drm/amdkfd: Consistently apply noretry setting

Apply the same setting to SH_MEM_CONFIG and VM_CONTEXT1_CNTL. This
makes the noretry param no longer KFD-specific. On GFX10 I'm not
changing SH_MEM_CONFIG in this commit because GFX10 has different
retry behaviour in the SQ and I don't have a way to test it at the
moment.
Suggested-by: Christian König <Christian.Koenig@amd.com>
CC: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by : Shaoyun.liu < Shaoyun.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/amdgpu.h

@@ -164,6 +164,7 @@ extern int amdgpu_async_gfx_ring;
 extern int amdgpu_mcbp;
 extern int amdgpu_discovery;
 extern int amdgpu_mes;
+extern int amdgpu_noretry;
 
 #ifdef CONFIG_DRM_AMDGPU_SI
 extern int amdgpu_si_support;

drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c

@@ -142,6 +142,7 @@ int amdgpu_async_gfx_ring = 1;
 int amdgpu_mcbp = 0;
 int amdgpu_discovery = -1;
 int amdgpu_mes = 0;
+int amdgpu_noretry;
 
 struct amdgpu_mgpu_info mgpu_info = {
 	.mutex = __MUTEX_INITIALIZER(mgpu_info.mutex),
@@ -608,6 +609,10 @@ MODULE_PARM_DESC(mes,
 	"Enable Micro Engine Scheduler (0 = disabled (default), 1 = enabled)");
 module_param_named(mes, amdgpu_mes, int, 0444);
 
+MODULE_PARM_DESC(noretry,
+	"Disable retry faults (0 = retry enabled (default), 1 = retry disabled)");
+module_param_named(noretry, amdgpu_noretry, int, 0644);
+
 #ifdef CONFIG_HSA_AMD
 /**
  * DOC: sched_policy (int)
@@ -683,17 +688,6 @@ module_param(ignore_crat, int, 0444);
 MODULE_PARM_DESC(ignore_crat,
 	"Ignore CRAT table during KFD initialization (0 = use CRAT (default), 1 = ignore CRAT)");
 
-/**
- * DOC: noretry (int)
- * This parameter sets sh_mem_config.retry_disable. Default value, 0, enables retry.
- * Setting 1 disables retry.
- * Retry is needed for recoverable page faults.
- */
-int noretry;
-module_param(noretry, int, 0644);
-MODULE_PARM_DESC(noretry,
-	"Set sh_mem_config.retry_disable on Vega10 (0 = retry enabled (default), 1 = retry disabled)");
-
 /**
  * DOC: halt_if_hws_hang (int)
  * Halt if HWS hang is detected. Default value, 0, disables the halt on hang.

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

@@ -1942,11 +1942,15 @@ static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
 		if (i == 0) {
 			tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE,
 					    SH_MEM_ALIGNMENT_MODE_UNALIGNED);
+			tmp = REG_SET_FIELD(tmp, SH_MEM_CONFIG, RETRY_DISABLE,
+					    !!amdgpu_noretry);
 			WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, tmp);
 			WREG32_SOC15_RLC(GC, 0, mmSH_MEM_BASES, 0);
 		} else {
 			tmp = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE,
 					    SH_MEM_ALIGNMENT_MODE_UNALIGNED);
+			tmp = REG_SET_FIELD(tmp, SH_MEM_CONFIG, RETRY_DISABLE,
+					    !!amdgpu_noretry);
 			WREG32_SOC15_RLC(GC, 0, mmSH_MEM_CONFIG, tmp);
 			tmp = REG_SET_FIELD(0, SH_MEM_BASES, PRIVATE_BASE,
 				(adev->gmc.private_aperture_start >> 48));

drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c

@@ -236,7 +236,8 @@ static void gfxhub_v1_0_setup_vmid_config(struct amdgpu_device *adev)
 				    block_size);
 		/* Send no-retry XNACK on fault to suppress VM fault storm. */
 		tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
-				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 1);
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
 		WREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT1_CNTL, i, tmp);
 		WREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, i*2, 0);
 		WREG32_SOC15_OFFSET(GC, 0, mmVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, i*2, 0);

drivers/gpu/drm/amd/amdgpu/gfxhub_v2_0.c

@@ -215,7 +215,8 @@ static void gfxhub_v2_0_setup_vmid_config(struct amdgpu_device *adev)
 				adev->vm_manager.block_size - 9);
 		/* Send no-retry XNACK on fault to suppress VM fault storm. */
 		tmp = REG_SET_FIELD(tmp, GCVM_CONTEXT1_CNTL,
-				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
 		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_CNTL, i, tmp);
 		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, i*2, 0);
 		WREG32_SOC15_OFFSET(GC, 0, mmGCVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, i*2, 0);

drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c

@@ -265,7 +265,8 @@ static void mmhub_v1_0_setup_vmid_config(struct amdgpu_device *adev)
 				    block_size);
 		/* Send no-retry XNACK on fault to suppress VM fault storm. */
 		tmp = REG_SET_FIELD(tmp, VM_CONTEXT1_CNTL,
-				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 1);
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
 		WREG32_SOC15_OFFSET(MMHUB, 0, mmVM_CONTEXT1_CNTL, i, tmp);
 		WREG32_SOC15_OFFSET(MMHUB, 0, mmVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, i*2, 0);
 		WREG32_SOC15_OFFSET(MMHUB, 0, mmVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, i*2, 0);

drivers/gpu/drm/amd/amdgpu/mmhub_v2_0.c

@@ -205,7 +205,8 @@ static void mmhub_v2_0_setup_vmid_config(struct amdgpu_device *adev)
 				    adev->vm_manager.block_size - 9);
 		/* Send no-retry XNACK on fault to suppress VM fault storm. */
 		tmp = REG_SET_FIELD(tmp, MMVM_CONTEXT1_CNTL,
-				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT, 0);
+				    RETRY_PERMISSION_OR_INVALID_PAGE_FAULT,
+				    !amdgpu_noretry);
 		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_CNTL, i, tmp);
 		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_LO32, i*2, 0);
 		WREG32_SOC15_OFFSET(MMHUB, 0, mmMMVM_CONTEXT1_PAGE_TABLE_START_ADDR_HI32, i*2, 0);

drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c

@@ -61,7 +61,7 @@ static int update_qpd_v9(struct device_queue_manager *dqm,
 		qpd->sh_mem_config =
 				SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
 					SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT;
-		if (noretry &&
+		if (amdgpu_noretry &&
 		    !dqm->dev->device_info->needs_iommu_device)
 			qpd->sh_mem_config |=
 				1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;

drivers/gpu/drm/amd/amdkfd/kfd_priv.h

@@ -157,7 +157,7 @@ extern int ignore_crat;
 /*
  * Set sh_mem_config.retry_disable on Vega10
  */
-extern int noretry;
+extern int amdgpu_noretry;
 
 /*
  * Halt if HWS hang is detected