[PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming from runtime PM

Subsystems: framebuffer layer, the rest

STALE5465d

10 messages, 4 authors, 2011-09-20 · open the first message on its own page

[PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming from runtime PM

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2011-07-11 12:35:54 

Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Magnus Damm <magnus.damm@gmail.com>
Cc: Damian Hobson-Garcia <redacted>
---
 drivers/video/sh_mobile_lcdcfb.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

This patch should fix the LCDC panning issue caused by runtime suspend. It
applies on top of Rafael J. Wysocki's pm-domains branch and Damian
Hobson-Garcia's MERAM patches.

diff --git a/drivers/video/sh_mobile_lcdcfb.c b/drivers/video/sh_mobile_lcdcfb.c
index 48a40e3..e9b80bc 100644
--- a/drivers/video/sh_mobile_lcdcfb.c
+++ b/drivers/video/sh_mobile_lcdcfb.c
@@ -256,9 +256,9 @@ struct sh_mobile_lcdc_sys_bus_ops sh_mobile_lcdc_sys_bus_ops = {
 static void sh_mobile_lcdc_clk_on(struct sh_mobile_lcdc_priv *priv)
 {
 	if (atomic_inc_and_test(&priv->hw_usecnt)) {
-		pm_runtime_get_sync(priv->dev);
 		if (priv->dot_clk)
 			clk_enable(priv->dot_clk);
+		pm_runtime_get_sync(priv->dev);
 		if (priv->meram_dev && priv->meram_dev->pdev)
 			pm_runtime_get_sync(&priv->meram_dev->pdev->dev);
 	}
@@ -267,11 +267,11 @@ static void sh_mobile_lcdc_clk_on(struct sh_mobile_lcdc_priv *priv)
 static void sh_mobile_lcdc_clk_off(struct sh_mobile_lcdc_priv *priv)
 {
 	if (atomic_sub_return(1, &priv->hw_usecnt) = -1) {
-		if (priv->dot_clk)
-			clk_disable(priv->dot_clk);
 		if (priv->meram_dev && priv->meram_dev->pdev)
 			pm_runtime_put_sync(&priv->meram_dev->pdev->dev);
 		pm_runtime_put(priv->dev);
+		if (priv->dot_clk)
+			clk_disable(priv->dot_clk);
 	}
 }
 
-- 
Regards,

Laurent Pinchart

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming

From: Guennadi Liakhovetski <hidden>
Date: 2011-07-12 09:33:09 

On Mon, 11 Jul 2011, Laurent Pinchart wrote:

Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Magnus Damm <magnus.damm@gmail.com>
Cc: Damian Hobson-Garcia <redacted>
(testing would have been a bit easier, if prerequisites had been specified 
in the patch;))

Tested-by: Guennadi Liakhovetski <redacted>

Thanks
Guennadi
quoted hunk
---
 drivers/video/sh_mobile_lcdcfb.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

This patch should fix the LCDC panning issue caused by runtime suspend. It
applies on top of Rafael J. Wysocki's pm-domains branch and Damian
Hobson-Garcia's MERAM patches.

diff --git a/drivers/video/sh_mobile_lcdcfb.c b/drivers/video/sh_mobile_lcdcfb.c
index 48a40e3..e9b80bc 100644
--- a/drivers/video/sh_mobile_lcdcfb.c
+++ b/drivers/video/sh_mobile_lcdcfb.c
@@ -256,9 +256,9 @@ struct sh_mobile_lcdc_sys_bus_ops sh_mobile_lcdc_sys_bus_ops = {
 static void sh_mobile_lcdc_clk_on(struct sh_mobile_lcdc_priv *priv)
 {
 	if (atomic_inc_and_test(&priv->hw_usecnt)) {
-		pm_runtime_get_sync(priv->dev);
 		if (priv->dot_clk)
 			clk_enable(priv->dot_clk);
+		pm_runtime_get_sync(priv->dev);
 		if (priv->meram_dev && priv->meram_dev->pdev)
 			pm_runtime_get_sync(&priv->meram_dev->pdev->dev);
 	}
@@ -267,11 +267,11 @@ static void sh_mobile_lcdc_clk_on(struct sh_mobile_lcdc_priv *priv)
 static void sh_mobile_lcdc_clk_off(struct sh_mobile_lcdc_priv *priv)
 {
 	if (atomic_sub_return(1, &priv->hw_usecnt) = -1) {
-		if (priv->dot_clk)
-			clk_disable(priv->dot_clk);
 		if (priv->meram_dev && priv->meram_dev->pdev)
 			pm_runtime_put_sync(&priv->meram_dev->pdev->dev);
 		pm_runtime_put(priv->dev);
+		if (priv->dot_clk)
+			clk_disable(priv->dot_clk);
 	}
 }
 
-- 
Regards,

Laurent Pinchart

--
To unsubscribe from this list: send the line "unsubscribe linux-sh" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming from runtime PM

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2011-07-12 09:37:57 

On Tuesday 12 July 2011 11:33:09 Guennadi Liakhovetski wrote:
On Mon, 11 Jul 2011, Laurent Pinchart wrote:
quoted
Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Magnus Damm <magnus.damm@gmail.com>
Cc: Damian Hobson-Garcia <redacted>
(testing would have been a bit easier, if prerequisites had been specified
in the patch;))
They're specified a couple of lines down ;-)

Tested-by: Guennadi Liakhovetski <redacted>
Thank you.
quoted
---

 drivers/video/sh_mobile_lcdcfb.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

This patch should fix the LCDC panning issue caused by runtime suspend.
It applies on top of Rafael J. Wysocki's pm-domains branch and Damian
Hobson-Garcia's MERAM patches.

diff --git a/drivers/video/sh_mobile_lcdcfb.c
b/drivers/video/sh_mobile_lcdcfb.c index 48a40e3..e9b80bc 100644
--- a/drivers/video/sh_mobile_lcdcfb.c
+++ b/drivers/video/sh_mobile_lcdcfb.c
@@ -256,9 +256,9 @@ struct sh_mobile_lcdc_sys_bus_ops
sh_mobile_lcdc_sys_bus_ops = {

 static void sh_mobile_lcdc_clk_on(struct sh_mobile_lcdc_priv *priv)
 {
 
 	if (atomic_inc_and_test(&priv->hw_usecnt)) {

-		pm_runtime_get_sync(priv->dev);

 		if (priv->dot_clk)
 		
 			clk_enable(priv->dot_clk);

+		pm_runtime_get_sync(priv->dev);

 		if (priv->meram_dev && priv->meram_dev->pdev)
 		
 			pm_runtime_get_sync(&priv->meram_dev->pdev->dev);
 	
 	}

@@ -267,11 +267,11 @@ static void sh_mobile_lcdc_clk_on(struct
sh_mobile_lcdc_priv *priv)

 static void sh_mobile_lcdc_clk_off(struct sh_mobile_lcdc_priv *priv)
 {
 
 	if (atomic_sub_return(1, &priv->hw_usecnt) = -1) {

-		if (priv->dot_clk)
-			clk_disable(priv->dot_clk);

 		if (priv->meram_dev && priv->meram_dev->pdev)
 		
 			pm_runtime_put_sync(&priv->meram_dev->pdev->dev);
 		
 		pm_runtime_put(priv->dev);

+		if (priv->dot_clk)
+			clk_disable(priv->dot_clk);

 	}
 
 }
-- 
Regards,

Laurent Pinchart

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming from runtime PM

From: Paul Mundt <hidden>
Date: 2011-07-13 08:21:13 

On Tue, Jul 12, 2011 at 11:37:57AM +0200, Laurent Pinchart wrote:
On Tuesday 12 July 2011 11:33:09 Guennadi Liakhovetski wrote:
quoted
On Mon, 11 Jul 2011, Laurent Pinchart wrote:
quoted
Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Magnus Damm <magnus.damm@gmail.com>
Cc: Damian Hobson-Garcia <redacted>
(testing would have been a bit easier, if prerequisites had been specified
in the patch;))
They're specified a couple of lines down ;-)
quoted
Tested-by: Guennadi Liakhovetski <redacted>
Thank you.
Folded in on top of the rest of the MERAM changes, thanks.

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming from runtime PM

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2011-09-20 19:30:56 

Hi Florian,

On Monday 11 July 2011 14:35:54 Laurent Pinchart wrote:
Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.
Could you please pick this patch for v3.2 ?

-- 
Regards,

Laurent Pinchart

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming

From: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Date: 2011-09-20 19:52:36 

Hi Laurent,

On 09/20/2011 07:30 PM, Laurent Pinchart wrote:
Hi Florian,

On Monday 11 July 2011 14:35:54 Laurent Pinchart wrote:
quoted
Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.
Could you please pick this patch for v3.2 ?
I assume you meant the currently developed 3.1? (as 3.2 is the next one for
which the patch is lying in my fbdev-next anyway)
Uhm, I originally didn't plan to ask Linus pulling anything before getting rid
of the old stuff, but okay, will go through my patches in the next days and look
what else is appropriate unless Linus beats me and releases 3.1 (although I do
not consider this likely as long as kernel.org is still down)


Best regards,

Florian Tobias Schandinat

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming from runtime PM

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2011-09-20 19:56:43 

Hi Florian,

On Tuesday 20 September 2011 21:52:36 Florian Tobias Schandinat wrote:
On 09/20/2011 07:30 PM, Laurent Pinchart wrote:
quoted
On Monday 11 July 2011 14:35:54 Laurent Pinchart wrote:
quoted
Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.
Could you please pick this patch for v3.2 ?
I assume you meant the currently developed 3.1? (as 3.2 is the next one for
which the patch is lying in my fbdev-next anyway)
No, I meant v3.2. As I haven't received any pull notification in response to 
the patch, and as I had no way to check your fbdev-next tree on kernel.org, I 
just wanted to make sure the patch would be queued for v3.2.

Please ignore this e-mail (and the other similar ones I've just sent) if the 
patches are already in your queue. And thank you for picking them.

Uhm, I originally didn't plan to ask Linus pulling anything before getting
rid of the old stuff, but okay, will go through my patches in the next
days and look what else is appropriate unless Linus beats me and releases
3.1 (although I do not consider this likely as long as kernel.org is still
down)
-- 
Regards,

Laurent Pinchart

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming

From: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Date: 2011-09-20 20:19:25 

Hi Laurent,

On 09/20/2011 07:56 PM, Laurent Pinchart wrote:
Hi Florian,

On Tuesday 20 September 2011 21:52:36 Florian Tobias Schandinat wrote:
quoted
On 09/20/2011 07:30 PM, Laurent Pinchart wrote:
quoted
On Monday 11 July 2011 14:35:54 Laurent Pinchart wrote:
quoted
Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.
Could you please pick this patch for v3.2 ?
I assume you meant the currently developed 3.1? (as 3.2 is the next one for
which the patch is lying in my fbdev-next anyway)
No, I meant v3.2. As I haven't received any pull notification in response to 
the patch, and as I had no way to check your fbdev-next tree on kernel.org, I 
just wanted to make sure the patch would be queued for v3.2.
Just for your info, my tree [1] never was on kernel.org, you can see the patch
in it here [2] (webfrontend). Looks like the patch I got come from your pull
request "SH mobile LCDC cleanups and fixes", hope that's the right version.


Best regards,

Florian Tobias Schandinat


[1] git://github.com/schandinat/linux-2.6.git
[2]
https://github.com/schandinat/linux-2.6/commit/f1ad90da5c0fcb8841cc5e6d66c56f4005d8c960

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming from runtime PM

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2011-09-20 22:44:10 

Hi Florian,

On Tuesday 20 September 2011 22:19:25 Florian Tobias Schandinat wrote:
On 09/20/2011 07:56 PM, Laurent Pinchart wrote:
quoted
On Tuesday 20 September 2011 21:52:36 Florian Tobias Schandinat wrote:
quoted
On 09/20/2011 07:30 PM, Laurent Pinchart wrote:
quoted
On Monday 11 July 2011 14:35:54 Laurent Pinchart wrote:
quoted
Resuming from runtime PM restores all LCDC registers. If the dot clock
is off at that time display panning information will be corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.
Could you please pick this patch for v3.2 ?
I assume you meant the currently developed 3.1? (as 3.2 is the next one
for which the patch is lying in my fbdev-next anyway)
No, I meant v3.2. As I haven't received any pull notification in response
to the patch, and as I had no way to check your fbdev-next tree on
kernel.org, I just wanted to make sure the patch would be queued for
v3.2.
Just for your info, my tree [1] never was on kernel.org, you can see the
patch in it here [2] (webfrontend). Looks like the patch I got come from
your pull request "SH mobile LCDC cleanups and fixes", hope that's the
right version.
Oops, my bad.

I'm rebasing my pending patches on top of your fbdev/fbdev-next tree. There 
are a couple of conflicts, I'll resent the rebased patches to the list.

-- 
Regards,

Laurent Pinchart

Re: [PATCH] fbdev: sh_mobile_lcdc: Turn dot clock on before resuming from runtime PM

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Date: 2011-09-20 23:03:26 

Hi Florian,

On Wednesday 21 September 2011 00:44:10 Laurent Pinchart wrote:
On Tuesday 20 September 2011 22:19:25 Florian Tobias Schandinat wrote:
quoted
On 09/20/2011 07:56 PM, Laurent Pinchart wrote:
quoted
On Tuesday 20 September 2011 21:52:36 Florian Tobias Schandinat wrote:
quoted
On 09/20/2011 07:30 PM, Laurent Pinchart wrote:
quoted
On Monday 11 July 2011 14:35:54 Laurent Pinchart wrote:
quoted
Resuming from runtime PM restores all LCDC registers. If the dot
clock is off at that time display panning information will be
corrupted.

Turn the dot clock on before resuming from runtime PM. Similarly,
turn the clock off after suspending the LCDC.
Could you please pick this patch for v3.2 ?
I assume you meant the currently developed 3.1? (as 3.2 is the next
one for which the patch is lying in my fbdev-next anyway)
No, I meant v3.2. As I haven't received any pull notification in
response to the patch, and as I had no way to check your fbdev-next
tree on kernel.org, I just wanted to make sure the patch would be
queued for v3.2.
Just for your info, my tree [1] never was on kernel.org, you can see the
patch in it here [2] (webfrontend). Looks like the patch I got come from
your pull request "SH mobile LCDC cleanups and fixes", hope that's the
right version.
Oops, my bad.

I'm rebasing my pending patches on top of your fbdev/fbdev-next tree. There
are a couple of conflicts, I'll resent the rebased patches to the list.
Conflicts are only in the WIP/RFC patches. Those are not meant for upstream 
inclusion now, I'll resend them later after getting feedback.

The only pending patches are about the new YUV API. They don't conflict with 
your fbdev-next branch.

-- 
Regards,

Laurent Pinchart
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help