drm/etnaviv: bring back progress check in job timeout handler

When the hangcheck handler was replaced by the DRM scheduler timeout
handling we dropped the forward progress check, as this might allow
clients to hog the GPU for a long time with a big job.

It turns out that even reasonably well behaved clients like the
Armada Xorg driver occasionally trip over the 500ms timeout. Bring
back the forward progress check to get rid of the userspace regression.

We would still like to fix userspace to submit smaller batches
if possible, but that is for another day.

Cc: <stable@vger.kernel.org>
Fixes: 6d7a20c07760 (drm/etnaviv: replace hangcheck with scheduler timeout)
Reported-by: Russell King <linux@armlinux.org.uk>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Eric Anholt <eric@anholt.net>

diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.h b/drivers/gpu/drm/etnaviv/etnaviv_gpu.h
index dd430f0f8ff5158975e21f26ad97fc3aa5bae2cb..90f17ff7888e7042b85f38e5ef114f77fa42dccd 100644 (file)
--- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.h
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.h
@@ -131,6 +131,9 @@ struct etnaviv_gpu {
        struct work_struct sync_point_work;
        int sync_point_event;
 
+       /* hang detection */
+       u32 hangcheck_dma_addr;
+
        void __iomem *mmio;
        int irq;
 

diff --git a/drivers/gpu/drm/etnaviv/etnaviv_sched.c b/drivers/gpu/drm/etnaviv/etnaviv_sched.c
index a74eb57af15bc65ba2ff4a2ed3906da29afe959b..50d6b88cb7aab3f99025829e1f544aae957d6d7f 100644 (file)
--- a/drivers/gpu/drm/etnaviv/etnaviv_sched.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_sched.c
@@ -10,6 +10,7 @@
 #include "etnaviv_gem.h"
 #include "etnaviv_gpu.h"
 #include "etnaviv_sched.h"
+#include "state.xml.h"
 
 static int etnaviv_job_hang_limit = 0;
 module_param_named(job_hang_limit, etnaviv_job_hang_limit, int , 0444);
@@ -85,6 +86,29 @@ static void etnaviv_sched_timedout_job(struct drm_sched_job *sched_job)
 {
        struct etnaviv_gem_submit *submit = to_etnaviv_submit(sched_job);
        struct etnaviv_gpu *gpu = submit->gpu;
+       u32 dma_addr;
+       int change;
+
+       /*
+        * If the GPU managed to complete this jobs fence, the timout is
+        * spurious. Bail out.
+        */
+       if (fence_completed(gpu, submit->out_fence->seqno))
+               return;
+
+       /*
+        * If the GPU is still making forward progress on the front-end (which
+        * should never loop) we shift out the timeout to give it a chance to
+        * finish the job.
+        */
+       dma_addr = gpu_read(gpu, VIVS_FE_DMA_ADDRESS);
+       change = dma_addr - gpu->hangcheck_dma_addr;
+       if (change < 0 || change > 16) {
+               gpu->hangcheck_dma_addr = dma_addr;
+               schedule_delayed_work(&sched_job->work_tdr,
+                                     sched_job->sched->timeout);
+               return;
+       }
 
        /* block scheduler */
        kthread_park(gpu->sched.thread);