Skip to contentSkip to conversation

drm/amdgpu: drop duplicated ecc check for vega10

PATCH In review Guchun Chen 2 participants 3 messages Sep 29, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/amdgpu: drop duplicated ecc check for vega10

The same ECC check has been executed in amdgpu_ras_init for vega10, prior to gmc_v9_0_late_init.

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 13 +++----------
 1 file changed, 3 insertions(+), 10 deletions(-)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +3 -10
index 3dddbc60fe3d..f72acb7cfea3 100644
@@ -1193,21 +1193,14 @@ static int gmc_v9_0_late_init(void *handle)
 	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
 	if (r)
 		return r;
-	/* Check if ecc is available */
+
+	/* apply DF parity workaround */
 	if (!amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_VEGA10)) {
 		r = amdgpu_atomfirmware_mem_ecc_supported(adev);
 		if (!r) {
-			DRM_INFO("ECC is not present.\n");
 			if (adev->df.funcs->enable_ecc_force_par_wr_rmw)
 				adev->df.funcs->enable_ecc_force_par_wr_rmw(adev, false);
-		} else
-			DRM_INFO("ECC is active.\n");
-
-		r = amdgpu_atomfirmware_sram_ecc_supported(adev);
-		if (!r)
-			DRM_INFO("SRAM ECC is not present.\n");
-		else
-			DRM_INFO("SRAM ECC is active.\n");
+		}
 	}
 
 	if (adev->mmhub.funcs && adev->mmhub.funcs->reset_ras_error_count)
Signature · 6 lines
2.17.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#1 lore raw

[AMD Public Use]

Although we unified check ecc capability to ras_init, we still need to query the ecc capability here to apply the vega10 specific workaround. So we don't need to call the atomfirmware helper again, but leverage adev->ras_features to decide apply the workaround or not.

Regards,
Hawking

-----Original Message-----
From: Chen, Guchun <[email protected]>
Sent: Tuesday, September 29, 2020 12:56
To: [email protected]; Deucher, Alexander <[email protected]>; Zhang, Hawking <[email protected]>; Li, Dennis <[email protected]>; Zhou1, Tao <[email protected]>; Clements, John <[email protected]>
Cc: Chen, Guchun <[email protected]>
Subject: [PATCH] drm/amdgpu: drop duplicated ecc check for vega10

The same ECC check has been executed in amdgpu_ras_init for vega10, prior to gmc_v9_0_late_init.

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 13 +++----------
 1 file changed, 3 insertions(+), 10 deletions(-)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +3 -10
index 3dddbc60fe3d..f72acb7cfea3 100644
@@ -1193,21 +1193,14 @@ static int gmc_v9_0_late_init(void *handle)
 	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
 	if (r)
 		return r;
-	/* Check if ecc is available */
+
+	/* apply DF parity workaround */
 	if (!amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_VEGA10)) {
 		r = amdgpu_atomfirmware_mem_ecc_supported(adev);
 		if (!r) {
-			DRM_INFO("ECC is not present.\n");
 			if (adev->df.funcs->enable_ecc_force_par_wr_rmw)
 				adev->df.funcs->enable_ecc_force_par_wr_rmw(adev, false);
-		} else
-			DRM_INFO("ECC is active.\n");
-
-		r = amdgpu_atomfirmware_sram_ecc_supported(adev);
-		if (!r)
-			DRM_INFO("SRAM ECC is not present.\n");
-		else
-			DRM_INFO("SRAM ECC is active.\n");
+		}
 	}
 
 	if (adev->mmhub.funcs && adev->mmhub.funcs->reset_ras_error_count)

--
2.17.1
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

#2 lore raw

[AMD Public Use]

Let me send v2 later to correct it.

Regards,
Guchun

-----Original Message-----
From: Zhang, Hawking <[email protected]>
Sent: Tuesday, September 29, 2020 2:05 PM
To: Chen, Guchun <[email protected]>; [email protected]; Deucher, Alexander <[email protected]>; Li, Dennis <[email protected]>; Zhou1, Tao <[email protected]>; Clements, John <[email protected]>
Subject: RE: [PATCH] drm/amdgpu: drop duplicated ecc check for vega10

[AMD Public Use]

Although we unified check ecc capability to ras_init, we still need to query the ecc capability here to apply the vega10 specific workaround. So we don't need to call the atomfirmware helper again, but leverage adev->ras_features to decide apply the workaround or not.

Regards,
Hawking

-----Original Message-----
From: Chen, Guchun <[email protected]>
Sent: Tuesday, September 29, 2020 12:56
To: [email protected]; Deucher, Alexander <[email protected]>; Zhang, Hawking <[email protected]>; Li, Dennis <[email protected]>; Zhou1, Tao <[email protected]>; Clements, John <[email protected]>
Cc: Chen, Guchun <[email protected]>
Subject: [PATCH] drm/amdgpu: drop duplicated ecc check for vega10

The same ECC check has been executed in amdgpu_ras_init for vega10, prior to gmc_v9_0_late_init.

---
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 13 +++----------
 1 file changed, 3 insertions(+), 10 deletions(-)
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +3 -10
index 3dddbc60fe3d..f72acb7cfea3 100644
@@ -1193,21 +1193,14 @@ static int gmc_v9_0_late_init(void *handle)
 	r = amdgpu_gmc_allocate_vm_inv_eng(adev);
 	if (r)
 		return r;
-	/* Check if ecc is available */
+
+	/* apply DF parity workaround */
 	if (!amdgpu_sriov_vf(adev) && (adev->asic_type == CHIP_VEGA10)) {
 		r = amdgpu_atomfirmware_mem_ecc_supported(adev);
 		if (!r) {
-			DRM_INFO("ECC is not present.\n");
 			if (adev->df.funcs->enable_ecc_force_par_wr_rmw)
 				adev->df.funcs->enable_ecc_force_par_wr_rmw(adev, false);
-		} else
-			DRM_INFO("ECC is active.\n");
-
-		r = amdgpu_atomfirmware_sram_ecc_supported(adev);
-		if (!r)
-			DRM_INFO("SRAM ECC is not present.\n");
-		else
-			DRM_INFO("SRAM ECC is active.\n");
+		}
 	}
 
 	if (adev->mmhub.funcs && adev->mmhub.funcs->reset_ras_error_count)

--
2.17.1
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx