[PATCH] mmc: tmio: fix PM in DMA mode

[PATCH] mmc: tmio: fix PM in DMA mode

[Guennadi Liakhovetski]

DMA mode has to be re-enabled during a resume.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
---
 drivers/mmc/host/tmio_mmc_dma.c |   21 ++++++++++++---------
 drivers/mmc/host/tmio_mmc_pio.c |    1 +
 2 files changed, 13 insertions(+), 9 deletions(-)

diff --git a/drivers/mmc/host/tmio_mmc_dma.c b/drivers/mmc/host/tmio_mmc_dma.c
index d3de74a..25f1ad6 100644
--- a/drivers/mmc/host/tmio_mmc_dma.c
+++ b/drivers/mmc/host/tmio_mmc_dma.c
@@ -256,7 +256,10 @@ static bool tmio_mmc_filter(struct dma_chan *chan, void *arg)
 void tmio_mmc_request_dma(struct tmio_mmc_host *host, struct tmio_mmc_data *pdata)
 {
 	/* We can only either use DMA for both Tx and Rx or not use it at all */
-	if (pdata->dma) {
+	if (!pdata->dma)
+		return;
+
+	if (!host->chan_tx && !host->chan_rx) {
 		dma_cap_mask_t mask;
 
 		dma_cap_zero(mask);
@@ -284,18 +287,18 @@ void tmio_mmc_request_dma(struct tmio_mmc_host *host, struct tmio_mmc_data *pdat
 
 		tasklet_init(&host->dma_complete, tmio_mmc_tasklet_fn, (unsigned long)host);
 		tasklet_init(&host->dma_issue, tmio_mmc_issue_tasklet_fn, (unsigned long)host);
+	}
 
-		tmio_mmc_enable_dma(host, true);
+	tmio_mmc_enable_dma(host, true);
+
+	return;
 
-		return;
 ebouncebuf:
-		dma_release_channel(host->chan_rx);
-		host->chan_rx = NULL;
+	dma_release_channel(host->chan_rx);
+	host->chan_rx = NULL;
 ereqrx:
-		dma_release_channel(host->chan_tx);
-		host->chan_tx = NULL;
-		return;
-	}
+	dma_release_channel(host->chan_tx);
+	host->chan_tx = NULL;
 }
 
 void tmio_mmc_release_dma(struct tmio_mmc_host *host)
diff --git a/drivers/mmc/host/tmio_mmc_pio.c b/drivers/mmc/host/tmio_mmc_pio.c
index 26598f1..27fc11c 100644
--- a/drivers/mmc/host/tmio_mmc_pio.c
+++ b/drivers/mmc/host/tmio_mmc_pio.c
@@ -995,6 +995,7 @@ int tmio_mmc_host_resume(struct device *dev)
 		pm_runtime_get_sync(dev);
 
 	tmio_mmc_reset(mmc_priv(mmc));
+	tmio_mmc_request_dma(host, host->pdata);
 
 	return mmc_resume_host(mmc);
 }
-- 
1.7.2.5

Re: [PATCH] mmc: tmio: fix PM in DMA mode

[Chris Ball]

Hi Guennadi,

On Fri, Apr 29 2011, Guennadi Liakhovetski wrote:
> DMA mode has to be re-enabled during a resume.
>
> Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
> ---
>  drivers/mmc/host/tmio_mmc_dma.c |   21 ++++++++++++---------
>  drivers/mmc/host/tmio_mmc_pio.c |    1 +
>  2 files changed, 13 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/mmc/host/tmio_mmc_dma.c b/drivers/mmc/host/tmio_mmc_dma.c
> index d3de74a..25f1ad6 100644
> --- a/drivers/mmc/host/tmio_mmc_dma.c
> +++ b/drivers/mmc/host/tmio_mmc_dma.c
> @@ -256,7 +256,10 @@ static bool tmio_mmc_filter(struct dma_chan *chan, void *arg)
>  void tmio_mmc_request_dma(struct tmio_mmc_host *host, struct tmio_mmc_data *pdata)
>  {
>  	/* We can only either use DMA for both Tx and Rx or not use it at all */
> -	if (pdata->dma) {
> +	if (!pdata->dma)
> +		return;
> +
> +	if (!host->chan_tx && !host->chan_rx) {
>  		dma_cap_mask_t mask;
>  
>  		dma_cap_zero(mask);
> @@ -284,18 +287,18 @@ void tmio_mmc_request_dma(struct tmio_mmc_host *host, struct tmio_mmc_data *pdat
>  
>  		tasklet_init(&host->dma_complete, tmio_mmc_tasklet_fn, (unsigned long)host);
>  		tasklet_init(&host->dma_issue, tmio_mmc_issue_tasklet_fn, (unsigned long)host);
> +	}
>  
> -		tmio_mmc_enable_dma(host, true);
> +	tmio_mmc_enable_dma(host, true);
> +
> +	return;
>  
> -		return;
>  ebouncebuf:
> -		dma_release_channel(host->chan_rx);
> -		host->chan_rx = NULL;
> +	dma_release_channel(host->chan_rx);
> +	host->chan_rx = NULL;
>  ereqrx:
> -		dma_release_channel(host->chan_tx);
> -		host->chan_tx = NULL;
> -		return;
> -	}
> +	dma_release_channel(host->chan_tx);
> +	host->chan_tx = NULL;
>  }
>  
>  void tmio_mmc_release_dma(struct tmio_mmc_host *host)
> diff --git a/drivers/mmc/host/tmio_mmc_pio.c b/drivers/mmc/host/tmio_mmc_pio.c
> index 26598f1..27fc11c 100644
> --- a/drivers/mmc/host/tmio_mmc_pio.c
> +++ b/drivers/mmc/host/tmio_mmc_pio.c
> @@ -995,6 +995,7 @@ int tmio_mmc_host_resume(struct device *dev)
>  		pm_runtime_get_sync(dev);
>  
>  	tmio_mmc_reset(mmc_priv(mmc));
> +	tmio_mmc_request_dma(host, host->pdata);
>  
>  	return mmc_resume_host(mmc);
>  }

This patch doesn't apply against current mmc-next.

Please could you send a full patchset against mmc-next for tmio for .40?

Thanks!

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: [PATCH] mmc: tmio: fix PM in DMA mode

[Guennadi Liakhovetski]

Hi Chris

On Tue, 10 May 2011, Chris Ball wrote:

> Hi Guennadi,
> 
> On Fri, Apr 29 2011, Guennadi Liakhovetski wrote:
> > DMA mode has to be re-enabled during a resume.
> >
> > Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
> > ---
> >  drivers/mmc/host/tmio_mmc_dma.c |   21 ++++++++++++---------
> >  drivers/mmc/host/tmio_mmc_pio.c |    1 +
> >  2 files changed, 13 insertions(+), 9 deletions(-)

[snip]

> Please could you send a full patchset against mmc-next for tmio for .40?

Can I also base my patches on next-2.6? I need both - latest mmc and pm, 
so, using next would make it much easier.

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: [PATCH] mmc: tmio: fix PM in DMA mode

[Guennadi Liakhovetski]

On Thu, 12 May 2011, Guennadi Liakhovetski wrote:

> Hi Chris
> 
> On Tue, 10 May 2011, Chris Ball wrote:
> 
> > Hi Guennadi,
> > 
> > On Fri, Apr 29 2011, Guennadi Liakhovetski wrote:
> > > DMA mode has to be re-enabled during a resume.
> > >
> > > Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
> > > ---
> > >  drivers/mmc/host/tmio_mmc_dma.c |   21 ++++++++++++---------
> > >  drivers/mmc/host/tmio_mmc_pio.c |    1 +
> > >  2 files changed, 13 insertions(+), 9 deletions(-)
> 
> [snip]
> 
> > Please could you send a full patchset against mmc-next for tmio for .40?
> 
> Can I also base my patches on next-2.6? I need both - latest mmc and pm, 
> so, using next would make it much easier.

I rebased my stack on top of today's next, no changes were required, so, 
here's the order, I have the patches in my tree:

https://patchwork.kernel.org/patch/711251/
https://patchwork.kernel.org/patch/757602/
https://patchwork.kernel.org/patch/757612/
https://patchwork.kernel.org/patch/724081/
https://patchwork.kernel.org/patch/761882/
https://patchwork.kernel.org/patch/761902/
https://patchwork.kernel.org/patch/761922/
https://patchwork.kernel.org/patch/761942/
https://patchwork.kernel.org/patch/777232/
https://patchwork.kernel.org/patch/777262/

As long as we manage to get these in 2.6.40 merge window early enough, 
we'll be able to push related arch patches via shmobile trees.

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: [PATCH] mmc: tmio: fix PM in DMA mode

[Chris Ball]

Hi Guennadi,

On Thu, May 12 2011, Guennadi Liakhovetski wrote:
>> > Please could you send a full patchset against mmc-next for tmio for .40?
>> 
>> Can I also base my patches on next-2.6? I need both - latest mmc and pm, 
>> so, using next would make it much easier.
>
> I rebased my stack on top of today's next, no changes were required, so, 
> here's the order, I have the patches in my tree:
>
> https://patchwork.kernel.org/patch/711251/
> https://patchwork.kernel.org/patch/757602/
> https://patchwork.kernel.org/patch/757612/
> https://patchwork.kernel.org/patch/724081/
> https://patchwork.kernel.org/patch/761882/
> https://patchwork.kernel.org/patch/761902/
> https://patchwork.kernel.org/patch/761922/
> https://patchwork.kernel.org/patch/761942/
> https://patchwork.kernel.org/patch/777232/
> https://patchwork.kernel.org/patch/777262/
>
> As long as we manage to get these in 2.6.40 merge window early enough, 
> we'll be able to push related arch patches via shmobile trees.

Thanks, all looks good to me -- pushed to mmc-next, and I'll try to
merge early in .40.

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: [PATCH] mmc: tmio: fix PM in DMA mode

[Chris Ball]

Hi Guennadi,

On Thu, May 12, 2011 at 12:19:10PM +0200, Guennadi Liakhovetski wrote:
> As long as we manage to get these in 2.6.40 merge window early enough, 
> we'll be able to push related arch patches via shmobile trees.

Linus has pulled the MMC merge now; fire away.

Thanks,

-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child