On Thu, May 13, 2021 at 1:32 AM Luben Tuikov luben.tuikov@amd.com wrote:
On QUERY2 IOCTL don't query counts of correctable and uncorrectable errors, since when RAS is enabled and supported on Vega20 server boards, this takes insurmountably long time, in O(n^3), which slows the system down to the point of it being unusable when we have GUI up.
Fixes: ae363a212b14 ("drm/amdgpu: Add a new flag to AMDGPU_CTX_OP_QUERY_STATE2") Cc: Alexander Deucher Alexander.Deucher@amd.com Cc: stable@vger.kernel.org Signed-off-by: Luben Tuikov luben.tuikov@amd.com
Reviewed-by: Alex Deucher alexander.deucher@amd.com
drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 16 ---------------- 1 file changed, 16 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c index 01fe60fedcbe..e1557020c49d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c @@ -337,7 +337,6 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev, { struct amdgpu_ctx *ctx; struct amdgpu_ctx_mgr *mgr;
unsigned long ras_counter; if (!fpriv) return -EINVAL;@@ -362,21 +361,6 @@ static int amdgpu_ctx_query2(struct amdgpu_device *adev, if (atomic_read(&ctx->guilty)) out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_GUILTY;
/*query ue count*/ras_counter = amdgpu_ras_query_error_count(adev, false);/*ras counter is monotonic increasing*/if (ras_counter != ctx->ras_counter_ue) {out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RAS_UE;ctx->ras_counter_ue = ras_counter;}/*query ce count*/ras_counter = amdgpu_ras_query_error_count(adev, true);if (ras_counter != ctx->ras_counter_ce) {out->state.flags |= AMDGPU_CTX_QUERY2_FLAGS_RAS_CE;ctx->ras_counter_ce = ras_counter;}mutex_unlock(&mgr->lock); return 0;}
2.31.1.527.g2d677e5b15
amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx