drm/amdgpu: add check_soft_reset ip func

It is used to identify if the ip block is hang.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

3 files changed, 26 insertions, 0 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 8ebc5f1eb4c0..a21b342ab2a3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1956,6 +1956,7 @@ struct amdgpu_ip_block_status {
         bool valid;
         bool sw;
         bool hw;
+        bool hang;
 };
 
 struct amdgpu_device {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index df7ab2458e50..fd7698a56f3b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -1937,6 +1937,24 @@ int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon)
         return 0;
 }
 
+static bool amdgpu_check_soft_reset(struct amdgpu_device *adev)
+{
+        int i;
+        bool asic_hang = false;
+
+        for (i = 0; i < adev->num_ip_blocks; i++) {
+                if (!adev->ip_block_status[i].valid)
+                        continue;
+                if (adev->ip_blocks[i].funcs->check_soft_reset)
+                        adev->ip_blocks[i].funcs->check_soft_reset(adev);
+                if (adev->ip_block_status[i].hang) {
+                        DRM_INFO("IP block:%d is hang!\n", i);
+                        asic_hang = true;
+                }
+        }
+        return asic_hang;
+}
+
 /**
  * amdgpu_gpu_reset - reset the asic
  *
@@ -1950,6 +1968,11 @@ int amdgpu_gpu_reset(struct amdgpu_device *adev)
         int i, r;
         int resched;
 
+        if (!amdgpu_check_soft_reset(adev)) {
+                DRM_INFO("No hardware hang detected. Did some blocks stall?\n");
+                return 0;
+        }
+
         atomic_inc(&adev->gpu_reset_counter);
 
         /* block TTM */
diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h
index a74a0d2ff1ca..fe4aa09252ab 100644
--- a/drivers/gpu/drm/amd/include/amd_shared.h
+++ b/drivers/gpu/drm/amd/include/amd_shared.h
@@ -159,6 +159,8 @@ struct amd_ip_funcs {
         bool (*is_idle)(void *handle);
         /* poll for idle */
         int (*wait_for_idle)(void *handle);
+        /* check soft reset the IP block */
+        int (*check_soft_reset)(void *handle);
         /* soft reset the IP block */
         int (*soft_reset)(void *handle);
         /* enable/disable cg for the IP block */