From e54def4ad8144ab15f826416e2e0f290ef1901b4 Mon Sep 17 00:00:00 2001 From: Daniel Baumann Date: Wed, 19 Jun 2024 23:00:30 +0200 Subject: Adding upstream version 6.9.2. Signed-off-by: Daniel Baumann --- drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c | 96 ++++++++++++++++++++++++++++---- 1 file changed, 86 insertions(+), 10 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c') diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c index 328682cbb..e708468ac 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c @@ -45,6 +45,8 @@ MODULE_FIRMWARE("amdgpu/sdma_4_4_2.bin"); +#define mmSMNAID_AID0_MCA_SMU 0x03b30400 + #define WREG32_SDMA(instance, offset, value) \ WREG32(sdma_v4_4_2_get_reg_offset(adev, (instance), (offset)), value) #define RREG32_SDMA(instance, offset) \ @@ -596,7 +598,7 @@ static uint32_t sdma_v4_4_2_rb_cntl(struct amdgpu_ring *ring, uint32_t rb_cntl) /* Set ring buffer size in dwords */ uint32_t rb_bufsz = order_base_2(ring->ring_size / 4); - barrier(); /* work around https://bugs.llvm.org/show_bug.cgi?id=42576 */ + barrier(); /* work around https://llvm.org/pr42576 */ rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_GFX_RB_CNTL, RB_SIZE, rb_bufsz); #ifdef __BIG_ENDIAN rb_cntl = REG_SET_FIELD(rb_cntl, SDMA_GFX_RB_CNTL, RB_SWAP_ENABLE, 1); @@ -1612,7 +1614,7 @@ static int sdma_v4_4_2_print_iv_entry(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) { int instance; - struct amdgpu_task_info task_info; + struct amdgpu_task_info *task_info; u64 addr; instance = sdma_v4_4_2_irq_id_to_seq(entry->client_id); @@ -1624,15 +1626,19 @@ static int sdma_v4_4_2_print_iv_entry(struct amdgpu_device *adev, addr = (u64)entry->src_data[0] << 12; addr |= ((u64)entry->src_data[1] & 0xf) << 44; - memset(&task_info, 0, sizeof(struct amdgpu_task_info)); - amdgpu_vm_get_task_info(adev, entry->pasid, &task_info); - dev_dbg_ratelimited(adev->dev, - "[sdma%d] address:0x%016llx src_id:%u ring:%u vmid:%u " - "pasid:%u, for process %s pid %d thread %s pid %d\n", - instance, addr, entry->src_id, entry->ring_id, entry->vmid, - entry->pasid, task_info.process_name, task_info.tgid, - task_info.task_name, task_info.pid); + "[sdma%d] address:0x%016llx src_id:%u ring:%u vmid:%u pasid:%u\n", + instance, addr, entry->src_id, entry->ring_id, entry->vmid, + entry->pasid); + + task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid); + if (task_info) { + dev_dbg_ratelimited(adev->dev, " for process %s pid %d thread %s pid %d\n", + task_info->process_name, task_info->tgid, + task_info->task_name, task_info->pid); + amdgpu_vm_put_task_info(task_info); + } + return 0; } @@ -2174,9 +2180,79 @@ static const struct amdgpu_ras_block_hw_ops sdma_v4_4_2_ras_hw_ops = { .reset_ras_error_count = sdma_v4_4_2_reset_ras_error_count, }; +static int sdma_v4_4_2_aca_bank_generate_report(struct aca_handle *handle, + struct aca_bank *bank, enum aca_error_type type, + struct aca_bank_report *report, void *data) +{ + u64 status, misc0; + int ret; + + status = bank->regs[ACA_REG_IDX_STATUS]; + if ((type == ACA_ERROR_TYPE_UE && + ACA_REG__STATUS__ERRORCODEEXT(status) == ACA_EXTERROR_CODE_FAULT) || + (type == ACA_ERROR_TYPE_CE && + ACA_REG__STATUS__ERRORCODEEXT(status) == ACA_EXTERROR_CODE_CE)) { + + ret = aca_bank_info_decode(bank, &report->info); + if (ret) + return ret; + + misc0 = bank->regs[ACA_REG_IDX_MISC0]; + report->count[type] = ACA_REG__MISC0__ERRCNT(misc0); + } + + return 0; +} + +/* CODE_SDMA0 - CODE_SDMA4, reference to smu driver if header file */ +static int sdma_v4_4_2_err_codes[] = { 33, 34, 35, 36 }; + +static bool sdma_v4_4_2_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank, + enum aca_error_type type, void *data) +{ + u32 instlo; + + instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]); + instlo &= GENMASK(31, 1); + + if (instlo != mmSMNAID_AID0_MCA_SMU) + return false; + + if (aca_bank_check_error_codes(handle->adev, bank, + sdma_v4_4_2_err_codes, + ARRAY_SIZE(sdma_v4_4_2_err_codes))) + return false; + + return true; +} + +static const struct aca_bank_ops sdma_v4_4_2_aca_bank_ops = { + .aca_bank_generate_report = sdma_v4_4_2_aca_bank_generate_report, + .aca_bank_is_valid = sdma_v4_4_2_aca_bank_is_valid, +}; + +static const struct aca_info sdma_v4_4_2_aca_info = { + .hwip = ACA_HWIP_TYPE_SMU, + .mask = ACA_ERROR_UE_MASK, + .bank_ops = &sdma_v4_4_2_aca_bank_ops, +}; + +static int sdma_v4_4_2_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block) +{ + int r; + + r = amdgpu_sdma_ras_late_init(adev, ras_block); + if (r) + return r; + + return amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__SDMA, + &sdma_v4_4_2_aca_info, NULL); +} + static struct amdgpu_sdma_ras sdma_v4_4_2_ras = { .ras_block = { .hw_ops = &sdma_v4_4_2_ras_hw_ops, + .ras_late_init = sdma_v4_4_2_ras_late_init, }, }; -- cgit v1.2.3