Skip to contentSkip to conversation

Add boot time error reporting

PATCH In review Hawking Zhang 5 participants 14 messages Jan 2, 2024 – Jan 3, 2024

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

cover Add boot time error reporting

For ASICs that support boot time error reporting, poll all the boot time errors cached in registers and make it available in kernel log.

Hawking Zhang (5):
drm/amdgpu: drop psp v13 query_boot_status implementation drm/amdgpu: Init pcie_index/data address as fallback drm/amdgpu: Add ras helper to query boot errors drm/amdgpu: Query boot status if discovery failed drm/amdgpu: Query boot status if boot failed

 drivers/gpu/drm/amd/amdgpu/amdgpu.h           |  3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c    | 22 +++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c |  6 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c       | 15 ---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h       |  4 -
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c       | 95 +++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h       | 15 ++-
 drivers/gpu/drm/amd/amdgpu/psp_v13_0.c        | 89 ++---------------
 8 files changed, 141 insertions(+), 108 deletions(-)
Signature · 1 line
2.17.1
#1 lore raw

PATCH 1/5 drm/amdgpu: drop psp v13 query_boot_status implementation

Will replace it with new implementation to cover
boot fails in ip discovery phase.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c |  2 -
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c    | 15 -----
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h    |  4 --
 drivers/gpu/drm/amd/amdgpu/psp_v13_0.c     | 78 ----------------------
 4 files changed, 99 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +0 -2
index 4b694696930e..001a35fa0f19 100644
@@ -1218,8 +1218,6 @@ static int amdgpu_device_asic_init(struct amdgpu_device *adev)
 	    amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0)) {
 		amdgpu_psp_wait_for_bootloader(adev);
 		ret = amdgpu_atomfirmware_asic_init(adev, true);
-		/* TODO: check the return val and stop device initialization if boot fails */
-		amdgpu_psp_query_boot_status(adev);
 		return ret;
 	} else {
 		return amdgpu_atom_asic_init(adev->mode_info.atom_context);
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +0 -15
index 1bf975b8d083..94b536e3cada 100644
@@ -2125,21 +2125,6 @@ int amdgpu_psp_wait_for_bootloader(struct amdgpu_device *adev)
 	return ret;
 }
 
-int amdgpu_psp_query_boot_status(struct amdgpu_device *adev)
-{
-	struct psp_context *psp = &adev->psp;
-	int ret = 0;
-
-	if (amdgpu_sriov_vf(adev) || (adev->flags & AMD_IS_APU))
-		return 0;
-
-	if (psp->funcs &&
-	    psp->funcs->query_boot_status)
-		ret = psp->funcs->query_boot_status(psp);
-
-	return ret;
-}
-
 static int psp_hw_start(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +0 -4
index c4d9cbde55b9..09d1f8f72a9c 100644
@@ -134,7 +134,6 @@ struct psp_funcs {
 	int (*update_spirom)(struct psp_context *psp, uint64_t fw_pri_mc_addr);
 	int (*vbflash_stat)(struct psp_context *psp);
 	int (*fatal_error_recovery_quirk)(struct psp_context *psp);
-	int (*query_boot_status)(struct psp_context *psp);
 };
 
 struct ta_funcs {
@@ -538,7 +537,4 @@ int psp_spatial_partition(struct psp_context *psp, int mode);
 int is_psp_fw_valid(struct psp_bin_desc bin);
 
 int amdgpu_psp_wait_for_bootloader(struct amdgpu_device *adev);
-
-int amdgpu_psp_query_boot_status(struct amdgpu_device *adev);
-
 #endif
drivers/gpu/drm/amd/amdgpu/psp_v13_0.c +0 -78

90 lines not on this page · raw message

Signature · 1 line
2.17.1
#2 lore raw

PATCH 2/5 drm/amdgpu: Init pcie_index/data address as fallback

In reply to Hawking Zhang

To allow using this helper for indirect access when nbio funcs is not available. For instance, in ip discovery phase.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +15 -5
index 001a35fa0f19..873419a5b9aa 100644
@@ -781,12 +781,22 @@ u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev,
 	void __iomem *pcie_index_hi_offset;
 	void __iomem *pcie_data_offset;
 
-	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
-	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
-	if ((reg_addr >> 32) && (adev->nbio.funcs->get_pcie_index_hi_offset))
-		pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
-	else
+	if (unlikely(!adev->nbio.funcs)) {
+		pcie_index = (0x38 >> 2);
+		pcie_data = (0x3C >> 2);
+	} else {
+		pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+		pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+	}
+
+	if (reg_addr >> 32) {
+		if (unlikely(!adev->nbio.funcs))
+			pcie_index_hi = (0x44 >> 2);
+		else
+			pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
+	} else {
 		pcie_index_hi = 0;
+	}
 
 	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
 	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
Signature · 1 line
2.17.1
#3 lore raw

Re: [PATCH 2/5] drm/amdgpu: Init pcie_index/data address as fallback

Quoted 24 lines · Hawking Zhang

To allow using this helper for indirect access when nbio funcs is not available. For instance, in ip discovery phase.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +1 -5
index 001a35fa0f19..873419a5b9aa 100644
@@ -781,12 +781,22 @@ u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev,
        void __iomem *pcie_index_hi_offset;
        void __iomem *pcie_data_offset;
 
-       pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
-       pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
-       if ((reg_addr >> 32) && (adev->nbio.funcs->get_pcie_index_hi_offset))
-               pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
-       else
+       if (unlikely(!adev->nbio.funcs)) {

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev, · line 794

Quoted 10 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +10 -0
+               pcie_index = (0x38 >> 2);
+               pcie_data = (0x3C >> 2);
+       } else {
+               pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+               pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+       }
+
+       if (reg_addr >> 32) {
+               if (unlikely(!adev->nbio.funcs))
+                       pcie_index_hi = (0x44 >> 2);

I'd still use a define for these, E.g.,

#define AMDGPU_PCIE_INDEX_FALLBACK (0x38 >> 2)
etc.
Or something similar.

Alex

Quoted 11 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +4 -1
+               else
+                       pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
+       } else {
                pcie_index_hi = 0;
+       }
 
        spin_lock_irqsave(&adev->pcie_idx_lock, flags);
        pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
--

2.17.1

#4 lore raw

RE: [PATCH 2/5] drm/amdgpu: Init pcie_index/data address as fallback

[AMD Official Use Only - General]

Sure, will do in v2

Regards,
Hawking

-----Original Message-----
From: Alex Deucher <[email protected]>
Sent: Wednesday, January 3, 2024 03:44
To: Zhang, Hawking <[email protected]>
Cc: [email protected]; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Wang, Yang(Kevin) <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>; Deucher, Alexander <[email protected]>; Ma, Le <[email protected]>; Lazar, Lijo <[email protected]>
Subject: Re: [PATCH 2/5] drm/amdgpu: Init pcie_index/data address as fallback

Quoted 24 lines · Hawking Zhang

To allow using this helper for indirect access when nbio funcs is not available. For instance, in ip discovery phase.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +1 -5
index 001a35fa0f19..873419a5b9aa 100644
@@ -781,12 +781,22 @@ u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev,
        void __iomem *pcie_index_hi_offset;
        void __iomem *pcie_data_offset;
 
-       pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
-       pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
-       if ((reg_addr >> 32) && (adev->nbio.funcs->get_pcie_index_hi_offset))
-               pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
-       else
+       if (unlikely(!adev->nbio.funcs)) {

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev, · line 794

Quoted 10 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +10 -0
+               pcie_index = (0x38 >> 2);
+               pcie_data = (0x3C >> 2);
+       } else {
+               pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+               pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+       }
+
+       if (reg_addr >> 32) {
+               if (unlikely(!adev->nbio.funcs))
+                       pcie_index_hi = (0x44 >> 2);

I'd still use a define for these, E.g.,

#define AMDGPU_PCIE_INDEX_FALLBACK (0x38 >> 2) etc. Or something similar.

Alex

Quoted 12 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +4 -0
+               else
+                       pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
+       } else {
                pcie_index_hi = 0;
+       }
 
        spin_lock_irqsave(&adev->pcie_idx_lock, flags);
        pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index *

4;
--
2.17.1

#5 lore raw

PATCH 3/5 drm/amdgpu: Add ras helper to query boot errors

In reply to Hawking Zhang

Add ras helper function to query boot time gpu
errors.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 95 +++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 15 +++-
 3 files changed, 112 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +3 -0
index 616b6c911767..db44ec857a31 100644
@@ -1328,6 +1328,9 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define WREG32_FIELD_OFFSET(reg, offset, field, val)	\
 	WREG32(mm##reg + offset, (RREG32(mm##reg + offset) & ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
 
+#define AMDGPU_SMN_TARGET_AID(x) ((u64)(x) << 32)
+#define AMDGPU_SMN_CROSS_AID (1ULL << 34)
+#define AMDGPU_GET_REG_FIELD(x, h, l) (((x) & GENMASK_ULL(h, l)) >> (l))
 /*
  * BIOS helpers.
  */
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +95 -0

98 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +14 -1
index 76fb85628716..5785b705c692 100644
@@ -32,6 +32,19 @@
 
 struct amdgpu_iv_entry;
 
+#define AMDGPU_RAS_GPU_ERR_MEM_TRAINING(x)		AMDGPU_GET_REG_FIELD(x, 0, 0)
+#define AMDGPU_RAS_GPU_ERR_FW_LOAD(x)			AMDGPU_GET_REG_FIELD(x, 1, 1)
+#define AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 2, 2)
+#define AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 3, 3)
+#define AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 4, 4)
+#define AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 5, 5)
+#define AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(x)		AMDGPU_GET_REG_FIELD(x, 6, 6)
+#define AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(x)		AMDGPU_GET_REG_FIELD(x, 7, 7)
+#define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x)			AMDGPU_GET_REG_FIELD(x, 10, 8)
+#define AMDGPU_RAS_GPU_ERR_AID_ID(x)			AMDGPU_GET_REG_FIELD(x, 12, 11)
+#define AMDGPU_RAS_GPU_ERR_HBM_ID(x)			AMDGPU_GET_REG_FIELD(x, 13, 13)
+#define AMDGPU_RAS_GPU_ERR_BOOT_STATUS(x)		AMDGPU_GET_REG_FIELD(x, 31, 31)
+
 #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS		(0x1 << 0)
 /* position of instance value in sub_block_index of
  * ta_ras_trigger_error_input, the sub block uses lower 12 bits
@@ -818,5 +831,5 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,
 int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
 		struct amdgpu_smuio_mcm_config_info *mcm_info,
 		struct ras_err_addr *err_addr, u64 count);
-
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 num_instances);
 #endif
Signature · 1 line
2.17.1
#6 lore raw

RE: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

[AMD Official Use Only - General]

-----Original Message-----
From: Hawking Zhang <[email protected]>
Sent: Tuesday, January 2, 2024 11:44 AM
To: [email protected]; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Wang, Yang(Kevin) <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>
Cc: Zhang, Hawking <[email protected]>; Deucher, Alexander <[email protected]>; Lazar, Lijo <[email protected]>; Ma, Le <[email protected]>
Subject: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

Add ras helper function to query boot time gpu errors.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 95 +++++++++++++++++++++++++  drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 15 +++-
 3 files changed, 112 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +3 -0
index 616b6c911767..db44ec857a31 100644
@@ -1328,6 +1328,9 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define WREG32_FIELD_OFFSET(reg, offset, field, val)   \
        WREG32(mm##reg + offset, (RREG32(mm##reg + offset) & ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
 
+#define AMDGPU_SMN_TARGET_AID(x) ((u64)(x) << 32) #define
+AMDGPU_SMN_CROSS_AID (1ULL << 34) #define AMDGPU_GET_REG_FIELD(x, h, l)
+(((x) & GENMASK_ULL(h, l)) >> (l))
 

[kevin]:
The macro GENMASK_ULL() will return a 64bit mask value, but the register is 32 bits (in this patch), do we need to change it to GENMASK() ? or you want to cover 64bit register cases.. Thanks.

Best Regards,
Kevin
/*
* BIOS helpers.
*/

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +94 -0

97 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +13 -0
index 76fb85628716..5785b705c692 100644
@@ -32,6 +32,19 @@
 
 struct amdgpu_iv_entry;
 
+#define AMDGPU_RAS_GPU_ERR_MEM_TRAINING(x)             AMDGPU_GET_REG_FIELD(x, 0, 0)
+#define AMDGPU_RAS_GPU_ERR_FW_LOAD(x)                  AMDGPU_GET_REG_FIELD(x, 1, 1)
+#define AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(x)       AMDGPU_GET_REG_FIELD(x, 2, 2)
+#define AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(x)       AMDGPU_GET_REG_FIELD(x, 3, 3)
+#define AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(x)     AMDGPU_GET_REG_FIELD(x, 4, 4)
+#define AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(x)     AMDGPU_GET_REG_FIELD(x, 5, 5)
+#define AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(x)             AMDGPU_GET_REG_FIELD(x, 6, 6)
+#define AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(x)            AMDGPU_GET_REG_FIELD(x, 7, 7)
+#define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x)                        AMDGPU_GET_REG_FIELD(x, 10, 8)
+#define AMDGPU_RAS_GPU_ERR_AID_ID(x)                   AMDGPU_GET_REG_FIELD(x, 12, 11)
+#define AMDGPU_RAS_GPU_ERR_HBM_ID(x)                   AMDGPU_GET_REG_FIELD(x, 13, 13)
+#define AMDGPU_RAS_GPU_ERR_BOOT_STATUS(x)              AMDGPU_GET_REG_FIELD(x, 31, 31)
+
 #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS          (0x1 << 0)
 /* position of instance value in sub_block_index of
  * ta_ras_trigger_error_input, the sub block uses lower 12 bits @@ -818,5 +831,5 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,  int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,

struct amdgpu_smuio_mcm_config_info *mcm_info, struct ras_err_addr *err_addr, u64 count);
-
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 +num_instances);
#endif
--
2.17.1

#7 lore raw

RE: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

[AMD Official Use Only - General]

I was hoping the macro could be used for 64bit registers as well. i.e., the aca regs.

Regards,
Hawking

-----Original Message-----
From: Wang, Yang(Kevin) <[email protected]>
Sent: Tuesday, January 2, 2024 13:24
To: Zhang, Hawking <[email protected]>; [email protected]; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>
Cc: Zhang, Hawking <[email protected]>; Deucher, Alexander <[email protected]>; Lazar, Lijo <[email protected]>; Ma, Le <[email protected]>
Subject: RE: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

[AMD Official Use Only - General]

-----Original Message-----
From: Hawking Zhang <[email protected]>
Sent: Tuesday, January 2, 2024 11:44 AM
To: [email protected]; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Wang, Yang(Kevin) <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>
Cc: Zhang, Hawking <[email protected]>; Deucher, Alexander <[email protected]>; Lazar, Lijo <[email protected]>; Ma, Le <[email protected]>
Subject: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

Add ras helper function to query boot time gpu errors.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 95 +++++++++++++++++++++++++  drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 15 +++-
 3 files changed, 112 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +3 -0
index 616b6c911767..db44ec857a31 100644
@@ -1328,6 +1328,9 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define WREG32_FIELD_OFFSET(reg, offset, field, val)   \
        WREG32(mm##reg + offset, (RREG32(mm##reg + offset) & ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
 
+#define AMDGPU_SMN_TARGET_AID(x) ((u64)(x) << 32) #define
+AMDGPU_SMN_CROSS_AID (1ULL << 34) #define AMDGPU_GET_REG_FIELD(x, h, l)
+(((x) & GENMASK_ULL(h, l)) >> (l))
 

[kevin]:
The macro GENMASK_ULL() will return a 64bit mask value, but the register is 32 bits (in this patch), do we need to change it to GENMASK() ? or you want to cover 64bit register cases.. Thanks.

Best Regards,
Kevin
/*
* BIOS helpers.
*/

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +93 -0

96 lines not on this page · raw message

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +13 -0
index 76fb85628716..5785b705c692 100644
@@ -32,6 +32,19 @@
 
 struct amdgpu_iv_entry;
 
+#define AMDGPU_RAS_GPU_ERR_MEM_TRAINING(x)             AMDGPU_GET_REG_FIELD(x, 0, 0)
+#define AMDGPU_RAS_GPU_ERR_FW_LOAD(x)                  AMDGPU_GET_REG_FIELD(x, 1, 1)
+#define AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(x)       AMDGPU_GET_REG_FIELD(x, 2, 2)
+#define AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(x)       AMDGPU_GET_REG_FIELD(x, 3, 3)
+#define AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(x)     AMDGPU_GET_REG_FIELD(x, 4, 4)
+#define AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(x)     AMDGPU_GET_REG_FIELD(x, 5, 5)
+#define AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(x)             AMDGPU_GET_REG_FIELD(x, 6, 6)
+#define AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(x)            AMDGPU_GET_REG_FIELD(x, 7, 7)
+#define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x)                        AMDGPU_GET_REG_FIELD(x, 10, 8)
+#define AMDGPU_RAS_GPU_ERR_AID_ID(x)                   AMDGPU_GET_REG_FIELD(x, 12, 11)
+#define AMDGPU_RAS_GPU_ERR_HBM_ID(x)                   AMDGPU_GET_REG_FIELD(x, 13, 13)
+#define AMDGPU_RAS_GPU_ERR_BOOT_STATUS(x)              AMDGPU_GET_REG_FIELD(x, 31, 31)
+
 #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS          (0x1 << 0)
 /* position of instance value in sub_block_index of
  * ta_ras_trigger_error_input, the sub block uses lower 12 bits @@ -818,5 +831,5 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,  int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,

struct amdgpu_smuio_mcm_config_info *mcm_info, struct ras_err_addr *err_addr, u64 count);
-
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 +num_instances);
#endif
--
2.17.1

#8 lore raw

RE: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

In reply to Hawking Zhang

[AMD Official Use Only - General]

Quoted 58 lines

-----Original Message-----
From: Hawking Zhang <[email protected]>
Sent: Tuesday, January 2, 2024 11:44 AM
To: [email protected]; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Wang, Yang(Kevin) <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>
Cc: Zhang, Hawking <[email protected]>; Deucher, Alexander <[email protected]>; Lazar, Lijo <[email protected]>; Ma, Le <[email protected]>
Subject: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

Add ras helper function to query boot time gpu errors.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 95
+++++++++++++++++++++++++  drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h |
15 +++-
 3 files changed, 112 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +3 -0
index 616b6c911767..db44ec857a31 100644
@@ -1328,6 +1328,9 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define WREG32_FIELD_OFFSET(reg, offset, field, val) \
      WREG32(mm##reg + offset, (RREG32(mm##reg + offset) &
 ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
 
+#define AMDGPU_SMN_TARGET_AID(x) ((u64)(x) << 32) #define
+AMDGPU_SMN_CROSS_AID (1ULL << 34) #define AMDGPU_GET_REG_FIELD(x,
 h, l)
+(((x) & GENMASK_ULL(h, l)) >> (l))
 /*

* BIOS helpers.
*/

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
index 39399d0f2ce5..5f302b7693b3 100644
@@ -3764,3 +3764,98 @@ int amdgpu_ras_error_statistic_ce_count(struct
 ras_err_data *err_data,
 
      return 0;
 }
+
+#define mmMP0_SMN_C2PMSG_92  0x1609C
+#define mmMP0_SMN_C2PMSG_126 0x160BE
+static void amdgpu_ras_boot_time_error_reporting(struct amdgpu_device

*adev,

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                                              u32 instance, u32 boot_error)
+{
+     u32 socket_id, aid_id, hbm_id;
+     u32 reg_data;

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data, · line 3784

Quoted 10 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +10 -0
+     u64 reg_addr;
+
+     socket_id = AMDGPU_RAS_GPU_ERR_SOCKET_ID(boot_error);
+     aid_id = AMDGPU_RAS_GPU_ERR_AID_ID(boot_error);
+     hbm_id = AMDGPU_RAS_GPU_ERR_HBM_ID(boot_error);
+
+     if (instance)
+             reg_addr = (mmMP0_SMN_C2PMSG_92 << 2) +
+                        AMDGPU_SMN_TARGET_AID(instance) +
+                        AMDGPU_SMN_CROSS_AID;

Hi Hawking,

We have asic function "aqua_vanjaram_encode_ext_smn_addressing" for this stuff, maybe it could also be re-used here.

Thanks.

Quoted 134 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5 -0
+     else
+             reg_addr = (mmMP0_SMN_C2PMSG_92 << 2);
+
+     reg_data = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+     dev_err(adev->dev, "socket: %d, aid: %d, firmware boot failed, fw

status is 0x%x\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+             socket_id, aid_id, reg_data);
+
+     if (AMDGPU_RAS_GPU_ERR_MEM_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, memory

training failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                      socket_id, aid_id, hbm_id);
+
+     if (AMDGPU_RAS_GPU_ERR_FW_LOAD(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, firmware load failed

at boot time\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, wafl link training

failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, xgmi link training

failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, usr cp link training

failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, usr dp link training

failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, hbm

memory test failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                      socket_id, aid_id, hbm_id);
+
+     if (AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, hbm bist

test failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +37 -0
+                      socket_id, aid_id, hbm_id);
+}
+
+static int amdgpu_ras_wait_for_boot_complete(struct amdgpu_device *adev,
+                                          u32 instance, u32 *boot_error) {
+     u32 reg_addr;
+     u32 reg_data;
+     int retry_loop;
+
+     if (instance)
+             reg_addr = (mmMP0_SMN_C2PMSG_126 << 2) +
+                        AMDGPU_SMN_TARGET_AID(instance) +
+                        AMDGPU_SMN_CROSS_AID;
+     else
+             reg_addr = (mmMP0_SMN_C2PMSG_126 << 2);
+
+     for (retry_loop = 0; retry_loop < 1000; retry_loop++) {
+             reg_data = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+             if (AMDGPU_RAS_GPU_ERR_BOOT_STATUS(reg_data)) {
+                     *boot_error = reg_data;
+                     return 0;
+             }
+             msleep(1);
+     }
+
+     *boot_error = reg_data;
+     return -ETIME;
+}
+
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32
+num_instances) {
+     u32 boot_error = 0;
+     u32 i;
+
+     for (i = 0; i < num_instances; i++) {
+             if (amdgpu_ras_wait_for_boot_complete(adev, i, &boot_error))
+                     amdgpu_ras_boot_time_error_reporting(adev, i,

boot_error);

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +2 -0
+     }
+}
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +14 -0
index 76fb85628716..5785b705c692 100644
@@ -32,6 +32,19 @@
 
 struct amdgpu_iv_entry;
 
+#define AMDGPU_RAS_GPU_ERR_MEM_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 0, 0)
+#define AMDGPU_RAS_GPU_ERR_FW_LOAD(x)
      AMDGPU_GET_REG_FIELD(x, 1, 1)
+#define AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 2, 2)
+#define AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 3, 3)
+#define AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 4, 4)
+#define AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 5, 5)
+#define AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(x)
      AMDGPU_GET_REG_FIELD(x, 6, 6)
+#define AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(x)
      AMDGPU_GET_REG_FIELD(x, 7, 7)
+#define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x)
      AMDGPU_GET_REG_FIELD(x, 10, 8)
+#define AMDGPU_RAS_GPU_ERR_AID_ID(x)
      AMDGPU_GET_REG_FIELD(x, 12, 11)
+#define AMDGPU_RAS_GPU_ERR_HBM_ID(x)
      AMDGPU_GET_REG_FIELD(x, 13, 13)
+#define AMDGPU_RAS_GPU_ERR_BOOT_STATUS(x)
      AMDGPU_GET_REG_FIELD(x, 31, 31)
+
 #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS                (0x1 << 0)
 /* position of instance value in sub_block_index of
  * ta_ras_trigger_error_input, the sub block uses lower 12 bits @@ -818,5
+831,5 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data

*err_data, int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
struct amdgpu_smuio_mcm_config_info *mcm_info, struct ras_err_addr *err_addr, u64 count);

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +2 -2
-
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32
+num_instances);
 #endif
--

2.17.1

#9 lore raw

RE: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

[AMD Official Use Only - General]

Good point, Le, will switch to the existing helper for the cross die access in v2.

Regards,
Hawking

-----Original Message-----
From: Ma, Le <[email protected]>
Sent: Tuesday, January 2, 2024 14:45
To: Zhang, Hawking <[email protected]>; [email protected]; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Wang, Yang(Kevin) <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>
Cc: Zhang, Hawking <[email protected]>; Deucher, Alexander <[email protected]>; Lazar, Lijo <[email protected]>
Subject: RE: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

[AMD Official Use Only - General]

Quoted 58 lines

-----Original Message-----
From: Hawking Zhang <[email protected]>
Sent: Tuesday, January 2, 2024 11:44 AM
To: [email protected]; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Wang, Yang(Kevin) <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>
Cc: Zhang, Hawking <[email protected]>; Deucher, Alexander <[email protected]>; Lazar, Lijo <[email protected]>; Ma, Le <[email protected]>
Subject: [PATCH 3/5] drm/amdgpu: Add ras helper to query boot errors

Add ras helper function to query boot time gpu errors.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  3 +
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 95
+++++++++++++++++++++++++  drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h |
15 +++-
 3 files changed, 112 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu.h +3 -0
index 616b6c911767..db44ec857a31 100644
@@ -1328,6 +1328,9 @@ int emu_soc_asic_init(struct amdgpu_device
 *adev);  #define WREG32_FIELD_OFFSET(reg, offset, field, val) \
      WREG32(mm##reg + offset, (RREG32(mm##reg + offset) &
 ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
 
+#define AMDGPU_SMN_TARGET_AID(x) ((u64)(x) << 32) #define
+AMDGPU_SMN_CROSS_AID (1ULL << 34) #define AMDGPU_GET_REG_FIELD(x,
 h, l)
+(((x) & GENMASK_ULL(h, l)) >> (l))
 /*

* BIOS helpers.
*/

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
index 39399d0f2ce5..5f302b7693b3 100644
@@ -3764,3 +3764,98 @@ int amdgpu_ras_error_statistic_ce_count(struct
 ras_err_data *err_data,
 
      return 0;
 }
+
+#define mmMP0_SMN_C2PMSG_92  0x1609C
+#define mmMP0_SMN_C2PMSG_126 0x160BE
+static void amdgpu_ras_boot_time_error_reporting(struct amdgpu_device

*adev,

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                                              u32 instance, u32
+boot_error) {
+     u32 socket_id, aid_id, hbm_id;
+     u32 reg_data;

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data, · line 3784

Quoted 10 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +10 -0
+     u64 reg_addr;
+
+     socket_id = AMDGPU_RAS_GPU_ERR_SOCKET_ID(boot_error);
+     aid_id = AMDGPU_RAS_GPU_ERR_AID_ID(boot_error);
+     hbm_id = AMDGPU_RAS_GPU_ERR_HBM_ID(boot_error);
+
+     if (instance)
+             reg_addr = (mmMP0_SMN_C2PMSG_92 << 2) +
+                        AMDGPU_SMN_TARGET_AID(instance) +
+                        AMDGPU_SMN_CROSS_AID;

Hi Hawking,

We have asic function "aqua_vanjaram_encode_ext_smn_addressing" for this stuff, maybe it could also be re-used here.

Thanks.

Quoted 142 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +6 -0
+     else
+             reg_addr = (mmMP0_SMN_C2PMSG_92 << 2);
+
+     reg_data = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+     dev_err(adev->dev, "socket: %d, aid: %d, firmware boot failed,
+ fw

status is 0x%x\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5 -0
+             socket_id, aid_id, reg_data);
+
+     if (AMDGPU_RAS_GPU_ERR_MEM_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d,
+ memory

training failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5 -0
+                      socket_id, aid_id, hbm_id);
+
+     if (AMDGPU_RAS_GPU_ERR_FW_LOAD(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, firmware load
+ failed

at boot time\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, wafl link
+ training

failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, xgmi link
+ training

failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, usr cp link
+ training

failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, usr dp link
+ training

failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +4 -0
+                      socket_id, aid_id);
+
+     if (AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, hbm

memory test failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5 -0
+                      socket_id, aid_id, hbm_id);
+
+     if (AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(boot_error))
+             dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, hbm
+ bist

test failed\n",

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +36 -0
+                      socket_id, aid_id, hbm_id); }
+
+static int amdgpu_ras_wait_for_boot_complete(struct amdgpu_device *adev,
+                                          u32 instance, u32 *boot_error) {
+     u32 reg_addr;
+     u32 reg_data;
+     int retry_loop;
+
+     if (instance)
+             reg_addr = (mmMP0_SMN_C2PMSG_126 << 2) +
+                        AMDGPU_SMN_TARGET_AID(instance) +
+                        AMDGPU_SMN_CROSS_AID;
+     else
+             reg_addr = (mmMP0_SMN_C2PMSG_126 << 2);
+
+     for (retry_loop = 0; retry_loop < 1000; retry_loop++) {
+             reg_data = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+             if (AMDGPU_RAS_GPU_ERR_BOOT_STATUS(reg_data)) {
+                     *boot_error = reg_data;
+                     return 0;
+             }
+             msleep(1);
+     }
+
+     *boot_error = reg_data;
+     return -ETIME;
+}
+
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32
+num_instances) {
+     u32 boot_error = 0;
+     u32 i;
+
+     for (i = 0; i < num_instances; i++) {
+             if (amdgpu_ras_wait_for_boot_complete(adev, i, &boot_error))
+                     amdgpu_ras_boot_time_error_reporting(adev, i,

boot_error);

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +2 -0
+     }
+}
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +14 -1
index 76fb85628716..5785b705c692 100644
@@ -32,6 +32,19 @@
 
 struct amdgpu_iv_entry;
 
+#define AMDGPU_RAS_GPU_ERR_MEM_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 0, 0)
+#define AMDGPU_RAS_GPU_ERR_FW_LOAD(x)
      AMDGPU_GET_REG_FIELD(x, 1, 1)
+#define AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 2, 2)
+#define AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 3, 3)
+#define AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 4, 4)
+#define AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(x)
      AMDGPU_GET_REG_FIELD(x, 5, 5)
+#define AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(x)
      AMDGPU_GET_REG_FIELD(x, 6, 6)
+#define AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(x)
      AMDGPU_GET_REG_FIELD(x, 7, 7)
+#define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x)
      AMDGPU_GET_REG_FIELD(x, 10, 8)
+#define AMDGPU_RAS_GPU_ERR_AID_ID(x)
      AMDGPU_GET_REG_FIELD(x, 12, 11)
+#define AMDGPU_RAS_GPU_ERR_HBM_ID(x)
      AMDGPU_GET_REG_FIELD(x, 13, 13)
+#define AMDGPU_RAS_GPU_ERR_BOOT_STATUS(x)
      AMDGPU_GET_REG_FIELD(x, 31, 31)
+
 #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS                (0x1 << 0)
 /* position of instance value in sub_block_index of
  * ta_ras_trigger_error_input, the sub block uses lower 12 bits @@
-818,5
+831,5 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data

*err_data, int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
struct amdgpu_smuio_mcm_config_info *mcm_info, struct ras_err_addr *err_addr, u64 count);

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +2 -2
-
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32
+num_instances);
 #endif
--

2.17.1

#10 lore raw

PATCH 4/5 drm/amdgpu: Query boot status if discovery failed

In reply to Hawking Zhang

Check and report boot status if discovery failed.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +5 -1
index b8fde08aec8e..302b71e9f1e2 100644
@@ -27,6 +27,7 @@
 #include "amdgpu_discovery.h"
 #include "soc15_hw_ip.h"
 #include "discovery.h"
+#include "amdgpu_ras.h"
 
 #include "soc15.h"
 #include "gfx_v9_0.h"
@@ -98,6 +99,7 @@
 #define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin"
 MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY);
 
+#define mmIP_DISCOVERY_VERSION  0x16A00
 #define mmRCC_CONFIG_MEMSIZE	0xde3
 #define mmMP0_SMN_C2PMSG_33	0x16061
 #define mmMM_INDEX		0x0
@@ -518,7 +520,9 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev)
 out:
 	kfree(adev->mman.discovery_bin);
 	adev->mman.discovery_bin = NULL;
-
+	if ((amdgpu_discovery != 2) &&
+	    (RREG32(mmIP_DISCOVERY_VERSION) == 4))
+		amdgpu_ras_query_boot_status(adev, 4);
 	return r;
 }
 
Signature · 1 line
2.17.1
#11 lore raw

Re: [PATCH 4/5] drm/amdgpu: Query boot status if discovery failed

[AMD Official Use Only - General]

Is mmIP_DISCOVERY_VERSION at the same offset across ASIC families?

Alex

________________________________
From: Hawking Zhang <[email protected]>
Sent: Monday, January 1, 2024 10:43 PM
To: [email protected] <[email protected]>; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Wang, Yang(Kevin) <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>
Cc: Zhang, Hawking <[email protected]>; Deucher, Alexander <[email protected]>; Lazar, Lijo <[email protected]>; Ma, Le <[email protected]>
Subject: [PATCH 4/5] drm/amdgpu: Query boot status if discovery failed

Check and report boot status if discovery failed.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +5 -1
index b8fde08aec8e..302b71e9f1e2 100644
@@ -27,6 +27,7 @@
 #include "amdgpu_discovery.h"
 #include "soc15_hw_ip.h"
 #include "discovery.h"
+#include "amdgpu_ras.h"
 
 #include "soc15.h"
 #include "gfx_v9_0.h"
@@ -98,6 +99,7 @@
 #define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin"
 MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY);
 
+#define mmIP_DISCOVERY_VERSION  0x16A00
 #define mmRCC_CONFIG_MEMSIZE    0xde3
 #define mmMP0_SMN_C2PMSG_33     0x16061
 #define mmMM_INDEX              0x0
@@ -518,7 +520,9 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev)
 out:
         kfree(adev->mman.discovery_bin);
         adev->mman.discovery_bin = NULL;
-
+       if ((amdgpu_discovery != 2) &&
+           (RREG32(mmIP_DISCOVERY_VERSION) == 4))
+               amdgpu_ras_query_boot_status(adev, 4);
         return r;
 }
 

--
2.17.1

#12 lore raw

RE: [PATCH 4/5] drm/amdgpu: Query boot status if discovery failed

[AMD Official Use Only - General]

Yes, it is.

Regards,
Hawking

From: Deucher, Alexander <[email protected]>
Sent: Wednesday, January 3, 2024 03:45
To: Zhang, Hawking <[email protected]>; [email protected]; Zhou1, Tao <[email protected]>; Yang, Stanley <[email protected]>; Wang, Yang(Kevin) <[email protected]>; Chai, Thomas <[email protected]>; Li, Candice <[email protected]>
Cc: Lazar, Lijo <[email protected]>; Ma, Le <[email protected]>
Subject: Re: [PATCH 4/5] drm/amdgpu: Query boot status if discovery failed

[AMD Official Use Only - General]

Is mmIP_DISCOVERY_VERSION at the same offset across ASIC families?

Alex

________________________________
From: Hawking Zhang <[email protected]<mailto:[email protected]>>
Sent: Monday, January 1, 2024 10:43 PM
To: [email protected]<mailto:[email protected]> <[email protected]<mailto:[email protected]>>; Zhou1, Tao <[email protected]<mailto:[email protected]>>; Yang, Stanley <[email protected]<mailto:[email protected]>>; Wang, Yang(Kevin) <[email protected]<mailto:[email protected]>>; Chai, Thomas <[email protected]<mailto:[email protected]>>; Li, Candice <[email protected]<mailto:[email protected]>>
Cc: Zhang, Hawking <[email protected]<mailto:[email protected]>>; Deucher, Alexander <[email protected]<mailto:[email protected]>>; Lazar, Lijo <[email protected]<mailto:[email protected]>>; Ma, Le <[email protected]<mailto:[email protected]>>
Subject: [PATCH 4/5] drm/amdgpu: Query boot status if discovery failed

Check and report boot status if discovery failed.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +5 -1
index b8fde08aec8e..302b71e9f1e2 100644
@@ -27,6 +27,7 @@
 #include "amdgpu_discovery.h"
 #include "soc15_hw_ip.h"
 #include "discovery.h"
+#include "amdgpu_ras.h"
 
 #include "soc15.h"
 #include "gfx_v9_0.h"
@@ -98,6 +99,7 @@
 #define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin"
 MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY);
 
+#define mmIP_DISCOVERY_VERSION  0x16A00
 #define mmRCC_CONFIG_MEMSIZE    0xde3
 #define mmMP0_SMN_C2PMSG_33     0x16061
 #define mmMM_INDEX              0x0
@@ -518,7 +520,9 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev)
 out:
         kfree(adev->mman.discovery_bin);
         adev->mman.discovery_bin = NULL;
-
+       if ((amdgpu_discovery != 2) &&
+           (RREG32(mmIP_DISCOVERY_VERSION) == 4))
+               amdgpu_ras_query_boot_status(adev, 4);
         return r;
 }
 

--
2.17.1

#13 lore raw

PATCH 5/5 drm/amdgpu: Query boot status if boot failed

In reply to Hawking Zhang

Check and report firmware boot status if it doesn't reach steady status.

---
 drivers/gpu/drm/amd/amdgpu/psp_v13_0.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/psp_v13_0.c +9 -2
index 6fad451a85be..676bec2cc157 100644
@@ -187,11 +187,18 @@ static int psp_v13_0_wait_for_bootloader(struct psp_context *psp)
 static int psp_v13_0_wait_for_bootloader_steady_state(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
+	int ret;
 
 	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6)) {
-		psp_v13_0_wait_for_vmbx_ready(psp);
+		ret = psp_v13_0_wait_for_vmbx_ready(psp);
+		if (ret)
+			amdgpu_ras_query_boot_status(adev, 4);
+
+		ret = psp_v13_0_wait_for_bootloader(psp);
+		if (ret)
+			amdgpu_ras_query_boot_status(adev, 4);
 
-		return psp_v13_0_wait_for_bootloader(psp);
+		return ret;
 	}
 
 	return 0;
Signature · 1 line
2.17.1