Lower the severity of the radeon lockup messages.
Subsystems (2)
abrt files a lot of bug reports when users get GPU lockups, but there's not really enough context to do anything useful with them. Given the lack of GPU context being dumped, this patch removes the stack trace, so that abrt ignores the messages.
drivers/gpu/drm/radeon/radeon_fence.c +1 -1
@@ -263,7 +263,7 @@ retry:*/ if (seq == rdev->fence_drv.last_seq &&radeon_gpu_is_lockup (rdev)) {/* good news we believe it's a lockup */ -WARN (1 ,"GPU lockup (waiting for 0x%08X last fence id 0x%08X)\n" , +printk (KERN_WARNING "GPU lockup (waiting for 0x%08X last fence id 0x%08X)\n" , fence->seq, seq);/* FIXME: what should we do ? marking everyone * as signaled for now
Quoted 11 lines · Dave Jones
abrt files a lot of bug reports when users get GPU lockups, but there's not really enough context to do anything useful with them. Given the lack of GPU context being dumped, this patch removes the stack trace, so that abrt ignores the messages.
Signed-off-by Dave Jones<[email protected]>drivers/gpu/drm/radeon/radeon_fence.c +0
-0
drivers/gpu/drm/radeon/radeon_fence.c retry: · line 269
Quoted 8 lines
drivers/gpu/drm/radeon/radeon_fence.c +1
-1@@ -263,7 +263,7 @@ retry:*/ if (seq == rdev->fence_drv.last_seq&&radeon_gpu_is_lockup (rdev)) {/* good news we believe it's a lockup */ -WARN (1 ,"GPU lockup (waiting for 0x%08X last fence id 0x%08X)\n" , +printk (KERN_WARNING "GPU lockup (waiting for 0x%08X last fence id 0x%08X)\n" , fence->seq, seq);/* FIXME: what should we do ? marking everyone * as signaled for now
At the risk of sounding obvious, wouldn't it be better to enhance the driver to dump out more useful debugging info in this case rather than just suppressing the ABRT report?
Quoted 2 lines
At the risk of sounding obvious, wouldn't it be better to enhance the driver to dump out more useful debugging info in this case rather than just suppressing the ABRT report?
In theory yes, in practice that is non-trivial. The amount of data you need to keep track off is in the MBs, it can range from any previous GPU command stream sent to the GPU, to the contents of any of the vertex buffers etc.
the problem with GPU lockups is you don't notice them until after you've queued up a tonne more stuff since they are such deep pipelines.
Dave.