm_substream *substream) -{ - snd_pcm_set_runtime_buffer(substream, NULL); - - return 0; -} - -static int tegra_pcm_trigger(struct snd_pcm_substream *substream, int cmd) -{ - struct snd_pcm_runtime *runtime = substream->runtime; - struct tegra_runtime_data *prtd = runtime->private_data; - unsigned long flags; - - switch (cmd) { - case SNDRV_PCM_TRIGGER_START: - prtd->dma_pos = 0; - prtd->dma_pos_end = frames_to_bytes(runtime, runtime->periods * runtime->period_size); - prtd->period_index = 0; - prtd->dma_req_idx = 0; - /* Fall-through */ - case SNDRV_PCM_TRIGGER_RESUME: - case SNDRV_PCM_TRIGGER_PAUSE_RELEASE: - spin_lock_irqsave(&prtd->lock, flags); - prtd->running = 1; - spin_unlock_irqrestore(&prtd->lock, flags); - tegra_pcm_queue_dma(prtd); - tegra_pcm_queue_dma(prtd); - break; - case SNDRV_PCM_TRIGGER_STOP: - case SNDRV_PCM_TRIGGER_SUSPEND: - case SNDRV_PCM_TRIGGER_PAUSE_PUSH: - spin_lock_irqsave(&prtd->lock, flags); - prtd->running = 0; - spin_unlock_irqrestore(&prtd->lock, flags); - tegra_dma_dequeue_req(prtd->dma_chan, &prtd->dma_req[0]); - tegra_dma_dequeue_req(prtd->dma_chan, &prtd->dma_req[1]); - break; - default: - return -EINVAL; - } - - return 0; -} - -static snd_pcm_uframes_t tegra_pcm_pointer(struct snd_pcm_substream *substream) -{ - struct snd_pcm_runtime *runtime = substream->runtime; - struct tegra_runtime_data *prtd = runtime->private_data; - - 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, - .ioctl = snd_pcm_lib_ioctl, - .hw_params = tegra_pcm_hw_params, - .hw_free = tegra_pcm_hw_free, - .trigger = tegra_pcm_trigger, - .pointer = tegra_pcm_pointer, - .mmap = tegra_pcm_mmap, -}; -#else static int tegra_pcm_open(struct snd_pcm_substream *substream) { struct snd_soc_pcm_runtime *rtd = substream->private_data; @@ -399,7 +168,6 @@ static struct snd_pcm_ops tegra_pcm_ops = { .pointer = snd_dmaengine_pcm_pointer, .mmap = tegra_pcm_mmap, }; -#endif static int tegra_pcm_preallocate_dma_buffer(struct snd_pcm *pcm, int stream) { diff --git a/sound/soc/tegra/tegra_pcm.h b/sound/soc/tegra/tegra_pcm.h index a3a4503..b40279b 100644 --- a/sound/soc/tegra/tegra_pcm.h +++ b/sound/soc/tegra/tegra_pcm.h @@ -40,20 +40,6 @@ struct tegra_pcm_dma_params { unsigned long req_sel; }; -#if defined(CONFIG_TEGRA_SYSTEM_DMA) -struct tegra_runtime_data { - struct snd_pcm_substream *substream; - spinlock_t lock; - int running; - int dma_pos; - int dma_pos_end; - int period_index; - int dma_req_idx; - struct tegra_dma_req dma_req[2]; - struct tegra_dma_channel *dma_chan; -}; -#endif - int tegra_pcm_platform_register(struct device *dev); void tegra_pcm_platform_unregister(struct device *dev);
Remove the support code which uses the legacy APB DMA driver for accessing the I2S FIFO. The driver will use the dmaengine based APB DMA driver for accessing reqding/writing to I2S FIFO. Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com> --- sound/soc/tegra/Kconfig | 2 +- sound/soc/tegra/tegra_pcm.c | 232 ------------------------------------------- sound/soc/tegra/tegra_pcm.h | 14 --- 3 files changed, 1 insertions(+), 247 deletions(-)