SWDEV-452795 - Disable RAS plugin, fix XGMI
RAS plugin loaded rocm-smi which is in conflict with amd-smi library Main source of grief was the map 'devInfoTypesStrings' that is defined in both rocm-smi and amd-smi We assume that rocm-smi would get lazy-loaded by RAS library and overwrite symbols defined in amd-smi. devInfoTypesStrings in rocm-smi contains different number of elements, the enums are also different. RDC relies on amd-smi's enums. One such enum is kDevGpuMetrics: rocm-smi: kDevGpuMetrics = 68 amd-smi: kDevGpuMetrics = 75 Example of overlapping map definitions: $ objdump --dynamic-syms /opt/rocm/lib/libamd_smi.so | grep devInfoTypesStrings 00000000003c4980 g DO .data.rel.ro0000000000000008 Base devInfoTypesStrings 00000000003db830 g DO .bss0000000000000030 Base _ZN3amd3smi6Device19devInfoTypesStringsE $ objdump --dynamic-syms /opt/rocm/lib/librocm_smi64.so | grep devInfoTypesStrings 00000000003dc590 g DO .bss0000000000000030 Base _ZN3amd3smi6Device19devInfoTypesStringsE 00000000003c9c68 g DO .data.rel.ro0000000000000008 Base devInfoTypesStrings Change-Id: Ib2f2db32b6abd7ebe84e7807c25581461eb86bae Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
Αυτή η υποβολή περιλαμβάνεται σε:
1 γονέας
a885944d97
υποβολή
d85657e5f2
184 files changed
+277
-25987
No files matched your search
+47
-41
@@ -185,47 +185,46 @@ typedef enum {
|
||||
RDC_FI_ECC_CORRECT_TOTAL = 600, //!< Accumulated correctable ECC errors
|
||||
RDC_FI_ECC_UNCORRECT_TOTAL, //!< Accumulated uncorrectable ECC errors
|
||||
|
||||
RDC_FI_ECC_SDMA_SEC, //!< SDMA Single Error Correction
|
||||
RDC_FI_ECC_SDMA_DED, //!< SDMA Double Error Detection
|
||||
|
||||
RDC_FI_ECC_GFX_SEC, //!< GFX Single Error Correction
|
||||
RDC_FI_ECC_GFX_DED, //!< GFX Double Error Detection
|
||||
|
||||
RDC_FI_ECC_MMHUB_SEC, //!< MMHUB Single Error Correction
|
||||
RDC_FI_ECC_MMHUB_DED, //!< MMHUB Double Error Detection
|
||||
|
||||
RDC_FI_ECC_ATHUB_SEC, //!< ATHUB Single Error Correction
|
||||
RDC_FI_ECC_ATHUB_DED, //!< ATHUB Double Error Detection
|
||||
|
||||
RDC_FI_ECC_BIF_SEC, //!< BIF Single Error Correction
|
||||
RDC_FI_ECC_BIF_DED, //!< BIF Double Error Detection
|
||||
|
||||
RDC_FI_ECC_HDP_SEC, //!< HDP Single Error Correction
|
||||
RDC_FI_ECC_HDP_DED, //!< HDP Double Error Detection
|
||||
|
||||
RDC_FI_ECC_XGMI_WAFL_SEC, //!< XGMI WAFL Single Error Correction
|
||||
RDC_FI_ECC_XGMI_WAFL_DED, //!< XGMI WAFL Double Error Detection
|
||||
|
||||
RDC_FI_ECC_DF_SEC, //!< DF Single Error Correction
|
||||
RDC_FI_ECC_DF_DED, //!< DF Double Error Detection
|
||||
|
||||
RDC_FI_ECC_SMN_SEC, //!< SMN Single Error Correction
|
||||
RDC_FI_ECC_SMN_DED, //!< SMN Double Error Detection
|
||||
|
||||
RDC_FI_ECC_SEM_SEC, //!< SEM Single Error Correction
|
||||
RDC_FI_ECC_SEM_DED, //!< SEM Double Error Detection
|
||||
|
||||
RDC_FI_ECC_MP0_SEC, //!< MP0 Single Error Correction
|
||||
RDC_FI_ECC_MP0_DED, //!< MP0 Double Error Detection
|
||||
|
||||
RDC_FI_ECC_MP1_SEC, //!< MP1 Single Error Correction
|
||||
RDC_FI_ECC_MP1_DED, //!< MP1 Double Error Detection
|
||||
|
||||
RDC_FI_ECC_FUSE_SEC, //!< FUSE Single Error Correction
|
||||
RDC_FI_ECC_FUSE_DED, //!< FUSE Double Error Detection
|
||||
|
||||
RDC_FI_ECC_UMC_SEC, //!< UMC Single Error Correction
|
||||
RDC_FI_ECC_UMC_DED, //!< UMC Double Error Detection
|
||||
RDC_FI_ECC_FIRST = 602, //!< FIRST Error Correction and Detection field
|
||||
RDC_FI_ECC_SDMA_CE = RDC_FI_ECC_FIRST,
|
||||
RDC_FI_ECC_SDMA_UE,
|
||||
RDC_FI_ECC_GFX_CE,
|
||||
RDC_FI_ECC_GFX_UE,
|
||||
RDC_FI_ECC_MMHUB_CE,
|
||||
RDC_FI_ECC_MMHUB_UE,
|
||||
RDC_FI_ECC_ATHUB_CE,
|
||||
RDC_FI_ECC_ATHUB_UE,
|
||||
RDC_FI_ECC_PCIE_BIF_CE,
|
||||
RDC_FI_ECC_PCIE_BIF_UE,
|
||||
RDC_FI_ECC_HDP_CE,
|
||||
RDC_FI_ECC_HDP_UE,
|
||||
RDC_FI_ECC_XGMI_WAFL_CE,
|
||||
RDC_FI_ECC_XGMI_WAFL_UE,
|
||||
RDC_FI_ECC_DF_CE,
|
||||
RDC_FI_ECC_DF_UE,
|
||||
RDC_FI_ECC_SMN_CE,
|
||||
RDC_FI_ECC_SMN_UE,
|
||||
RDC_FI_ECC_SEM_CE,
|
||||
RDC_FI_ECC_SEM_UE,
|
||||
RDC_FI_ECC_MP0_CE,
|
||||
RDC_FI_ECC_MP0_UE,
|
||||
RDC_FI_ECC_MP1_CE,
|
||||
RDC_FI_ECC_MP1_UE,
|
||||
RDC_FI_ECC_FUSE_CE,
|
||||
RDC_FI_ECC_FUSE_UE,
|
||||
RDC_FI_ECC_UMC_CE,
|
||||
RDC_FI_ECC_UMC_UE,
|
||||
RDC_FI_ECC_MCA_CE,
|
||||
RDC_FI_ECC_MCA_UE,
|
||||
RDC_FI_ECC_VCN_CE,
|
||||
RDC_FI_ECC_VCN_UE,
|
||||
RDC_FI_ECC_JPEG_CE,
|
||||
RDC_FI_ECC_JPEG_UE,
|
||||
RDC_FI_ECC_IH_CE,
|
||||
RDC_FI_ECC_IH_UE,
|
||||
RDC_FI_ECC_MPIO_CE,
|
||||
RDC_FI_ECC_MPIO_UE,
|
||||
RDC_FI_ECC_LAST = RDC_FI_ECC_MPIO_UE,
|
||||
|
||||
// In new ASCI, such as MI300, the XGMI events is not supported
|
||||
// Using below XGMI related fields to calculate the bandwidth.
|
||||
@@ -320,6 +319,13 @@ typedef enum {
|
||||
|
||||
RDC_EVNT_NOTIF_LAST = RDC_EVNT_NOTIF_RING_HANG,
|
||||
} rdc_field_t;
|
||||
|
||||
// even and odd numbers are used for correctable and uncorrectable errors
|
||||
static_assert(RDC_FI_ECC_SDMA_CE % 2 == 0, "Correctable Error enum is not even");
|
||||
static_assert(RDC_FI_ECC_SDMA_UE % 2 == 1, "Uncorrectable Error enum is not odd");
|
||||
static_assert(RDC_FI_ECC_MPIO_CE % 2 == 0, "Correctable Error enum is not even");
|
||||
static_assert(RDC_FI_ECC_MPIO_UE % 2 == 1, "Uncorrectable Error enum is not odd");
|
||||
|
||||
#define RDC_EVNT_IS_NOTIF_FIELD(FIELD) \
|
||||
((FIELD) >= RDC_EVNT_NOTIF_FIRST && (FIELD) <= RDC_EVNT_NOTIF_LAST)
|
||||
/**
|
||||
|
||||
Reference in new issue
Block a user