diff mbox series

dma: add support for reporting pause and resume separately

Message ID 20180702130819eucas1p2e11d8fabec93ba78d9e84f4ca6e5fd65~9jzNWZMqQ1216712167eucas1p27@eucas1p2.samsung.com
State Accepted
Commit d8095f94e19581057bcad35b8a725aa739e77595
Headers show
Series dma: add support for reporting pause and resume separately | expand

Commit Message

Marek Szyprowski July 2, 2018, 1:08 p.m. UTC
'cmd_pause' DMA channel capability means that respective DMA engine
supports both pausing and resuming given DMA channel. However, in some
cases it is important to know if DMA channel can be paused without the
need to resume it. This is a typical requirement for proper residue
reading on transfer timeout in UART drivers. There are also some DMA
engines with limited hardware, which doesn't really support resuming.

Reporting pause and resume capabilities separately allows UART drivers to
properly check for the really required capabilities and operate in DMA
mode also in systems with limited DMA hardware. On the other hand drivers,
which rely on full channel suspend/resume support, should now check for
both 'pause' and 'resume' features.

Existing clients of dma_get_slave_caps() have been checked and the only
driver which rely on proper channel resuming is soc-generic-dmaengine-pcm
driver, which has been updated to check the newly added capability.
Existing 'cmd_pause' now only indicates that DMA engine support pausing
given DMA channel.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>

---
This patch is a follow-up of the Vinod's advise from the following
discussion:
https://www.spinics.net/lists/linux-samsung-soc/msg63166.html
---
 drivers/dma/dmaengine.c               | 8 ++------
 include/linux/dmaengine.h             | 5 ++++-
 sound/soc/soc-generic-dmaengine-pcm.c | 2 +-
 3 files changed, 7 insertions(+), 8 deletions(-)

-- 
2.17.1

Comments

Vinod Koul July 4, 2018, 7 a.m. UTC | #1
Hi Marek,

On 02-07-18, 15:08, Marek Szyprowski wrote:
> 'cmd_pause' DMA channel capability means that respective DMA engine

> supports both pausing and resuming given DMA channel. However, in some

> cases it is important to know if DMA channel can be paused without the

> need to resume it. This is a typical requirement for proper residue

> reading on transfer timeout in UART drivers. There are also some DMA

> engines with limited hardware, which doesn't really support resuming.


Am curious given that your hardware does not support resume, what was motivation
for adding pause?

> Reporting pause and resume capabilities separately allows UART drivers to

> properly check for the really required capabilities and operate in DMA

> mode also in systems with limited DMA hardware. On the other hand drivers,

> which rely on full channel suspend/resume support, should now check for

> both 'pause' and 'resume' features.

> 

> Existing clients of dma_get_slave_caps() have been checked and the only

> driver which rely on proper channel resuming is soc-generic-dmaengine-pcm

> driver, which has been updated to check the newly added capability.

> Existing 'cmd_pause' now only indicates that DMA engine support pausing

> given DMA channel.


The change looks fine to me. I was hoping that serial would also check
this..

Mark, Lars you okay with this?

> 

> Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>

> ---

> This patch is a follow-up of the Vinod's advise from the following

> discussion:

> https://www.spinics.net/lists/linux-samsung-soc/msg63166.html

> ---

>  drivers/dma/dmaengine.c               | 8 ++------

>  include/linux/dmaengine.h             | 5 ++++-

>  sound/soc/soc-generic-dmaengine-pcm.c | 2 +-

>  3 files changed, 7 insertions(+), 8 deletions(-)

> 

> diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c

> index 08ba8473a284..84ac38dbdb65 100644

> --- a/drivers/dma/dmaengine.c

> +++ b/drivers/dma/dmaengine.c

> @@ -500,12 +500,8 @@ int dma_get_slave_caps(struct dma_chan *chan, struct dma_slave_caps *caps)

>  	caps->max_burst = device->max_burst;

>  	caps->residue_granularity = device->residue_granularity;

>  	caps->descriptor_reuse = device->descriptor_reuse;

> -

> -	/*

> -	 * Some devices implement only pause (e.g. to get residuum) but no

> -	 * resume. However cmd_pause is advertised as pause AND resume.

> -	 */

> -	caps->cmd_pause = !!(device->device_pause && device->device_resume);

> +	caps->cmd_pause = !!device->device_pause;

> +	caps->cmd_resume = !!device->device_resume;

>  	caps->cmd_terminate = !!device->device_terminate_all;

>  

>  	return 0;

> diff --git a/include/linux/dmaengine.h b/include/linux/dmaengine.h

> index 861be5cab1df..c8c3a7a93802 100644

> --- a/include/linux/dmaengine.h

> +++ b/include/linux/dmaengine.h

> @@ -415,7 +415,9 @@ enum dma_residue_granularity {

>   *	each type, the dma controller should set BIT(<TYPE>) and same

>   *	should be checked by controller as well

>   * @max_burst: max burst capability per-transfer

> - * @cmd_pause: true, if pause and thereby resume is supported

> + * @cmd_pause: true, if pause is supported (i.e. for reading residue or

> + *	       for resume later)

> + * @cmd_resume: true, if resume is supported

>   * @cmd_terminate: true, if terminate cmd is supported

>   * @residue_granularity: granularity of the reported transfer residue

>   * @descriptor_reuse: if a descriptor can be reused by client and

> @@ -427,6 +429,7 @@ struct dma_slave_caps {

>  	u32 directions;

>  	u32 max_burst;

>  	bool cmd_pause;

> +	bool cmd_resume;

>  	bool cmd_terminate;

>  	enum dma_residue_granularity residue_granularity;

>  	bool descriptor_reuse;

> diff --git a/sound/soc/soc-generic-dmaengine-pcm.c b/sound/soc/soc-generic-dmaengine-pcm.c

> index 56a541b9ff9e..76c46d793843 100644

> --- a/sound/soc/soc-generic-dmaengine-pcm.c

> +++ b/sound/soc/soc-generic-dmaengine-pcm.c

> @@ -156,7 +156,7 @@ static int dmaengine_pcm_set_runtime_hwparams(struct snd_pcm_substream *substrea

>  

>  	ret = dma_get_slave_caps(chan, &dma_caps);

>  	if (ret == 0) {

> -		if (dma_caps.cmd_pause)

> +		if (dma_caps.cmd_pause && dma_caps.cmd_resume)

>  			hw.info |= SNDRV_PCM_INFO_PAUSE | SNDRV_PCM_INFO_RESUME;

>  		if (dma_caps.residue_granularity <= DMA_RESIDUE_GRANULARITY_SEGMENT)

>  			hw.info |= SNDRV_PCM_INFO_BATCH;

> -- 

> 2.17.1

> 

> --

> To unsubscribe from this list: send the line "unsubscribe dmaengine" in

> the body of a message to majordomo@vger.kernel.org

> More majordomo info at  http://vger.kernel.org/majordomo-info.html


-- 
~Vinod
Andy Shevchenko July 4, 2018, 10:38 a.m. UTC | #2
On Wed, Jul 4, 2018 at 10:00 AM, Vinod <vkoul@kernel.org> wrote:
> Hi Marek,

>

> On 02-07-18, 15:08, Marek Szyprowski wrote:

>> 'cmd_pause' DMA channel capability means that respective DMA engine

>> supports both pausing and resuming given DMA channel. However, in some

>> cases it is important to know if DMA channel can be paused without the

>> need to resume it. This is a typical requirement for proper residue

>> reading on transfer timeout in UART drivers. There are also some DMA

>> engines with limited hardware, which doesn't really support resuming.

>

> Am curious given that your hardware does not support resume, what was motivation

> for adding pause?

>

>> Reporting pause and resume capabilities separately allows UART drivers to

>> properly check for the really required capabilities and operate in DMA

>> mode also in systems with limited DMA hardware. On the other hand drivers,

>> which rely on full channel suspend/resume support, should now check for

>> both 'pause' and 'resume' features.

>>

>> Existing clients of dma_get_slave_caps() have been checked and the only

>> driver which rely on proper channel resuming is soc-generic-dmaengine-pcm

>> driver, which has been updated to check the newly added capability.

>> Existing 'cmd_pause' now only indicates that DMA engine support pausing

>> given DMA channel.

>

> The change looks fine to me. I was hoping that serial would also check

> this..


AFAIU serial does not need to resume.

-- 
With Best Regards,
Andy Shevchenko
Vinod Koul July 6, 2018, 6 a.m. UTC | #3
On 04-07-18, 13:38, Andy Shevchenko wrote:
> On Wed, Jul 4, 2018 at 10:00 AM, Vinod <vkoul@kernel.org> wrote:

> > Hi Marek,

> >

> > On 02-07-18, 15:08, Marek Szyprowski wrote:

> >> 'cmd_pause' DMA channel capability means that respective DMA engine

> >> supports both pausing and resuming given DMA channel. However, in some

> >> cases it is important to know if DMA channel can be paused without the

> >> need to resume it. This is a typical requirement for proper residue

> >> reading on transfer timeout in UART drivers. There are also some DMA

> >> engines with limited hardware, which doesn't really support resuming.

> >

> > Am curious given that your hardware does not support resume, what was motivation

> > for adding pause?

> >

> >> Reporting pause and resume capabilities separately allows UART drivers to

> >> properly check for the really required capabilities and operate in DMA

> >> mode also in systems with limited DMA hardware. On the other hand drivers,

> >> which rely on full channel suspend/resume support, should now check for

> >> both 'pause' and 'resume' features.

> >>

> >> Existing clients of dma_get_slave_caps() have been checked and the only

> >> driver which rely on proper channel resuming is soc-generic-dmaengine-pcm

> >> driver, which has been updated to check the newly added capability.

> >> Existing 'cmd_pause' now only indicates that DMA engine support pausing

> >> given DMA channel.

> >

> > The change looks fine to me. I was hoping that serial would also check

> > this..

> 

> AFAIU serial does not need to resume.


I meant the serial should check for Pause capability..

-- 
~Vinod
Marek Szyprowski July 6, 2018, 6:22 a.m. UTC | #4
Hi Vinod,

On 2018-07-06 08:00, Vinod wrote:
> On 04-07-18, 13:38, Andy Shevchenko wrote:

>> On Wed, Jul 4, 2018 at 10:00 AM, Vinod <vkoul@kernel.org> wrote:

>>> Hi Marek,

>>>

>>> On 02-07-18, 15:08, Marek Szyprowski wrote:

>>>> 'cmd_pause' DMA channel capability means that respective DMA engine

>>>> supports both pausing and resuming given DMA channel. However, in some

>>>> cases it is important to know if DMA channel can be paused without the

>>>> need to resume it. This is a typical requirement for proper residue

>>>> reading on transfer timeout in UART drivers. There are also some DMA

>>>> engines with limited hardware, which doesn't really support resuming.

>>> Am curious given that your hardware does not support resume, what was motivation

>>> for adding pause?

>>>

>>>> Reporting pause and resume capabilities separately allows UART drivers to

>>>> properly check for the really required capabilities and operate in DMA

>>>> mode also in systems with limited DMA hardware. On the other hand drivers,

>>>> which rely on full channel suspend/resume support, should now check for

>>>> both 'pause' and 'resume' features.

>>>>

>>>> Existing clients of dma_get_slave_caps() have been checked and the only

>>>> driver which rely on proper channel resuming is soc-generic-dmaengine-pcm

>>>> driver, which has been updated to check the newly added capability.

>>>> Existing 'cmd_pause' now only indicates that DMA engine support pausing

>>>> given DMA channel.

>>> The change looks fine to me. I was hoping that serial would also check

>>> this..

>> AFAIU serial does not need to resume.

> I meant the serial should check for Pause capability..


Yes, I want to add such check to Samsung Exynos serial driver once this 
change is
merged. PL330 DMA engine (which is used in Samsung Exynos SoCs) supports 
only
pause, it is not possible to check that functionality from serial driver 
yet.

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland
Mark Brown July 9, 2018, 11:17 a.m. UTC | #5
On Wed, Jul 04, 2018 at 12:30:33PM +0530, Vinod wrote:

> > Existing clients of dma_get_slave_caps() have been checked and the only

> > driver which rely on proper channel resuming is soc-generic-dmaengine-pcm

> > driver, which has been updated to check the newly added capability.

> > Existing 'cmd_pause' now only indicates that DMA engine support pausing

> > given DMA channel.


> The change looks fine to me. I was hoping that serial would also check

> this..


> Mark, Lars you okay with this?


Acked-by: Mark Brown <broonie@kernel.org>
Vinod Koul July 9, 2018, 5:29 p.m. UTC | #6
On 02-07-18, 15:08, Marek Szyprowski wrote:
> 'cmd_pause' DMA channel capability means that respective DMA engine

> supports both pausing and resuming given DMA channel. However, in some

> cases it is important to know if DMA channel can be paused without the

> need to resume it. This is a typical requirement for proper residue

> reading on transfer timeout in UART drivers. There are also some DMA

> engines with limited hardware, which doesn't really support resuming.

> 

> Reporting pause and resume capabilities separately allows UART drivers to

> properly check for the really required capabilities and operate in DMA

> mode also in systems with limited DMA hardware. On the other hand drivers,

> which rely on full channel suspend/resume support, should now check for

> both 'pause' and 'resume' features.

> 

> Existing clients of dma_get_slave_caps() have been checked and the only

> driver which rely on proper channel resuming is soc-generic-dmaengine-pcm

> driver, which has been updated to check the newly added capability.

> Existing 'cmd_pause' now only indicates that DMA engine support pausing

> given DMA channel.


Applied after fixing the subsystem name, thanks

-- 
~Vinod
diff mbox series

Patch

diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c
index 08ba8473a284..84ac38dbdb65 100644
--- a/drivers/dma/dmaengine.c
+++ b/drivers/dma/dmaengine.c
@@ -500,12 +500,8 @@  int dma_get_slave_caps(struct dma_chan *chan, struct dma_slave_caps *caps)
 	caps->max_burst = device->max_burst;
 	caps->residue_granularity = device->residue_granularity;
 	caps->descriptor_reuse = device->descriptor_reuse;
-
-	/*
-	 * Some devices implement only pause (e.g. to get residuum) but no
-	 * resume. However cmd_pause is advertised as pause AND resume.
-	 */
-	caps->cmd_pause = !!(device->device_pause && device->device_resume);
+	caps->cmd_pause = !!device->device_pause;
+	caps->cmd_resume = !!device->device_resume;
 	caps->cmd_terminate = !!device->device_terminate_all;
 
 	return 0;
diff --git a/include/linux/dmaengine.h b/include/linux/dmaengine.h
index 861be5cab1df..c8c3a7a93802 100644
--- a/include/linux/dmaengine.h
+++ b/include/linux/dmaengine.h
@@ -415,7 +415,9 @@  enum dma_residue_granularity {
  *	each type, the dma controller should set BIT(<TYPE>) and same
  *	should be checked by controller as well
  * @max_burst: max burst capability per-transfer
- * @cmd_pause: true, if pause and thereby resume is supported
+ * @cmd_pause: true, if pause is supported (i.e. for reading residue or
+ *	       for resume later)
+ * @cmd_resume: true, if resume is supported
  * @cmd_terminate: true, if terminate cmd is supported
  * @residue_granularity: granularity of the reported transfer residue
  * @descriptor_reuse: if a descriptor can be reused by client and
@@ -427,6 +429,7 @@  struct dma_slave_caps {
 	u32 directions;
 	u32 max_burst;
 	bool cmd_pause;
+	bool cmd_resume;
 	bool cmd_terminate;
 	enum dma_residue_granularity residue_granularity;
 	bool descriptor_reuse;
diff --git a/sound/soc/soc-generic-dmaengine-pcm.c b/sound/soc/soc-generic-dmaengine-pcm.c
index 56a541b9ff9e..76c46d793843 100644
--- a/sound/soc/soc-generic-dmaengine-pcm.c
+++ b/sound/soc/soc-generic-dmaengine-pcm.c
@@ -156,7 +156,7 @@  static int dmaengine_pcm_set_runtime_hwparams(struct snd_pcm_substream *substrea
 
 	ret = dma_get_slave_caps(chan, &dma_caps);
 	if (ret == 0) {
-		if (dma_caps.cmd_pause)
+		if (dma_caps.cmd_pause && dma_caps.cmd_resume)
 			hw.info |= SNDRV_PCM_INFO_PAUSE | SNDRV_PCM_INFO_RESUME;
 		if (dma_caps.residue_granularity <= DMA_RESIDUE_GRANULARITY_SEGMENT)
 			hw.info |= SNDRV_PCM_INFO_BATCH;