<!DOCTYPE html><html><head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> </head> <body> Hi Jesse, Please ,add to the commit message as below. The change looks good to me. Reviewed-by : Vitaly Prosyak <a class="moz-txt-link-rfc2396E" href="mailto:luben.tuikov@amd.com"><vitaly.prosyak@amd.com></a> -- Thanks, Vitaly On 2023-10-10 02:50, Jesse Zhang wrote: <blockquote type="cite" cite="mid:20231010065001.1785964-1-jesse.zhang@amd.com"> <pre class="moz-quote-pre" wrap="">Dispatch a bad program on gfx/compute ring. Check wether they will hang.</pre> </blockquote> Expose the additional parameter 'hung' to the upper layer. Dispatch existent a binary shader that waits for the register to be changed, but it does not occur and as a result, we have GPU reset and check this. Improve test description. <blockquote type="cite" cite="mid:20231010065001.1785964-1-jesse.zhang@amd.com"> <pre class="moz-quote-pre" wrap=""> V2: -add detail description and fix code style(Kamil) Cc: Vitaly Prosyak <a class="moz-txt-link-rfc2396E" href="mailto:vitaly.prosyak@amd.com"><vitaly.prosyak@amd.com></a> Cc: Luben Tuikov <a class="moz-txt-link-rfc2396E" href="mailto:luben.tuikov@amd.com"><luben.tuikov@amd.com></a> Cc: Alex Deucher <a class="moz-txt-link-rfc2396E" href="mailto:alexander.deucher@amd.com"><alexander.deucher@amd.com></a> Cc: Christian Koenig <a class="moz-txt-link-rfc2396E" href="mailto:christian.koenig@amd.com"><christian.koenig@amd.com></a> Cc: Kamil Konieczny <a class="moz-txt-link-rfc2396E" href="mailto:kamil.konieczny@linux.intel.com"><kamil.konieczny@linux.intel.com></a> --- lib/amdgpu/amd_dispatch.c | 4 ++-- lib/amdgpu/amd_dispatch.h | 2 +- tests/amdgpu/amd_dispatch.c | 38 ++++++++++++++++++++++++++++++++----- 3 files changed, 36 insertions(+), 8 deletions(-) diff --git a/lib/amdgpu/amd_dispatch.c b/lib/amdgpu/amd_dispatch.c index 9de3986ba..040381a11 100644 --- a/lib/amdgpu/amd_dispatch.c +++ b/lib/amdgpu/amd_dispatch.c @@ -524,7 +524,7 @@ amdgpu_dispatch_hang_slow_helper(amdgpu_device_handle device_handle, } } -void amdgpu_gfx_dispatch_test(amdgpu_device_handle device_handle, uint32_t ip_type) +void amdgpu_gfx_dispatch_test(amdgpu_device_handle device_handle, uint32_t ip_type, int hang) { int r; struct drm_amdgpu_info_hw_ip info; @@ -547,7 +547,7 @@ void amdgpu_gfx_dispatch_test(amdgpu_device_handle device_handle, uint32_t ip_ty amdgpu_memset_dispatch_test(device_handle, ip_type, ring_id, version); amdgpu_memcpy_dispatch_test(device_handle, ip_type, ring_id, - version, 0); + version, hang); } } diff --git a/lib/amdgpu/amd_dispatch.h b/lib/amdgpu/amd_dispatch.h index c500b5b3f..4df8b1355 100644 --- a/lib/amdgpu/amd_dispatch.h +++ b/lib/amdgpu/amd_dispatch.h @@ -27,7 +27,7 @@ #include <amdgpu.h> void amdgpu_gfx_dispatch_test(amdgpu_device_handle device_handle, - uint32_t ip_type); + uint32_t ip_type, int hang); void amdgpu_memcpy_dispatch_test(amdgpu_device_handle device_handle, uint32_t ip_type, diff --git a/tests/amdgpu/amd_dispatch.c b/tests/amdgpu/amd_dispatch.c index 769f26cd0..323284306 100644 --- a/tests/amdgpu/amd_dispatch.c +++ b/tests/amdgpu/amd_dispatch.c @@ -24,6 +24,18 @@ amdgpu_dispatch_hang_slow_compute(amdgpu_device_handle device_handle) amdgpu_dispatch_hang_slow_helper(device_handle, AMDGPU_HW_IP_COMPUTE); } +static void +amdgpu_dispatch_hang_gfx(amdgpu_device_handle device_handle) +{ + amdgpu_gfx_dispatch_test(device_handle, AMDGPU_HW_IP_GFX, 1); +} + +static void +amdgpu_dispatch_hang_compute(amdgpu_device_handle device_handle) +{ + amdgpu_gfx_dispatch_test(device_handle, AMDGPU_HW_IP_COMPUTE, 1); +} + static void amdgpu_gpu_reset_test(amdgpu_device_handle device_handle, int drm_amdgpu) { @@ -54,8 +66,8 @@ amdgpu_gpu_reset_test(amdgpu_device_handle device_handle, int drm_amdgpu) r = amdgpu_cs_ctx_free(context_handle); igt_assert_eq(r, 0); - amdgpu_gfx_dispatch_test(device_handle, AMDGPU_HW_IP_GFX); - amdgpu_gfx_dispatch_test(device_handle, AMDGPU_HW_IP_COMPUTE); + amdgpu_gfx_dispatch_test(device_handle, AMDGPU_HW_IP_GFX, 0); + amdgpu_gfx_dispatch_test(device_handle, AMDGPU_HW_IP_COMPUTE, 0); } igt_main @@ -85,7 +97,7 @@ igt_main asic_rings_readness(device, 1, arr_cap); } - igt_describe("Test-GPU-reset-using-a-binary-shader-to-hang-the-job-on-compute-ring"); + igt_describe("Test GPU reset using a binary shader to slow hang the job on compute ring"); igt_subtest_with_dynamic("amdgpu-dispatch-test-compute-with-IP-COMPUTE") { if (arr_cap[AMD_IP_COMPUTE]) { igt_dynamic_f("amdgpu-dispatch-test-compute") @@ -93,7 +105,7 @@ igt_main } } - igt_describe("Test-GPU-reset-using-a-binary-shader-to-hang-the-job-on-gfx-ring"); + igt_describe("Test GPU reset using a binary shader to slow hang the job on gfx ring"); igt_subtest_with_dynamic("amdgpu-dispatch-test-gfx-with-IP-GFX") { if (arr_cap[AMD_IP_GFX]) { igt_dynamic_f("amdgpu-dispatch-test-gfx") @@ -101,7 +113,23 @@ igt_main } } - igt_describe("Test-GPU-reset-using-amdgpu-debugfs-to-hang-the-job-on-gfx-ring"); + igt_describe("Test GPU reset using a binary shader to hang the job on gfx ring"); + igt_subtest_with_dynamic("amdgpu-dispatch-hang-test-gfx-with-IP-GFX") { + if (arr_cap[AMD_IP_GFX]) { + igt_dynamic_f("amdgpu-dispatch-hang-test-gfx") + amdgpu_dispatch_hang_gfx(device); + } + } + + igt_describe("Test GPU reset using a binary shader to hang the job on compute ring"); + igt_subtest_with_dynamic("amdgpu-dispatch-hang-test-compute-with-IP-COMPUTE") { + if (arr_cap[AMD_IP_COMPUTE]) { + igt_dynamic_f("amdgpu-dispatch-hang-test-compute") + amdgpu_dispatch_hang_compute(device); + } + } + + igt_describe("Test GPU reset using amdgpu debugfs to hang the job on gfx ring"); igt_subtest_with_dynamic("amdgpu-reset-test-gfx-with-IP-GFX-and-COMPUTE") { if (arr_cap[AMD_IP_GFX] && arr_cap[AMD_IP_COMPUTE]) { igt_dynamic_f("amdgpu-reset-gfx-compute") </pre> </blockquote> </body> </html>