[PATCH 11/13] drm/amdgpu: Report vm fault status HI32 when avaialble

Alex Deucher <[email protected]>
Newsgroups org.freedesktop.lists.amd-gfx
Message-ID <[email protected]> 
From: Hawking Zhang <[email protected]>

VM fault status is 64 bits on gfxhub v12_1 and
mmhub v4_2, with fields such as FED located in the
upper status register.

Add an optional HI32 fault status callback to help
identify FED status on gfxhub v12_1 and mmhub v4_2

Signed-off-by: Hawking Zhang <[email protected]>
Reviewed-by: Horatio Zhang <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h   |  3 +++
 drivers/gpu/drm/amd/amdgpu/gfxhub_v12_1.c | 20 +++++++++++++++++---
 drivers/gpu/drm/amd/amdgpu/gmc_v12_1.c    | 10 ++++++++--
 drivers/gpu/drm/amd/amdgpu/mmhub_v4_2_0.c | 20 +++++++++++++++++---
 4 files changed, 45 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
index c0884797dd544..cffd5ca3fc66b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
@@ -114,6 +114,8 @@ struct amdgpu_gmc_fault {
 struct amdgpu_vmhub_funcs {
 	void (*print_l2_protection_fault_status)(struct amdgpu_device *adev,
 						 uint32_t status);
+	void (*print_l2_protection_fault_status_hi)(struct amdgpu_device *adev,
+						    uint32_t status);
 	uint32_t (*get_invalidate_req)(unsigned int vmid, uint32_t flush_type);
 };
 
@@ -125,6 +127,7 @@ struct amdgpu_vmhub {
 	uint32_t	vm_inv_eng0_ack;
 	uint32_t	vm_context0_cntl;
 	uint32_t	vm_l2_pro_fault_status;
+	uint32_t	vm_l2_pro_fault_status_hi;
 	uint32_t	vm_l2_pro_fault_cntl;
 
 	/*
diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_1.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_1.c
index 0d57c106f9d1c..1f7e57792d4dc 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v12_1.c
@@ -702,8 +702,18 @@ static const char *gfxhub_v12_1_client_ids[] = {
 	"PA"
 };
 
-/*TODO: l2 protection fault status is increased to 64bits.
- * some critical fields like FED are moved to STATUS_HI32 */
+static void
+gfxhub_v12_1_print_l2_protection_fault_status_hi(struct amdgpu_device *adev,
+						 uint32_t status)
+{
+	dev_err(adev->dev,
+		"GCVM_L2_PROTECTION_FAULT_STATUS_HI32:0x%08X\n",
+		status);
+	dev_err(adev->dev, "\t FED: 0x%lx\n",
+		REG_GET_FIELD(status,
+			      GCVM_L2_PROTECTION_FAULT_STATUS_HI32, FED));
+}
+
 static void gfxhub_v12_1_print_l2_protection_fault_status(struct amdgpu_device *adev,
 							  uint32_t status)
 {
@@ -736,6 +746,8 @@ static void gfxhub_v12_1_print_l2_protection_fault_status(struct amdgpu_device *
 
 static const struct amdgpu_vmhub_funcs gfxhub_v12_1_vmhub_funcs = {
 	.print_l2_protection_fault_status = gfxhub_v12_1_print_l2_protection_fault_status,
+	.print_l2_protection_fault_status_hi =
+		gfxhub_v12_1_print_l2_protection_fault_status_hi,
 	.get_invalidate_req = gfxhub_v12_1_get_invalidate_req,
 };
 
@@ -765,10 +777,12 @@ static void gfxhub_v12_1_xcc_init(struct amdgpu_device *adev, uint32_t xcc_mask)
 		hub->vm_context0_cntl =
 			SOC15_REG_OFFSET(GC, GET_INST(GC, i),
 				regGCVM_CONTEXT0_CNTL);
-		/* TODO: add a new member to accomandate additional fault status/cntl reg */
 		hub->vm_l2_pro_fault_status =
 			SOC15_REG_OFFSET(GC, GET_INST(GC, i),
 				regGCVM_L2_PROTECTION_FAULT_STATUS_LO32);
+		hub->vm_l2_pro_fault_status_hi =
+			SOC15_REG_OFFSET(GC, GET_INST(GC, i),
+					 regGCVM_L2_PROTECTION_FAULT_STATUS_HI32);
 		hub->vm_l2_pro_fault_cntl =
 			SOC15_REG_OFFSET(GC, GET_INST(GC, i),
 				regGCVM_L2_PROTECTION_FAULT_CNTL_LO32);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v12_1.c b/drivers/gpu/drm/amd/amdgpu/gmc_v12_1.c
index e8e35d1b22765..b0962a14fdb4d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v12_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v12_1.c
@@ -128,7 +128,7 @@ static int gmc_v12_1_process_interrupt(struct amdgpu_device *adev,
 	uint32_t cam_index = 0;
 	const char *hub_name;
 	int ret, xcc_id = 0;
-	uint32_t status = 0;
+	uint32_t status = 0, status_hi = 0;
 	const char *die_name;
 	char die_name_buf[32];
 	u64 addr;
@@ -243,11 +243,14 @@ static int gmc_v12_1_process_interrupt(struct amdgpu_device *adev,
 		RREG32(hub->vm_l2_pro_fault_status);
 
 	status = RREG32(hub->vm_l2_pro_fault_status);
+	if (hub->vmhub_funcs &&
+	    hub->vmhub_funcs->print_l2_protection_fault_status_hi)
+		status_hi = RREG32(hub->vm_l2_pro_fault_status_hi);
 
 	/* Only print L2 fault status if the status register could be read and
 	 * contains useful information
 	 */
-	if (!status)
+	if (!status && !status_hi)
 		return 0;
 
 	WREG32_P(hub->vm_l2_pro_fault_cntl, 1, ~1);
@@ -260,6 +263,9 @@ static int gmc_v12_1_process_interrupt(struct amdgpu_device *adev,
 	}
 
 	hub->vmhub_funcs->print_l2_protection_fault_status(adev, status);
+	if (hub->vmhub_funcs->print_l2_protection_fault_status_hi)
+		hub->vmhub_funcs->print_l2_protection_fault_status_hi(adev,
+								     status_hi);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v4_2_0.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v4_2_0.c
index 8bed4addba426..b2a3276750deb 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v4_2_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v4_2_0.c
@@ -732,8 +732,18 @@ static uint32_t mmhub_v4_2_0_get_invalidate_req(unsigned int vmid,
 	return req;
 }
 
-/*TODO: l2 protection fault status is increased to 64bits.
- * some critical fields like FED are moved to STATUS_HI32 */
+static void
+mmhub_v4_2_0_print_l2_protection_fault_status_hi(struct amdgpu_device *adev,
+						 uint32_t status)
+{
+	dev_err(adev->dev,
+		"MMVM_L2_PROTECTION_FAULT_STATUS_HI32:0x%08X\n",
+		status);
+	dev_err(adev->dev, "\t FED: 0x%lx\n",
+		REG_GET_FIELD(status,
+			      MMVM_L2_PROTECTION_FAULT_STATUS_HI32, FED));
+}
+
 static void
 mmhub_v4_2_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
 					      uint32_t status)
@@ -770,6 +780,8 @@ mmhub_v4_2_0_print_l2_protection_fault_status(struct amdgpu_device *adev,
 
 static const struct amdgpu_vmhub_funcs mmhub_v4_2_0_vmhub_funcs = {
 	.print_l2_protection_fault_status = mmhub_v4_2_0_print_l2_protection_fault_status,
+	.print_l2_protection_fault_status_hi =
+		mmhub_v4_2_0_print_l2_protection_fault_status_hi,
 	.get_invalidate_req = mmhub_v4_2_0_get_invalidate_req,
 };
 
@@ -800,10 +812,12 @@ static void mmhub_v4_2_0_mid_init(struct amdgpu_device *adev,
 		hub->vm_context0_cntl =
 			SOC15_REG_OFFSET(MMHUB, GET_INST(MMHUB, i),
 					 regMMVM_CONTEXT0_CNTL);
-		/* TODO: add a new member to accomandate additional fault status/cntl reg */
 		hub->vm_l2_pro_fault_status =
 			SOC15_REG_OFFSET(MMHUB, GET_INST(MMHUB, i),
 					 regMMVM_L2_PROTECTION_FAULT_STATUS_LO32);
+		hub->vm_l2_pro_fault_status_hi =
+			SOC15_REG_OFFSET(MMHUB, GET_INST(MMHUB, i),
+					 regMMVM_L2_PROTECTION_FAULT_STATUS_HI32);
 		hub->vm_l2_pro_fault_cntl =
 			SOC15_REG_OFFSET(MMHUB, GET_INST(MMHUB, i),
 					 regMMVM_L2_PROTECTION_FAULT_CNTL_LO32);
-- 
2.55.0
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.