| Message ID | 1424410412-24910-2-git-send-email-benjamin.widawsky@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
At first glance, this is interesting. On reflection I think I still only care about the register values at the time of the error. Otherwise, I would rather have the last few registers written - and only those of the interesting set as defined by the error state - their values and when. We would also then require the timestamp of batch execution to marry with the register updates if we suspect them of causing GPU hangs. -Chris
Tested-By: PRC QA PRTS (Patch Regression Test System Contact: shuang.he@intel.com)
Task id: 5799
-------------------------------------Summary-------------------------------------
Platform Delta drm-intel-nightly Series Applied
PNV -2 277/277 275/277
ILK 313/313 313/313
SNB -1 309/309 308/309
IVB 382/382 382/382
BYT 296/296 296/296
HSW -1 425/425 424/425
BDW -1 318/318 317/318
-------------------------------------Detailed-------------------------------------
Platform Test drm-intel-nightly Series Applied
PNV igt_gem_userptr_blits_coherency-sync NO_RESULT(1)CRASH(5)NRUN(1)PASS(6) CRASH(2)
PNV igt_gem_userptr_blits_coherency-unsync CRASH(3)NRUN(1)PASS(4) CRASH(2)
*SNB igt_kms_plane_plane-panning-top-left-pipe-B-plane-2 PASS(2) TIMEOUT(1)PASS(1)
*HSW igt_gem_storedw_batches_loop_secure-dispatch PASS(2) DMESG_WARN(1)PASS(1)
*BDW igt_gem_gtt_hog PASS(17) DMESG_WARN(1)PASS(1)
Note: You need to pay more attention to line start with '*'
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 6fa22db..49bc296 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -323,6 +323,7 @@ struct drm_i915_error_state { struct timeval time; char error_msg[128]; + u32 last_written[NR_CPUS]; u32 reset_count; u32 suspend_count; diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 2c87a79..f7737c5 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -356,6 +356,9 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m, error->ring[i].pid); } } + for_each_possible_cpu(i) + err_printf(m, "Last written register (cpu=%d): 0x%08x\n", + i, error->last_written[i]); err_printf(m, "Reset count: %u\n", error->reset_count); err_printf(m, "Suspend count: %u\n", error->suspend_count); err_printf(m, "PCI ID: 0x%04x\n", dev->pdev->device); @@ -1253,6 +1256,13 @@ static void i915_error_capture_msg(struct drm_device *dev, wedged ? "reset" : "continue"); } +static void i915_capture_cpu_state(struct drm_i915_error_state *error) +{ + int cpu; + for_each_possible_cpu(cpu) + error->last_written[cpu] = per_cpu(i915_last_written, cpu); +} + static void i915_capture_gen_state(struct drm_i915_private *dev_priv, struct drm_i915_error_state *error) { @@ -1286,6 +1296,7 @@ void i915_capture_error_state(struct drm_device *dev, bool wedged, kref_init(&error->ref); + i915_capture_cpu_state(error); i915_capture_gen_state(dev_priv, error); i915_capture_reg_state(dev_priv, error); i915_gem_capture_buffers(dev_priv, error);
Signed-off-by: Ben Widawsky <ben@bwidawsk.net> --- drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/i915_gpu_error.c | 11 +++++++++++ 2 files changed, 12 insertions(+)