| Message ID | 1400617569-7187-1-git-send-email-anarsoul@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, May 20, 2014 at 11:26:08PM +0300, Vasily Khoruzhick wrote:
> Use dmaengine instead of legacy s3c24xx DMA API for s3c24xx and s3c2412
Do these depend on any other code or can they just be applied
immediately?
On Wed, May 21, 2014 at 1:19 AM, Mark Brown <broonie@kernel.org> wrote: > On Tue, May 20, 2014 at 11:26:08PM +0300, Vasily Khoruzhick wrote: >> Use dmaengine instead of legacy s3c24xx DMA API for s3c24xx and s3c2412 > > Do these depend on any other code or can they just be applied > immediately? For compile time everything is present, for runtime they depend on cyclic transfers support for s3c24xx-dmac (already submitted). Also s3c24xx machines lack a platform device DMA (one-liner, not submitted yet). Regards Vasily
On Wed, May 21, 2014 at 09:42:21AM +0300, Vasily Khoruzhick wrote: > On Wed, May 21, 2014 at 1:19 AM, Mark Brown <broonie@kernel.org> wrote: > > On Tue, May 20, 2014 at 11:26:08PM +0300, Vasily Khoruzhick wrote: > >> Use dmaengine instead of legacy s3c24xx DMA API for s3c24xx and s3c2412 > > Do these depend on any other code or can they just be applied > > immediately? > For compile time everything is present, for runtime they depend on > cyclic transfers support for s3c24xx-dmac (already submitted). > Also s3c24xx machines lack a platform device DMA (one-liner, not submitted yet). OK, in that case I think it might be better to hold off until either the DMA patches are merged or after the merge window (should be very soon) so that the driver still works in v3.16 - does that seem sensible? The changes look OK, just a question of getting them merged safely.
On Wed, May 21, 2014 at 1:49 PM, Mark Brown <broonie@kernel.org> wrote: > On Wed, May 21, 2014 at 09:42:21AM +0300, Vasily Khoruzhick wrote: >> On Wed, May 21, 2014 at 1:19 AM, Mark Brown <broonie@kernel.org> wrote: >> > On Tue, May 20, 2014 at 11:26:08PM +0300, Vasily Khoruzhick wrote: >> >> Use dmaengine instead of legacy s3c24xx DMA API for s3c24xx and s3c2412 > >> > Do these depend on any other code or can they just be applied >> > immediately? > >> For compile time everything is present, for runtime they depend on >> cyclic transfers support for s3c24xx-dmac (already submitted). >> Also s3c24xx machines lack a platform device DMA (one-liner, not submitted yet). > > OK, in that case I think it might be better to hold off until either the > DMA patches are merged or after the merge window (should be very soon) > so that the driver still works in v3.16 - does that seem sensible? The > changes look OK, just a question of getting them merged safely. Mark, legacy dma does not work for me as for 3.14, tried on h1940 (s3c2410) and on rx1950 (s3c2442). So merging them won't break anything, it's already broken. Regards Vasily
Hi Vasily, Please remember to keep linux-samsung-soc ML on Cc when sending patches related to Samsung platforms. It is essential to let Samsung people review them. Please resend the series with this in mind. The patch in general looks good to me. Might be a good idea to specify on what board(s) and how it was tested, though. Best regards, Tomasz On 20.05.2014 22:26, Vasily Khoruzhick wrote: > Use dmaengine instead of legacy s3c24xx DMA API for s3c24xx and s3c2412 > > Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com> > --- > v2: use hardcoded dma channel number > > sound/soc/samsung/Kconfig | 9 +++---- > sound/soc/samsung/dmaengine.c | 3 +++ > sound/soc/samsung/s3c-i2s-v2.c | 17 +------------ > sound/soc/samsung/s3c2412-i2s.c | 47 +++++++++++++++++----------------- > sound/soc/samsung/s3c24xx-i2s.c | 56 +++++++++++++++++++---------------------- > 5 files changed, 56 insertions(+), 76 deletions(-) > > diff --git a/sound/soc/samsung/Kconfig b/sound/soc/samsung/Kconfig > index 3507574..022c7a4 100644 > --- a/sound/soc/samsung/Kconfig > +++ b/sound/soc/samsung/Kconfig > @@ -1,11 +1,10 @@ > config SND_SOC_SAMSUNG > tristate "ASoC support for Samsung" > depends on PLAT_SAMSUNG > - select S3C2410_DMA if ARCH_S3C24XX > + select S3C24XX_DMAC if ARCH_S3C24XX > select S3C64XX_PL080 if ARCH_S3C64XX > - select SND_S3C_DMA if !ARCH_S3C24XX > - select SND_S3C_DMA_LEGACY if ARCH_S3C24XX > - select SND_SOC_GENERIC_DMAENGINE_PCM if !ARCH_S3C24XX > + select SND_S3C_DMA > + select SND_SOC_GENERIC_DMAENGINE_PCM > help > Say Y or M if you want to add support for codecs attached to > the Samsung SoCs' Audio interfaces. You will also need to > @@ -19,7 +18,6 @@ config SND_S3C_DMA_LEGACY > > config SND_S3C24XX_I2S > tristate > - select S3C24XX_DMA > > config SND_S3C_I2SV2_SOC > tristate > @@ -27,7 +25,6 @@ config SND_S3C_I2SV2_SOC > config SND_S3C2412_SOC_I2S > tristate > select SND_S3C_I2SV2_SOC > - select S3C2410_DMA > > config SND_SAMSUNG_PCM > tristate > diff --git a/sound/soc/samsung/dmaengine.c b/sound/soc/samsung/dmaengine.c > index 750ce58..72f27d1 100644 > --- a/sound/soc/samsung/dmaengine.c > +++ b/sound/soc/samsung/dmaengine.c > @@ -17,6 +17,7 @@ > > #include <linux/module.h> > #include <linux/amba/pl08x.h> > +#include <linux/platform_data/dma-s3c24xx.h> > > #include <sound/core.h> > #include <sound/pcm.h> > @@ -29,6 +30,8 @@ > > #ifdef CONFIG_ARCH_S3C64XX > #define filter_fn pl08x_filter_id > +#elif defined(CONFIG_ARCH_S3C24XX) > +#define filter_fn s3c24xx_dma_filter > #else > #define filter_fn NULL > #endif > diff --git a/sound/soc/samsung/s3c-i2s-v2.c b/sound/soc/samsung/s3c-i2s-v2.c > index 79e7efb..1a74051 100644 > --- a/sound/soc/samsung/s3c-i2s-v2.c > +++ b/sound/soc/samsung/s3c-i2s-v2.c > @@ -392,8 +392,6 @@ static int s3c2412_i2s_trigger(struct snd_pcm_substream *substream, int cmd, > int capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); > unsigned long irqs; > int ret = 0; > - struct s3c_dma_params *dma_data = > - snd_soc_dai_get_dma_data(rtd->cpu_dai, substream); > > pr_debug("Entered %s\n", __func__); > > @@ -424,13 +422,6 @@ static int s3c2412_i2s_trigger(struct snd_pcm_substream *substream, int cmd, > > local_irq_restore(irqs); > > - /* > - * Load the next buffer to DMA to meet the reqirement > - * of the auto reload mechanism of S3C24XX. > - * This call won't bother S3C64XX. > - */ > - s3c2410_dma_ctrl(dma_data->channel, S3C2410_DMAOP_STARTED); > - > break; > > case SNDRV_PCM_TRIGGER_STOP: > @@ -644,12 +635,6 @@ int s3c_i2sv2_probe(struct snd_soc_dai *dai, > /* record our i2s structure for later use in the callbacks */ > snd_soc_dai_set_drvdata(dai, i2s); > > - i2s->regs = ioremap(base, 0x100); > - if (i2s->regs == NULL) { > - dev_err(dev, "cannot ioremap registers\n"); > - return -ENXIO; > - } > - > i2s->iis_pclk = clk_get(dev, "iis"); > if (IS_ERR(i2s->iis_pclk)) { > dev_err(dev, "failed to get iis_clock\n"); > @@ -729,7 +714,7 @@ int s3c_i2sv2_register_component(struct device *dev, int id, > struct snd_soc_component_driver *cmp_drv, > struct snd_soc_dai_driver *dai_drv) > { > - struct snd_soc_dai_ops *ops = dai_drv->ops; > + struct snd_soc_dai_ops *ops = (struct snd_soc_dai_ops *)dai_drv->ops; > > ops->trigger = s3c2412_i2s_trigger; > if (!ops->hw_params) > diff --git a/sound/soc/samsung/s3c2412-i2s.c b/sound/soc/samsung/s3c2412-i2s.c > index d079445..ba084e3 100644 > --- a/sound/soc/samsung/s3c2412-i2s.c > +++ b/sound/soc/samsung/s3c2412-i2s.c > @@ -33,25 +33,19 @@ > #include "regs-i2s-v2.h" > #include "s3c2412-i2s.h" > > -static struct s3c2410_dma_client s3c2412_dma_client_out = { > - .name = "I2S PCM Stereo out" > -}; > - > -static struct s3c2410_dma_client s3c2412_dma_client_in = { > - .name = "I2S PCM Stereo in" > -}; > - > static struct s3c_dma_params s3c2412_i2s_pcm_stereo_out = { > - .client = &s3c2412_dma_client_out, > + .client = > + (struct s3c2410_dma_client *)&s3c2412_i2s_pcm_stereo_out, > .channel = DMACH_I2S_OUT, > - .dma_addr = S3C2410_PA_IIS + S3C2412_IISTXD, > + .ch_name = "tx", > .dma_size = 4, > }; > > static struct s3c_dma_params s3c2412_i2s_pcm_stereo_in = { > - .client = &s3c2412_dma_client_in, > + .client = > + (struct s3c2410_dma_client *)&s3c2412_i2s_pcm_stereo_in, > .channel = DMACH_I2S_IN, > - .dma_addr = S3C2410_PA_IIS + S3C2412_IISRXD, > + .ch_name = "rx", > .dma_size = 4, > }; > > @@ -63,6 +57,9 @@ static int s3c2412_i2s_probe(struct snd_soc_dai *dai) > > pr_debug("Entered %s\n", __func__); > > + samsung_asoc_init_dma_data(dai, &s3c2412_i2s_pcm_stereo_out, > + &s3c2412_i2s_pcm_stereo_in); > + > ret = s3c_i2sv2_probe(dai, &s3c2412_i2s, S3C2410_PA_IIS); > if (ret) > return ret; > @@ -70,10 +67,9 @@ static int s3c2412_i2s_probe(struct snd_soc_dai *dai) > s3c2412_i2s.dma_capture = &s3c2412_i2s_pcm_stereo_in; > s3c2412_i2s.dma_playback = &s3c2412_i2s_pcm_stereo_out; > > - s3c2412_i2s.iis_cclk = clk_get(dai->dev, "i2sclk"); > + s3c2412_i2s.iis_cclk = devm_clk_get(dai->dev, "i2sclk"); > if (IS_ERR(s3c2412_i2s.iis_cclk)) { > pr_err("failed to get i2sclk clock\n"); > - iounmap(s3c2412_i2s.regs); > return PTR_ERR(s3c2412_i2s.iis_cclk); > } > > @@ -94,8 +90,6 @@ static int s3c2412_i2s_probe(struct snd_soc_dai *dai) > static int s3c2412_i2s_remove(struct snd_soc_dai *dai) > { > clk_disable(s3c2412_i2s.iis_cclk); > - clk_put(s3c2412_i2s.iis_cclk); > - iounmap(s3c2412_i2s.regs); > > return 0; > } > @@ -105,18 +99,10 @@ static int s3c2412_i2s_hw_params(struct snd_pcm_substream *substream, > struct snd_soc_dai *cpu_dai) > { > struct s3c_i2sv2_info *i2s = snd_soc_dai_get_drvdata(cpu_dai); > - struct s3c_dma_params *dma_data; > u32 iismod; > > pr_debug("Entered %s\n", __func__); > > - if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) > - dma_data = i2s->dma_playback; > - else > - dma_data = i2s->dma_capture; > - > - snd_soc_dai_set_dma_data(cpu_dai, substream, dma_data); > - > iismod = readl(i2s->regs + S3C2412_IISMOD); > pr_debug("%s: r: IISMOD: %x\n", __func__, iismod); > > @@ -169,6 +155,19 @@ static const struct snd_soc_component_driver s3c2412_i2s_component = { > static int s3c2412_iis_dev_probe(struct platform_device *pdev) > { > int ret = 0; > + struct resource *res; > + > + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > + if (!res) { > + dev_err(&pdev->dev, "Can't get IO resource.\n"); > + return -ENOENT; > + } > + s3c2412_i2s.regs = devm_ioremap_resource(&pdev->dev, res); > + if (s3c2412_i2s.regs == NULL) > + return -ENXIO; > + > + s3c2412_i2s_pcm_stereo_out.dma_addr = res->start + S3C2412_IISTXD; > + s3c2412_i2s_pcm_stereo_in.dma_addr = res->start + S3C2412_IISRXD; > > ret = s3c_i2sv2_register_component(&pdev->dev, -1, > &s3c2412_i2s_component, > diff --git a/sound/soc/samsung/s3c24xx-i2s.c b/sound/soc/samsung/s3c24xx-i2s.c > index f31e916..97600df 100644 > --- a/sound/soc/samsung/s3c24xx-i2s.c > +++ b/sound/soc/samsung/s3c24xx-i2s.c > @@ -31,25 +31,19 @@ > #include "dma.h" > #include "s3c24xx-i2s.h" > > -static struct s3c2410_dma_client s3c24xx_dma_client_out = { > - .name = "I2S PCM Stereo out" > -}; > - > -static struct s3c2410_dma_client s3c24xx_dma_client_in = { > - .name = "I2S PCM Stereo in" > -}; > - > static struct s3c_dma_params s3c24xx_i2s_pcm_stereo_out = { > - .client = &s3c24xx_dma_client_out, > + .client = > + (struct s3c2410_dma_client *)&s3c24xx_i2s_pcm_stereo_out, > .channel = DMACH_I2S_OUT, > - .dma_addr = S3C2410_PA_IIS + S3C2410_IISFIFO, > + .ch_name = "tx", > .dma_size = 2, > }; > > static struct s3c_dma_params s3c24xx_i2s_pcm_stereo_in = { > - .client = &s3c24xx_dma_client_in, > + .client = > + (struct s3c2410_dma_client *)&s3c24xx_i2s_pcm_stereo_in, > .channel = DMACH_I2S_IN, > - .dma_addr = S3C2410_PA_IIS + S3C2410_IISFIFO, > + .ch_name = "rx", > .dma_size = 2, > }; > > @@ -231,18 +225,12 @@ static int s3c24xx_i2s_hw_params(struct snd_pcm_substream *substream, > struct snd_pcm_hw_params *params, > struct snd_soc_dai *dai) > { > - struct snd_soc_pcm_runtime *rtd = substream->private_data; > - struct s3c_dma_params *dma_data; > + struct snd_dmaengine_dai_dma_data *dma_data; > u32 iismod; > > pr_debug("Entered %s\n", __func__); > > - if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) > - dma_data = &s3c24xx_i2s_pcm_stereo_out; > - else > - dma_data = &s3c24xx_i2s_pcm_stereo_in; > - > - snd_soc_dai_set_dma_data(rtd->cpu_dai, substream, dma_data); > + dma_data = snd_soc_dai_get_dma_data(dai, substream); > > /* Working copies of register */ > iismod = readl(s3c24xx_i2s.regs + S3C2410_IISMOD); > @@ -251,11 +239,11 @@ static int s3c24xx_i2s_hw_params(struct snd_pcm_substream *substream, > switch (params_format(params)) { > case SNDRV_PCM_FORMAT_S8: > iismod &= ~S3C2410_IISMOD_16BIT; > - dma_data->dma_size = 1; > + dma_data->addr_width = 1; > break; > case SNDRV_PCM_FORMAT_S16_LE: > iismod |= S3C2410_IISMOD_16BIT; > - dma_data->dma_size = 2; > + dma_data->addr_width = 2; > break; > default: > return -EINVAL; > @@ -270,8 +258,6 @@ static int s3c24xx_i2s_trigger(struct snd_pcm_substream *substream, int cmd, > struct snd_soc_dai *dai) > { > int ret = 0; > - struct s3c_dma_params *dma_data = > - snd_soc_dai_get_dma_data(dai, substream); > > pr_debug("Entered %s\n", __func__); > > @@ -290,7 +276,6 @@ static int s3c24xx_i2s_trigger(struct snd_pcm_substream *substream, int cmd, > else > s3c24xx_snd_txctrl(1); > > - s3c2410_dma_ctrl(dma_data->channel, S3C2410_DMAOP_STARTED); > break; > case SNDRV_PCM_TRIGGER_STOP: > case SNDRV_PCM_TRIGGER_SUSPEND: > @@ -380,14 +365,12 @@ static int s3c24xx_i2s_probe(struct snd_soc_dai *dai) > { > pr_debug("Entered %s\n", __func__); > > - s3c24xx_i2s.regs = ioremap(S3C2410_PA_IIS, 0x100); > - if (s3c24xx_i2s.regs == NULL) > - return -ENXIO; > + samsung_asoc_init_dma_data(dai, &s3c24xx_i2s_pcm_stereo_out, > + &s3c24xx_i2s_pcm_stereo_in); > > - s3c24xx_i2s.iis_clk = clk_get(dai->dev, "iis"); > + s3c24xx_i2s.iis_clk = devm_clk_get(dai->dev, "iis"); > if (IS_ERR(s3c24xx_i2s.iis_clk)) { > pr_err("failed to get iis_clock\n"); > - iounmap(s3c24xx_i2s.regs); > return PTR_ERR(s3c24xx_i2s.iis_clk); > } > clk_enable(s3c24xx_i2s.iis_clk); > @@ -474,6 +457,19 @@ static const struct snd_soc_component_driver s3c24xx_i2s_component = { > static int s3c24xx_iis_dev_probe(struct platform_device *pdev) > { > int ret = 0; > + struct resource *res; > + > + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > + if (!res) { > + dev_err(&pdev->dev, "Can't get IO resource.\n"); > + return -ENOENT; > + } > + s3c24xx_i2s.regs = devm_ioremap_resource(&pdev->dev, res); > + if (s3c24xx_i2s.regs == NULL) > + return -ENXIO; > + > + s3c24xx_i2s_pcm_stereo_out.dma_addr = res->start + S3C2410_IISFIFO; > + s3c24xx_i2s_pcm_stereo_in.dma_addr = res->start + S3C2410_IISFIFO; > > ret = snd_soc_register_component(&pdev->dev, &s3c24xx_i2s_component, > &s3c24xx_i2s_dai, 1); >
On Wed, May 21, 2014 at 4:28 PM, Tomasz Figa <t.figa@samsung.com> wrote: > Hi Vasily, Hi Tomasz, > Please remember to keep linux-samsung-soc ML on Cc when sending patches > related to Samsung platforms. It is essential to let Samsung people > review them. Ok, I'll keep it in mind. I've just checked and get_maintainer.pl outputs linux-samsung-soc ML for these patches. I've omitted it occasionally, because it wasn't necessary to keep it in CC back in 2011 (year of my last sound-related patch for s3c24xx). > Please resend the series with this in mind. Is it really necessary? This patch is supposed to go through ASoC tree, not through Samsung's. And I really doubt that Samsung guys test kernel for s3c24xx boards nowadays. > The patch in general looks good to me. Might be a good idea to specify > on what board(s) and how it was tested, though. It was tested on s3c2410-based iPAQ h1940, s3c2412 part was only compile-tested, as I don't have s3c2412 hardware. > Best regards, > Tomasz Regards Vasily
On 21.05.2014 16:53, Vasily Khoruzhick wrote: > On Wed, May 21, 2014 at 4:28 PM, Tomasz Figa <t.figa@samsung.com> wrote: >> Hi Vasily, > > Hi Tomasz, > >> Please remember to keep linux-samsung-soc ML on Cc when sending patches >> related to Samsung platforms. It is essential to let Samsung people >> review them. > > Ok, I'll keep it in mind. I've just checked and get_maintainer.pl > outputs linux-samsung-soc ML for these > patches. I've omitted it occasionally, because it wasn't necessary to > keep it in CC back in 2011 > (year of my last sound-related patch for s3c24xx). > >> Please resend the series with this in mind. > > Is it really necessary? This patch is supposed to go through ASoC > tree, not through Samsung's. > And I really doubt that Samsung guys test kernel for s3c24xx boards nowadays. Maybe not test, but at least review the code. Also there might be other people following that list that might be interested in looking at this series, while not subscribed to alsa-devel - I can name at least Heiko Stübner, who should be able to test it on further s3c24xx-based boards. > >> The patch in general looks good to me. Might be a good idea to specify >> on what board(s) and how it was tested, though. > > It was tested on s3c2410-based iPAQ h1940, s3c2412 part was only > compile-tested, as I don't have s3c2412 hardware. OK, thanks. Best regards, Tomasz
On Wed, May 21, 2014 at 6:07 PM, Tomasz Figa <t.figa@samsung.com> wrote: >> >> Is it really necessary? This patch is supposed to go through ASoC >> tree, not through Samsung's. >> And I really doubt that Samsung guys test kernel for s3c24xx boards nowadays. > > Maybe not test, but at least review the code. Also there might be other > people following that list that might be interested in looking at this > series, while not subscribed to alsa-devel - I can name at least Heiko > Stübner, who should be able to test it on further s3c24xx-based boards. Ok, will resend tonight. Regards Vasily
Am Mittwoch, 21. Mai 2014, 17:07:47 schrieb Tomasz Figa: > On 21.05.2014 16:53, Vasily Khoruzhick wrote: > > On Wed, May 21, 2014 at 4:28 PM, Tomasz Figa <t.figa@samsung.com> wrote: > >> Please remember to keep linux-samsung-soc ML on Cc when sending patches > >> related to Samsung platforms. It is essential to let Samsung people > >> review them. > > > > Ok, I'll keep it in mind. I've just checked and get_maintainer.pl > > outputs linux-samsung-soc ML for these > > patches. I've omitted it occasionally, because it wasn't necessary to > > keep it in CC back in 2011 > > (year of my last sound-related patch for s3c24xx). > > > >> Please resend the series with this in mind. > > > > Is it really necessary? This patch is supposed to go through ASoC > > tree, not through Samsung's. > > And I really doubt that Samsung guys test kernel for s3c24xx boards > > nowadays. > Maybe not test, but at least review the code. Also there might be other > people following that list that might be interested in looking at this > series, while not subscribed to alsa-devel - I can name at least Heiko > Stübner, who should be able to test it on further s3c24xx-based boards. Sadly, I don't have a working sound setup on my 2 s3c boards. Heiko
On Wed, May 21, 2014 at 6:29 PM, Heiko Stübner <heiko@sntech.de> wrote: > Am Mittwoch, 21. Mai 2014, 17:07:47 schrieb Tomasz Figa: >> On 21.05.2014 16:53, Vasily Khoruzhick wrote: >> > On Wed, May 21, 2014 at 4:28 PM, Tomasz Figa <t.figa@samsung.com> wrote: >> >> Please remember to keep linux-samsung-soc ML on Cc when sending patches >> >> related to Samsung platforms. It is essential to let Samsung people >> >> review them. >> > >> > Ok, I'll keep it in mind. I've just checked and get_maintainer.pl >> > outputs linux-samsung-soc ML for these >> > patches. I've omitted it occasionally, because it wasn't necessary to >> > keep it in CC back in 2011 >> > (year of my last sound-related patch for s3c24xx). >> > >> >> Please resend the series with this in mind. >> > >> > Is it really necessary? This patch is supposed to go through ASoC >> > tree, not through Samsung's. >> > And I really doubt that Samsung guys test kernel for s3c24xx boards >> > nowadays. >> Maybe not test, but at least review the code. Also there might be other >> people following that list that might be interested in looking at this >> series, while not subscribed to alsa-devel - I can name at least Heiko >> Stübner, who should be able to test it on further s3c24xx-based boards. > > Sadly, I don't have a working sound setup on my 2 s3c boards. What are they btw? Is it too time-consuming to make sound working on them? It would be nice if anyone except me could test these patches. > Heiko Regards Vasily
Am Mittwoch, 21. Mai 2014, 18:39:26 schrieb Vasily Khoruzhick: > On Wed, May 21, 2014 at 6:29 PM, Heiko Stübner <heiko@sntech.de> wrote: > > Am Mittwoch, 21. Mai 2014, 17:07:47 schrieb Tomasz Figa: > >> On 21.05.2014 16:53, Vasily Khoruzhick wrote: > >> > On Wed, May 21, 2014 at 4:28 PM, Tomasz Figa <t.figa@samsung.com> wrote: > >> >> Please remember to keep linux-samsung-soc ML on Cc when sending > >> >> patches > >> >> related to Samsung platforms. It is essential to let Samsung people > >> >> review them. > >> > > >> > Ok, I'll keep it in mind. I've just checked and get_maintainer.pl > >> > outputs linux-samsung-soc ML for these > >> > patches. I've omitted it occasionally, because it wasn't necessary to > >> > keep it in CC back in 2011 > >> > (year of my last sound-related patch for s3c24xx). > >> > > >> >> Please resend the series with this in mind. > >> > > >> > Is it really necessary? This patch is supposed to go through ASoC > >> > tree, not through Samsung's. > >> > And I really doubt that Samsung guys test kernel for s3c24xx boards > >> > nowadays. > >> > >> Maybe not test, but at least review the code. Also there might be other > >> people following that list that might be interested in looking at this > >> series, while not subscribed to alsa-devel - I can name at least Heiko > >> Stübner, who should be able to test it on further s3c24xx-based boards. > > > > Sadly, I don't have a working sound setup on my 2 s3c boards. > > What are they btw? Is it too time-consuming to make sound working on them? > It would be nice if anyone except me could test these patches. Openmoko Freerunner (S3C2440) where there may be sound reachable someway and a "Oyo" ebook-reader (S3C2416) whose i2s is I think compatible with the s3c64xx instead of the s3c24xx ones and with a completely unsupported codec. But in general I'm fairly confident that if you hear sounds on your device your conversion would be fine test-wise :-) Heiko
On Wed, May 21, 2014 at 6:56 PM, Heiko Stübner <heiko@sntech.de> wrote: > > Openmoko Freerunner (S3C2440) where there may be sound reachable someway and a > "Oyo" ebook-reader (S3C2416) whose i2s is I think compatible with the s3c64xx > instead of the s3c24xx ones and with a completely unsupported codec. Freerunner is s3c2442, and actually I can test s3c2442 on my iPAQ rx1950. s3c2416 may be supported by s3c2412-i2s, I have no spec at the hand, so I can't check. Regards Vasily > But in general I'm fairly confident that if you hear sounds on your device your > conversion would be fine test-wise :-) > > > Heiko
On Wed, May 21, 2014 at 07:09:19PM +0300, Vasily Khoruzhick wrote: > On Wed, May 21, 2014 at 6:56 PM, Heiko Stübner <heiko@sntech.de> wrote: > > Openmoko Freerunner (S3C2440) where there may be sound reachable someway and a > > "Oyo" ebook-reader (S3C2416) whose i2s is I think compatible with the s3c64xx > > instead of the s3c24xx ones and with a completely unsupported codec. > Freerunner is s3c2442, and actually I can test s3c2442 on my iPAQ rx1950. > s3c2416 may be supported by s3c2412-i2s, I have no spec at the hand, > so I can't check. The OpenMoko should work in mainline with an appropraite audio route set up - there's an earpiece and (2.5mm sadly) headphone jack.
Am Mittwoch, 21. Mai 2014, 19:09:19 schrieb Vasily Khoruzhick: > On Wed, May 21, 2014 at 6:56 PM, Heiko Stübner <heiko@sntech.de> wrote: > > Openmoko Freerunner (S3C2440) where there may be sound reachable someway > > and a "Oyo" ebook-reader (S3C2416) whose i2s is I think compatible with > > the s3c64xx instead of the s3c24xx ones and with a completely unsupported > > codec. > Freerunner is s3c2442, and actually I can test s3c2442 on my iPAQ rx1950. > s3c2416 may be supported by s3c2412-i2s, I have no spec at the hand, > so I can't check. I'm quite sure, s3c2416 shares the s3c64xx i2s ... at some point I matched the registers and it also shares the hsspi, sdhci and fb drivers with the newer generations (s3c64xx onward). Heiko
diff --git a/sound/soc/samsung/Kconfig b/sound/soc/samsung/Kconfig index 3507574..022c7a4 100644 --- a/sound/soc/samsung/Kconfig +++ b/sound/soc/samsung/Kconfig @@ -1,11 +1,10 @@ config SND_SOC_SAMSUNG tristate "ASoC support for Samsung" depends on PLAT_SAMSUNG - select S3C2410_DMA if ARCH_S3C24XX + select S3C24XX_DMAC if ARCH_S3C24XX select S3C64XX_PL080 if ARCH_S3C64XX - select SND_S3C_DMA if !ARCH_S3C24XX - select SND_S3C_DMA_LEGACY if ARCH_S3C24XX - select SND_SOC_GENERIC_DMAENGINE_PCM if !ARCH_S3C24XX + select SND_S3C_DMA + select SND_SOC_GENERIC_DMAENGINE_PCM help Say Y or M if you want to add support for codecs attached to the Samsung SoCs' Audio interfaces. You will also need to @@ -19,7 +18,6 @@ config SND_S3C_DMA_LEGACY config SND_S3C24XX_I2S tristate - select S3C24XX_DMA config SND_S3C_I2SV2_SOC tristate @@ -27,7 +25,6 @@ config SND_S3C_I2SV2_SOC config SND_S3C2412_SOC_I2S tristate select SND_S3C_I2SV2_SOC - select S3C2410_DMA config SND_SAMSUNG_PCM tristate diff --git a/sound/soc/samsung/dmaengine.c b/sound/soc/samsung/dmaengine.c index 750ce58..72f27d1 100644 --- a/sound/soc/samsung/dmaengine.c +++ b/sound/soc/samsung/dmaengine.c @@ -17,6 +17,7 @@ #include <linux/module.h> #include <linux/amba/pl08x.h> +#include <linux/platform_data/dma-s3c24xx.h> #include <sound/core.h> #include <sound/pcm.h> @@ -29,6 +30,8 @@ #ifdef CONFIG_ARCH_S3C64XX #define filter_fn pl08x_filter_id +#elif defined(CONFIG_ARCH_S3C24XX) +#define filter_fn s3c24xx_dma_filter #else #define filter_fn NULL #endif diff --git a/sound/soc/samsung/s3c-i2s-v2.c b/sound/soc/samsung/s3c-i2s-v2.c index 79e7efb..1a74051 100644 --- a/sound/soc/samsung/s3c-i2s-v2.c +++ b/sound/soc/samsung/s3c-i2s-v2.c @@ -392,8 +392,6 @@ static int s3c2412_i2s_trigger(struct snd_pcm_substream *substream, int cmd, int capture = (substream->stream == SNDRV_PCM_STREAM_CAPTURE); unsigned long irqs; int ret = 0; - struct s3c_dma_params *dma_data = - snd_soc_dai_get_dma_data(rtd->cpu_dai, substream); pr_debug("Entered %s\n", __func__); @@ -424,13 +422,6 @@ static int s3c2412_i2s_trigger(struct snd_pcm_substream *substream, int cmd, local_irq_restore(irqs); - /* - * Load the next buffer to DMA to meet the reqirement - * of the auto reload mechanism of S3C24XX. - * This call won't bother S3C64XX. - */ - s3c2410_dma_ctrl(dma_data->channel, S3C2410_DMAOP_STARTED); - break; case SNDRV_PCM_TRIGGER_STOP: @@ -644,12 +635,6 @@ int s3c_i2sv2_probe(struct snd_soc_dai *dai, /* record our i2s structure for later use in the callbacks */ snd_soc_dai_set_drvdata(dai, i2s); - i2s->regs = ioremap(base, 0x100); - if (i2s->regs == NULL) { - dev_err(dev, "cannot ioremap registers\n"); - return -ENXIO; - } - i2s->iis_pclk = clk_get(dev, "iis"); if (IS_ERR(i2s->iis_pclk)) { dev_err(dev, "failed to get iis_clock\n"); @@ -729,7 +714,7 @@ int s3c_i2sv2_register_component(struct device *dev, int id, struct snd_soc_component_driver *cmp_drv, struct snd_soc_dai_driver *dai_drv) { - struct snd_soc_dai_ops *ops = dai_drv->ops; + struct snd_soc_dai_ops *ops = (struct snd_soc_dai_ops *)dai_drv->ops; ops->trigger = s3c2412_i2s_trigger; if (!ops->hw_params) diff --git a/sound/soc/samsung/s3c2412-i2s.c b/sound/soc/samsung/s3c2412-i2s.c index d079445..ba084e3 100644 --- a/sound/soc/samsung/s3c2412-i2s.c +++ b/sound/soc/samsung/s3c2412-i2s.c @@ -33,25 +33,19 @@ #include "regs-i2s-v2.h" #include "s3c2412-i2s.h" -static struct s3c2410_dma_client s3c2412_dma_client_out = { - .name = "I2S PCM Stereo out" -}; - -static struct s3c2410_dma_client s3c2412_dma_client_in = { - .name = "I2S PCM Stereo in" -}; - static struct s3c_dma_params s3c2412_i2s_pcm_stereo_out = { - .client = &s3c2412_dma_client_out, + .client = + (struct s3c2410_dma_client *)&s3c2412_i2s_pcm_stereo_out, .channel = DMACH_I2S_OUT, - .dma_addr = S3C2410_PA_IIS + S3C2412_IISTXD, + .ch_name = "tx", .dma_size = 4, }; static struct s3c_dma_params s3c2412_i2s_pcm_stereo_in = { - .client = &s3c2412_dma_client_in, + .client = + (struct s3c2410_dma_client *)&s3c2412_i2s_pcm_stereo_in, .channel = DMACH_I2S_IN, - .dma_addr = S3C2410_PA_IIS + S3C2412_IISRXD, + .ch_name = "rx", .dma_size = 4, }; @@ -63,6 +57,9 @@ static int s3c2412_i2s_probe(struct snd_soc_dai *dai) pr_debug("Entered %s\n", __func__); + samsung_asoc_init_dma_data(dai, &s3c2412_i2s_pcm_stereo_out, + &s3c2412_i2s_pcm_stereo_in); + ret = s3c_i2sv2_probe(dai, &s3c2412_i2s, S3C2410_PA_IIS); if (ret) return ret; @@ -70,10 +67,9 @@ static int s3c2412_i2s_probe(struct snd_soc_dai *dai) s3c2412_i2s.dma_capture = &s3c2412_i2s_pcm_stereo_in; s3c2412_i2s.dma_playback = &s3c2412_i2s_pcm_stereo_out; - s3c2412_i2s.iis_cclk = clk_get(dai->dev, "i2sclk"); + s3c2412_i2s.iis_cclk = devm_clk_get(dai->dev, "i2sclk"); if (IS_ERR(s3c2412_i2s.iis_cclk)) { pr_err("failed to get i2sclk clock\n"); - iounmap(s3c2412_i2s.regs); return PTR_ERR(s3c2412_i2s.iis_cclk); } @@ -94,8 +90,6 @@ static int s3c2412_i2s_probe(struct snd_soc_dai *dai) static int s3c2412_i2s_remove(struct snd_soc_dai *dai) { clk_disable(s3c2412_i2s.iis_cclk); - clk_put(s3c2412_i2s.iis_cclk); - iounmap(s3c2412_i2s.regs); return 0; } @@ -105,18 +99,10 @@ static int s3c2412_i2s_hw_params(struct snd_pcm_substream *substream, struct snd_soc_dai *cpu_dai) { struct s3c_i2sv2_info *i2s = snd_soc_dai_get_drvdata(cpu_dai); - struct s3c_dma_params *dma_data; u32 iismod; pr_debug("Entered %s\n", __func__); - if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) - dma_data = i2s->dma_playback; - else - dma_data = i2s->dma_capture; - - snd_soc_dai_set_dma_data(cpu_dai, substream, dma_data); - iismod = readl(i2s->regs + S3C2412_IISMOD); pr_debug("%s: r: IISMOD: %x\n", __func__, iismod); @@ -169,6 +155,19 @@ static const struct snd_soc_component_driver s3c2412_i2s_component = { static int s3c2412_iis_dev_probe(struct platform_device *pdev) { int ret = 0; + struct resource *res; + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!res) { + dev_err(&pdev->dev, "Can't get IO resource.\n"); + return -ENOENT; + } + s3c2412_i2s.regs = devm_ioremap_resource(&pdev->dev, res); + if (s3c2412_i2s.regs == NULL) + return -ENXIO; + + s3c2412_i2s_pcm_stereo_out.dma_addr = res->start + S3C2412_IISTXD; + s3c2412_i2s_pcm_stereo_in.dma_addr = res->start + S3C2412_IISRXD; ret = s3c_i2sv2_register_component(&pdev->dev, -1, &s3c2412_i2s_component, diff --git a/sound/soc/samsung/s3c24xx-i2s.c b/sound/soc/samsung/s3c24xx-i2s.c index f31e916..97600df 100644 --- a/sound/soc/samsung/s3c24xx-i2s.c +++ b/sound/soc/samsung/s3c24xx-i2s.c @@ -31,25 +31,19 @@ #include "dma.h" #include "s3c24xx-i2s.h" -static struct s3c2410_dma_client s3c24xx_dma_client_out = { - .name = "I2S PCM Stereo out" -}; - -static struct s3c2410_dma_client s3c24xx_dma_client_in = { - .name = "I2S PCM Stereo in" -}; - static struct s3c_dma_params s3c24xx_i2s_pcm_stereo_out = { - .client = &s3c24xx_dma_client_out, + .client = + (struct s3c2410_dma_client *)&s3c24xx_i2s_pcm_stereo_out, .channel = DMACH_I2S_OUT, - .dma_addr = S3C2410_PA_IIS + S3C2410_IISFIFO, + .ch_name = "tx", .dma_size = 2, }; static struct s3c_dma_params s3c24xx_i2s_pcm_stereo_in = { - .client = &s3c24xx_dma_client_in, + .client = + (struct s3c2410_dma_client *)&s3c24xx_i2s_pcm_stereo_in, .channel = DMACH_I2S_IN, - .dma_addr = S3C2410_PA_IIS + S3C2410_IISFIFO, + .ch_name = "rx", .dma_size = 2, }; @@ -231,18 +225,12 @@ static int s3c24xx_i2s_hw_params(struct snd_pcm_substream *substream, struct snd_pcm_hw_params *params, struct snd_soc_dai *dai) { - struct snd_soc_pcm_runtime *rtd = substream->private_data; - struct s3c_dma_params *dma_data; + struct snd_dmaengine_dai_dma_data *dma_data; u32 iismod; pr_debug("Entered %s\n", __func__); - if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) - dma_data = &s3c24xx_i2s_pcm_stereo_out; - else - dma_data = &s3c24xx_i2s_pcm_stereo_in; - - snd_soc_dai_set_dma_data(rtd->cpu_dai, substream, dma_data); + dma_data = snd_soc_dai_get_dma_data(dai, substream); /* Working copies of register */ iismod = readl(s3c24xx_i2s.regs + S3C2410_IISMOD); @@ -251,11 +239,11 @@ static int s3c24xx_i2s_hw_params(struct snd_pcm_substream *substream, switch (params_format(params)) { case SNDRV_PCM_FORMAT_S8: iismod &= ~S3C2410_IISMOD_16BIT; - dma_data->dma_size = 1; + dma_data->addr_width = 1; break; case SNDRV_PCM_FORMAT_S16_LE: iismod |= S3C2410_IISMOD_16BIT; - dma_data->dma_size = 2; + dma_data->addr_width = 2; break; default: return -EINVAL; @@ -270,8 +258,6 @@ static int s3c24xx_i2s_trigger(struct snd_pcm_substream *substream, int cmd, struct snd_soc_dai *dai) { int ret = 0; - struct s3c_dma_params *dma_data = - snd_soc_dai_get_dma_data(dai, substream); pr_debug("Entered %s\n", __func__); @@ -290,7 +276,6 @@ static int s3c24xx_i2s_trigger(struct snd_pcm_substream *substream, int cmd, else s3c24xx_snd_txctrl(1); - s3c2410_dma_ctrl(dma_data->channel, S3C2410_DMAOP_STARTED); break; case SNDRV_PCM_TRIGGER_STOP: case SNDRV_PCM_TRIGGER_SUSPEND: @@ -380,14 +365,12 @@ static int s3c24xx_i2s_probe(struct snd_soc_dai *dai) { pr_debug("Entered %s\n", __func__); - s3c24xx_i2s.regs = ioremap(S3C2410_PA_IIS, 0x100); - if (s3c24xx_i2s.regs == NULL) - return -ENXIO; + samsung_asoc_init_dma_data(dai, &s3c24xx_i2s_pcm_stereo_out, + &s3c24xx_i2s_pcm_stereo_in); - s3c24xx_i2s.iis_clk = clk_get(dai->dev, "iis"); + s3c24xx_i2s.iis_clk = devm_clk_get(dai->dev, "iis"); if (IS_ERR(s3c24xx_i2s.iis_clk)) { pr_err("failed to get iis_clock\n"); - iounmap(s3c24xx_i2s.regs); return PTR_ERR(s3c24xx_i2s.iis_clk); } clk_enable(s3c24xx_i2s.iis_clk); @@ -474,6 +457,19 @@ static const struct snd_soc_component_driver s3c24xx_i2s_component = { static int s3c24xx_iis_dev_probe(struct platform_device *pdev) { int ret = 0; + struct resource *res; + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!res) { + dev_err(&pdev->dev, "Can't get IO resource.\n"); + return -ENOENT; + } + s3c24xx_i2s.regs = devm_ioremap_resource(&pdev->dev, res); + if (s3c24xx_i2s.regs == NULL) + return -ENXIO; + + s3c24xx_i2s_pcm_stereo_out.dma_addr = res->start + S3C2410_IISFIFO; + s3c24xx_i2s_pcm_stereo_in.dma_addr = res->start + S3C2410_IISFIFO; ret = snd_soc_register_component(&pdev->dev, &s3c24xx_i2s_component, &s3c24xx_i2s_dai, 1);
Use dmaengine instead of legacy s3c24xx DMA API for s3c24xx and s3c2412 Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com> --- v2: use hardcoded dma channel number sound/soc/samsung/Kconfig | 9 +++---- sound/soc/samsung/dmaengine.c | 3 +++ sound/soc/samsung/s3c-i2s-v2.c | 17 +------------ sound/soc/samsung/s3c2412-i2s.c | 47 +++++++++++++++++----------------- sound/soc/samsung/s3c24xx-i2s.c | 56 +++++++++++++++++++---------------------- 5 files changed, 56 insertions(+), 76 deletions(-)