[PATCH] S3C24XX DMA resume regression fix

[PATCH] S3C24XX DMA resume regression fix

[Gusakov Andrey]

In commit bb072c3cf21d1c9a5a2eeb5a00679ee7bf39675b suspend/resume
hooks were added. s3c2410_dma_suspend suspends channels from 0 to
dma_channels. s3c2410_dma_resume resumes channels in reverse
order. So pointer should be decremented insted of being incremented.

Signed-off-by: Gusakov Andrey <dron0gus@gmail.com>
---
 arch/arm/plat-s3c24xx/dma.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/arch/arm/plat-s3c24xx/dma.c b/arch/arm/plat-s3c24xx/dma.c
index 57abec9..16510d5 100644
--- a/arch/arm/plat-s3c24xx/dma.c
+++ b/arch/arm/plat-s3c24xx/dma.c
@@ -1249,7 +1249,7 @@ static void s3c2410_dma_resume(void)
 	struct s3c2410_dma_chan *cp = s3c2410_chans + dma_channels - 1;
 	int channel;
 
-	for (channel = dma_channels - 1; channel >= 0; cp++, channel--)
+	for (channel = dma_channels - 1; channel >= 0; cp--, channel--)
 		s3c2410_dma_resume_chan(cp);
 }
 
-- 
1.7.0.4

[PATCH] S3C24XX DMA resume regression fix

[Sergei Shtylyov]

Hello.

On 13-01-2012 17:47, Gusakov Andrey wrote:

> In commit bb072c3cf21d1c9a5a2eeb5a00679ee7bf39675b suspend/resume

    Please also specify that commit's summary in parens.

> hooks were added. s3c2410_dma_suspend suspends channels from 0 to
> dma_channels. s3c2410_dma_resume resumes channels in reverse
> order. So pointer should be decremented insted of being incremented.

> Signed-off-by: Gusakov Andrey<dron0gus@gmail.com>

WBR, Sergei

[PATCH] S3C24XX DMA resume regression fix

[Sergey Lapin]

On Sun, Jan 15, 2012 at 01:52:17AM +0400, Sergei Shtylyov wrote:
> Hello.
> 
> On 13-01-2012 17:47, Gusakov Andrey wrote:
> 
> >In commit bb072c3cf21d1c9a5a2eeb5a00679ee7bf39675b suspend/resume
> 
>    Please also specify that commit's summary in parens.

Could you please point me in direction of patch submission
guide where that is explained.

The commit message summary is
ARM / Samsung: Use struct syscore_ops for "core" power management

But that is not revert of that commit, so I don't really
see a point. This is bug fix which fixes resume code path.

Thanks a lot,
S.

[PATCH] S3C24XX DMA resume regression fix

[Russell King - ARM Linux]

On Fri, Jan 20, 2012 at 06:49:02AM -0500, Sergey Lapin wrote:
> On Sun, Jan 15, 2012 at 01:52:17AM +0400, Sergei Shtylyov wrote:
> > Hello.
> > 
> > On 13-01-2012 17:47, Gusakov Andrey wrote:
> > 
> > >In commit bb072c3cf21d1c9a5a2eeb5a00679ee7bf39675b suspend/resume
> > 
> >    Please also specify that commit's summary in parens.
> 
> Could you please point me in direction of patch submission
> guide where that is explained.

It's something that's been asked for by Linus - and as Linus can refuse
to pull anything he doesn't like, it's something that we'd better satisfy
if we want patches to hit mainline.

Like everything in kernel development, what's written in documentation is
never the full story; things change through discussions on mailing lists,
new guidelines get created, and the documents become stale.  That's the
nature of this.

Anyone is free to update the documentation, so if you think it's got a
shortcoming, please feel free to send a documentation patch.

> The commit message summary is
> ARM / Samsung: Use struct syscore_ops for "core" power management
> 
> But that is not revert of that commit, so I don't really
> see a point. This is bug fix which fixes resume code path.

The point is that if you include a commit ID, then you should include
the summary, so people reading the commit message don't have to faff
around trying to find out what you're referring to.

See, for example, http://lkml.org/lkml/2010/6/1/498

[PATCH] S3C24XX DMA resume regression fix

[Sergey Lapin]

> > > >In commit bb072c3cf21d1c9a5a2eeb5a00679ee7bf39675b suspend/resume
> > > 
> > >    Please also specify that commit's summary in parens.
> > 
> > Could you please point me in direction of patch submission
> > guide where that is explained.
> 
> It's something that's been asked for by Linus - and as Linus can refuse
> to pull anything he doesn't like, it's something that we'd better satisfy
> if we want patches to hit mainline.
> 
> Like everything in kernel development, what's written in documentation is
> never the full story; things change through discussions on mailing lists,
> new guidelines get created, and the documents become stale.  That's the
> nature of this.
...
> The point is that if you include a commit ID, then you should include
> the summary, so people reading the commit message don't have to faff
> around trying to find out what you're referring to.
> 
> See, for example, http://lkml.org/lkml/2010/6/1/498

Thanks a lot for good explanation,

S.