Skip to contentSkip to conversation

drm/amdgpu: default noretry=0 for navi1x and newer

PATCH In review Alex Deucher 2 participants 3 messages Nov 30, 2020

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

PATCH 1/1 drm/amdgpu: default noretry=0 for navi1x and newer

There are no performance advantages to setting it to 1 and it causes stability issues in some cases.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +5 -5
index fede84509dbc..02b2ee9b8939 100644
@@ -422,11 +422,6 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
 
 	switch (adev->asic_type) {
 	case CHIP_VEGA20:
-	case CHIP_NAVI10:
-	case CHIP_NAVI14:
-	case CHIP_SIENNA_CICHLID:
-	case CHIP_NAVY_FLOUNDER:
-	case CHIP_DIMGREY_CAVEFISH:
 		/*
 		 * noretry = 0 will cause kfd page fault tests fail
 		 * for some ASICs, so set default to 1 for these ASICs.
@@ -437,6 +432,11 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
 			gmc->noretry = amdgpu_noretry;
 		break;
 	case CHIP_RAVEN:
+	case CHIP_NAVI10:
+	case CHIP_NAVI14:
+	case CHIP_SIENNA_CICHLID:
+	case CHIP_NAVY_FLOUNDER:
+	case CHIP_DIMGREY_CAVEFISH:
 	default:
 		/* Raven currently has issues with noretry
 		 * regardless of what we decide for other
Signature · 6 lines
2.25.4

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
Acked-by on v1 1/1
#1 lore raw

Am 2020-11-30 um 1:58 p.m. schrieb Alex Deucher:

Quoted 26 lines

There are no performance advantages to setting it to 1 and it causes stability issues in some cases.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +0 -5
index fede84509dbc..02b2ee9b8939 100644
@@ -422,11 +422,6 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
 
 	switch (adev->asic_type) {
 	case CHIP_VEGA20:
-	case CHIP_NAVI10:
-	case CHIP_NAVI14:
-	case CHIP_SIENNA_CICHLID:
-	case CHIP_NAVY_FLOUNDER:
-	case CHIP_DIMGREY_CAVEFISH:
 		/*
 		 * noretry = 0 will cause kfd page fault tests fail
 		 * for some ASICs, so set default to 1 for these ASICs.

drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c void amdgpu_gmc_noretry_set(struct amdgpu_device *adev) · line 439

Quoted 8 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +5 -0
@@ -437,6 +432,11 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
 			gmc->noretry = amdgpu_noretry;
 		break;
 	case CHIP_RAVEN:
+	case CHIP_NAVI10:
+	case CHIP_NAVI14:
+	case CHIP_SIENNA_CICHLID:
+	case CHIP_NAVY_FLOUNDER:
+	case CHIP_DIMGREY_CAVEFISH:

Is there any point listing those chips here explicitly, given that it's the default case? Anyway, the patch is

drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c void amdgpu_gmc_noretry_set(struct amdgpu_device *adev) · line 442

Quoted 3 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +0 -0
 	default:
 		/* Raven currently has issues with noretry
 		 * regardless of what we decide for other

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

#2 lore raw
Quoted 31 lines · Felix Kuehling

Am 2020-11-30 um 1:58 p.m. schrieb Alex Deucher:

Quoted 29 lines

There are no performance advantages to setting it to 1 and it causes stability issues in some cases.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +0 -5
index fede84509dbc..02b2ee9b8939 100644
@@ -422,11 +422,6 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
 
      switch (adev->asic_type) {
      case CHIP_VEGA20:
-     case CHIP_NAVI10:
-     case CHIP_NAVI14:
-     case CHIP_SIENNA_CICHLID:
-     case CHIP_NAVY_FLOUNDER:
-     case CHIP_DIMGREY_CAVEFISH:
              /*
               * noretry = 0 will cause kfd page fault tests fail
               * for some ASICs, so set default to 1 for these ASICs.
@@ -437,6 +432,11 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
                      gmc->noretry = amdgpu_noretry;
              break;
      case CHIP_RAVEN:
Quoted 8 lines

drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c void amdgpu_gmc_noretry_set(struct amdgpu_device *adev) · line 439

Quoted 5 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +5 -0
+     case CHIP_NAVI10:
+     case CHIP_NAVI14:
+     case CHIP_SIENNA_CICHLID:
+     case CHIP_NAVY_FLOUNDER:
+     case CHIP_DIMGREY_CAVEFISH:

Is there any point listing those chips here explicitly, given that it's the default case? Anyway, the patch is

Just noticed that myself and sent a v2.

Alex

Quoted 11 lines

drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c void amdgpu_gmc_noretry_set(struct amdgpu_device *adev) · line 442

Quoted 3 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +0 -0
      default:
              /* Raven currently has issues with noretry
               * regardless of what we decide for other

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

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