Patchwork [Quantal] Fix graphics corruption regression on older ATI

login
register
mail settings
Submitter Stefan Bader
Date Sept. 25, 2012, 9:15 a.m.
Message ID <1348564534-5489-1-git-send-email-stefan.bader@canonical.com>
Download mbox | patch
Permalink /patch/186745/
State New
Headers show

Comments

Stefan Bader - Sept. 25, 2012, 9:15 a.m.
I saw this on my older Dell 1521. Graphics were unusable and often
lead to complete lockups. Quick testing showed that this seemed to
be fixed in between 3.6-rc5 and 3.6-rc6.
From the few patches in between, the below stood out the most likely
one. It has been submitted to stable on Sep-20 but not yet was
picked up. But testing with this applied would make my Dell work
again.

-Stefan


From b9a1600b7a1378f5182b6ca7b8d0d9d175950168 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Christian=20K=C3=B6nig?= <deathsimple@vodafone.de>
Date: Thu, 20 Sep 2012 16:57:00 +0200
Subject: [PATCH] UBUNTU: (pre-stable) drm/radeon: make 64bit fences more robust v3 (3.5 stable)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Only increase the higher 32bits if we really detect a wrap around.

v2: instead of increasing the higher 32bits just use the higher
    32bits from the last emitted fence.
v3: also use last emitted fence value as upper limit.

The intention of this patch is to make fences as robust as
they where before introducing 64bit fences. This is
necessary because on older systems it looks like the fence
value gets corrupted on initialization.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=51344

Should also fix:
https://bugs.freedesktop.org/show_bug.cgi?id=54129
https://bugs.freedesktop.org/show_bug.cgi?id=54662
https://bugzilla.redhat.com/show_bug.cgi?id=846505
https://bugzilla.redhat.com/show_bug.cgi?id=845639

This is the 3.5 stable version of the patch:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=f492c171a38d77fc13a8998a0721f2da50835224
3.4 and previous kernels do not need to be patched.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

BugLink: http://bugs.launchpad.net/bugs/1029582

(cherry-picked from stable mailing list submission)
Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
---
 drivers/gpu/drm/radeon/radeon_fence.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
Colin King - Sept. 25, 2012, 9:21 a.m.
On 25/09/12 10:15, Stefan Bader wrote:
> I saw this on my older Dell 1521. Graphics were unusable and often
> lead to complete lockups. Quick testing showed that this seemed to
> be fixed in between 3.6-rc5 and 3.6-rc6.
>  From the few patches in between, the below stood out the most likely
> one. It has been submitted to stable on Sep-20 but not yet was
> picked up. But testing with this applied would make my Dell work
> again.
>
> -Stefan
>
>
>  From b9a1600b7a1378f5182b6ca7b8d0d9d175950168 Mon Sep 17 00:00:00 2001
> From: =?UTF-8?q?Christian=20K=C3=B6nig?= <deathsimple@vodafone.de>
> Date: Thu, 20 Sep 2012 16:57:00 +0200
> Subject: [PATCH] UBUNTU: (pre-stable) drm/radeon: make 64bit fences more robust v3 (3.5 stable)
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
>
> Only increase the higher 32bits if we really detect a wrap around.
>
> v2: instead of increasing the higher 32bits just use the higher
>      32bits from the last emitted fence.
> v3: also use last emitted fence value as upper limit.
>
> The intention of this patch is to make fences as robust as
> they where before introducing 64bit fences. This is
> necessary because on older systems it looks like the fence
> value gets corrupted on initialization.
>
> Fixes:
> https://bugs.freedesktop.org/show_bug.cgi?id=51344
>
> Should also fix:
> https://bugs.freedesktop.org/show_bug.cgi?id=54129
> https://bugs.freedesktop.org/show_bug.cgi?id=54662
> https://bugzilla.redhat.com/show_bug.cgi?id=846505
> https://bugzilla.redhat.com/show_bug.cgi?id=845639
>
> This is the 3.5 stable version of the patch:
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=f492c171a38d77fc13a8998a0721f2da50835224
> 3.4 and previous kernels do not need to be patched.
>
> Signed-off-by: Christian König <deathsimple@vodafone.de>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
>
> BugLink: http://bugs.launchpad.net/bugs/1029582
>
> (cherry-picked from stable mailing list submission)
> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
> ---
>   drivers/gpu/drm/radeon/radeon_fence.c |    8 +++++---
>   1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
> index 11f5f40..71298ce 100644
> --- a/drivers/gpu/drm/radeon/radeon_fence.c
> +++ b/drivers/gpu/drm/radeon/radeon_fence.c
> @@ -75,7 +75,7 @@ int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence)
>
>   void radeon_fence_process(struct radeon_device *rdev, int ring)
>   {
> -	uint64_t seq, last_seq;
> +	uint64_t seq, last_seq, last_emitted;
>   	unsigned count_loop = 0;
>   	bool wake = false;
>
> @@ -102,13 +102,15 @@ void radeon_fence_process(struct radeon_device *rdev, int ring)
>   	 */
>   	last_seq = atomic64_read(&rdev->fence_drv[ring].last_seq);
>   	do {
> +		last_emitted = rdev->fence_drv[ring].seq;
>   		seq = radeon_fence_read(rdev, ring);
>   		seq |= last_seq & 0xffffffff00000000LL;
>   		if (seq < last_seq) {
> -			seq += 0x100000000LL;
> +			seq &= 0xffffffff;
> +			seq |= last_emitted & 0xffffffff00000000LL;
>   		}
>
> -		if (seq == last_seq) {
> +		if (seq <= last_seq || seq > last_emitted) {
>   			break;
>   		}
>   		/* If we loop over we don't want to return without
>

Nice find.  Acked-by: Colin Ian King <colin.king@canonical.com>
Tim Gardner - Sept. 25, 2012, 12:24 p.m.

Patch

diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index 11f5f40..71298ce 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -75,7 +75,7 @@  int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence)
 
 void radeon_fence_process(struct radeon_device *rdev, int ring)
 {
-	uint64_t seq, last_seq;
+	uint64_t seq, last_seq, last_emitted;
 	unsigned count_loop = 0;
 	bool wake = false;
 
@@ -102,13 +102,15 @@  void radeon_fence_process(struct radeon_device *rdev, int ring)
 	 */
 	last_seq = atomic64_read(&rdev->fence_drv[ring].last_seq);
 	do {
+		last_emitted = rdev->fence_drv[ring].seq;
 		seq = radeon_fence_read(rdev, ring);
 		seq |= last_seq & 0xffffffff00000000LL;
 		if (seq < last_seq) {
-			seq += 0x100000000LL;
+			seq &= 0xffffffff;
+			seq |= last_emitted & 0xffffffff00000000LL;
 		}
 
-		if (seq == last_seq) {
+		if (seq <= last_seq || seq > last_emitted) {
 			break;
 		}
 		/* If we loop over we don't want to return without