From 6da15a236c8c80d9f87e4c5216e00ad8f1cace2d Mon Sep 17 00:00:00 2001 From: Candice Li Date: Wed, 7 Sep 2022 15:52:04 +0800 Subject: drm/amdgpu: Skip reset error status for psp v13_0_0 No need to reset error status since only umc ras supported on psp v13_0_0. Signed-off-by: Candice Li Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index ab9ba5a9c33d..e55f106621ef 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1811,7 +1811,8 @@ static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev) amdgpu_ras_query_error_status(adev, &info); if (adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 2) && - adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 4)) { + adev->ip_versions[MP0_HWIP][0] != IP_VERSION(11, 0, 4) && + adev->ip_versions[MP0_HWIP][0] != IP_VERSION(13, 0, 0)) { if (amdgpu_ras_reset_error_status(adev, info.head.block)) dev_warn(adev->dev, "Failed to reset error counter and error status"); } -- cgit v1.2.3 From 642c0401135d9877d6888fdd2b0fbbee3b134409 Mon Sep 17 00:00:00 2001 From: YiPeng Chai Date: Thu, 8 Sep 2022 09:48:11 +0800 Subject: drm/amdgpu: Fixed ras warning when uninstalling amdgpu For the asic using smu v13_0_2, there is the following warning when uninstalling amdgpu: amdgpu: ras disable gfx failed poison:1 ret:-22. [Why]: For the asic using smu v13_0_2, the psp .suspend and mode1reset is called before amdgpu_ras_pre_fini during amdgpu uninstall, it has disabled all ras features and reset the psp. Since the psp is reset, calling amdgpu_ras_disable_all_features in amdgpu_ras_pre_fini to disable ras features will fail. [How]: If all ras features are disabled, amdgpu_ras_disable_all_features will not be called to disable all ras features again. Signed-off-by: YiPeng Chai Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index e55f106621ef..3deb716710e6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -2720,7 +2720,8 @@ int amdgpu_ras_pre_fini(struct amdgpu_device *adev) /* Need disable ras on all IPs here before ip [hw/sw]fini */ - amdgpu_ras_disable_all_features(adev, 0); + if (con->features) + amdgpu_ras_disable_all_features(adev, 0); amdgpu_ras_recovery_fini(adev); return 0; } -- cgit v1.2.3 From cd4c99f103af5a85fcae5d31a9e85361679ab979 Mon Sep 17 00:00:00 2001 From: Tao Zhou Date: Wed, 21 Sep 2022 16:46:19 +0800 Subject: drm/amdgpu: use RAS error address convert api in mca notifier Use the convert interface to simplify code. Signed-off-by: Tao Zhou Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 31 +++++++++++++++---------------- 1 file changed, 15 insertions(+), 16 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 3deb716710e6..ccebd8e2a2d8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -2834,11 +2834,8 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb, struct mce *m = (struct mce *)data; struct amdgpu_device *adev = NULL; uint32_t gpu_id = 0; - uint32_t umc_inst = 0; - uint32_t ch_inst, channel_index = 0; + uint32_t umc_inst = 0, ch_inst = 0; struct ras_err_data err_data = {0, 0, 0, NULL}; - struct eeprom_table_record err_rec; - uint64_t retired_page; /* * If the error was generated in UMC_V2, which belongs to GPU UMCs, @@ -2877,21 +2874,22 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb, dev_info(adev->dev, "Uncorrectable error detected in UMC inst: %d, chan_idx: %d", umc_inst, ch_inst); + err_data.err_addr = + kcalloc(adev->umc.max_ras_err_cnt_per_query, + sizeof(struct eeprom_table_record), GFP_KERNEL); + if(!err_data.err_addr) { + dev_warn(adev->dev, "Failed to alloc memory for " + "umc error address record in mca notifier!\n"); + return NOTIFY_DONE; + } + /* * Translate UMC channel address to Physical address */ - channel_index = - adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num - + ch_inst]; - - retired_page = ADDR_OF_8KB_BLOCK(m->addr) | - ADDR_OF_256B_BLOCK(channel_index) | - OFFSET_IN_256B_BLOCK(m->addr); - - memset(&err_rec, 0x0, sizeof(struct eeprom_table_record)); - err_data.err_addr = &err_rec; - amdgpu_umc_fill_error_record(&err_data, m->addr, - retired_page, channel_index, umc_inst); + if (adev->umc.ras && + adev->umc.ras->convert_ras_error_address) + adev->umc.ras->convert_ras_error_address(adev, + &err_data, 0, ch_inst, umc_inst, m->addr); if (amdgpu_bad_page_threshold != 0) { amdgpu_ras_add_bad_pages(adev, err_data.err_addr, @@ -2899,6 +2897,7 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb, amdgpu_ras_save_bad_pages(adev); } + kfree(err_data.err_addr); return NOTIFY_OK; } -- cgit v1.2.3 From 44420ac5f855f5704d8f939926ed145f99e49e55 Mon Sep 17 00:00:00 2001 From: Tao Zhou Date: Tue, 27 Sep 2022 11:36:46 +0800 Subject: drm/amdgpu: define RAS convert_error_address API Make the code reusable and remove redundant code. Signed-off-by: Tao Zhou Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h | 7 +- drivers/gpu/drm/amd/amdgpu/umc_v6_7.c | 148 +++++++++++++------------------- 3 files changed, 64 insertions(+), 93 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index ccebd8e2a2d8..c2f9970e851c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -2889,7 +2889,7 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb, if (adev->umc.ras && adev->umc.ras->convert_ras_error_address) adev->umc.ras->convert_ras_error_address(adev, - &err_data, 0, ch_inst, umc_inst, m->addr); + &err_data, m->addr, ch_inst, umc_inst); if (amdgpu_bad_page_threshold != 0) { amdgpu_ras_add_bad_pages(adev, err_data.err_addr, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h index 2fb4951a6433..e46439274f3a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h @@ -22,8 +22,6 @@ #define __AMDGPU_UMC_H__ #include "amdgpu_ras.h" -#define UMC_INVALID_ADDR 0x1ULL - /* * (addr / 256) * 4096, the higher 26 bits in ErrorAddr * is the index of 4KB block @@ -54,9 +52,8 @@ struct amdgpu_umc_ras { void (*err_cnt_init)(struct amdgpu_device *adev); bool (*query_ras_poison_mode)(struct amdgpu_device *adev); void (*convert_ras_error_address)(struct amdgpu_device *adev, - struct ras_err_data *err_data, - uint32_t umc_reg_offset, uint32_t ch_inst, - uint32_t umc_inst, uint64_t mca_addr); + struct ras_err_data *err_data, uint64_t err_addr, + uint32_t ch_inst, uint32_t umc_inst); void (*ecc_info_query_ras_error_count)(struct amdgpu_device *adev, void *ras_error_status); void (*ecc_info_query_ras_error_address)(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c index 64d760eb92a3..5d5d031c9e7d 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c +++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c @@ -187,20 +187,51 @@ static void umc_v6_7_ecc_info_query_ras_error_count(struct amdgpu_device *adev, } } +static void umc_v6_7_convert_error_address(struct amdgpu_device *adev, + struct ras_err_data *err_data, uint64_t err_addr, + uint32_t ch_inst, uint32_t umc_inst) +{ + uint32_t channel_index; + uint64_t soc_pa, retired_page, column; + + channel_index = + adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst]; + /* translate umc channel address to soc pa, 3 parts are included */ + soc_pa = ADDR_OF_8KB_BLOCK(err_addr) | + ADDR_OF_256B_BLOCK(channel_index) | + OFFSET_IN_256B_BLOCK(err_addr); + + /* The umc channel bits are not original values, they are hashed */ + SET_CHANNEL_HASH(channel_index, soc_pa); + + /* clear [C4 C3 C2] in soc physical address */ + soc_pa &= ~(0x7ULL << UMC_V6_7_PA_C2_BIT); + + /* loop for all possibilities of [C4 C3 C2] */ + for (column = 0; column < UMC_V6_7_NA_MAP_PA_NUM; column++) { + retired_page = soc_pa | (column << UMC_V6_7_PA_C2_BIT); + dev_info(adev->dev, "Error Address(PA): 0x%llx\n", retired_page); + amdgpu_umc_fill_error_record(err_data, err_addr, + retired_page, channel_index, umc_inst); + + /* shift R14 bit */ + retired_page ^= (0x1ULL << UMC_V6_7_PA_R14_BIT); + dev_info(adev->dev, "Error Address(PA): 0x%llx\n", retired_page); + amdgpu_umc_fill_error_record(err_data, err_addr, + retired_page, channel_index, umc_inst); + } +} + static void umc_v6_7_ecc_info_query_error_address(struct amdgpu_device *adev, struct ras_err_data *err_data, uint32_t ch_inst, uint32_t umc_inst) { - uint64_t mc_umc_status, err_addr, soc_pa, retired_page, column; - uint32_t channel_index; + uint64_t mc_umc_status, err_addr; uint32_t eccinfo_table_idx; struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); eccinfo_table_idx = umc_inst * adev->umc.channel_inst_num + ch_inst; - channel_index = - adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst]; - mc_umc_status = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_status; if (mc_umc_status == 0) @@ -216,30 +247,8 @@ static void umc_v6_7_ecc_info_query_error_address(struct amdgpu_device *adev, err_addr = ras->umc_ecc.ecc[eccinfo_table_idx].mca_umc_addr; err_addr = REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr); - /* translate umc channel address to soc pa, 3 parts are included */ - soc_pa = ADDR_OF_8KB_BLOCK(err_addr) | - ADDR_OF_256B_BLOCK(channel_index) | - OFFSET_IN_256B_BLOCK(err_addr); - - /* The umc channel bits are not original values, they are hashed */ - SET_CHANNEL_HASH(channel_index, soc_pa); - - /* clear [C4 C3 C2] in soc physical address */ - soc_pa &= ~(0x7ULL << UMC_V6_7_PA_C2_BIT); - - /* loop for all possibilities of [C4 C3 C2] */ - for (column = 0; column < UMC_V6_7_NA_MAP_PA_NUM; column++) { - retired_page = soc_pa | (column << UMC_V6_7_PA_C2_BIT); - dev_info(adev->dev, "Error Address(PA): 0x%llx\n", retired_page); - amdgpu_umc_fill_error_record(err_data, err_addr, - retired_page, channel_index, umc_inst); - - /* shift R14 bit */ - retired_page ^= (0x1ULL << UMC_V6_7_PA_R14_BIT); - dev_info(adev->dev, "Error Address(PA): 0x%llx\n", retired_page); - amdgpu_umc_fill_error_record(err_data, err_addr, - retired_page, channel_index, umc_inst); - } + umc_v6_7_convert_error_address(adev, err_data, err_addr, + ch_inst, umc_inst); } } @@ -448,75 +457,40 @@ static void umc_v6_7_query_ras_error_count(struct amdgpu_device *adev, static void umc_v6_7_query_error_address(struct amdgpu_device *adev, struct ras_err_data *err_data, uint32_t umc_reg_offset, uint32_t ch_inst, - uint32_t umc_inst, uint64_t mca_addr) + uint32_t umc_inst) { uint32_t mc_umc_status_addr; - uint32_t channel_index; - uint64_t mc_umc_status = 0, mc_umc_addrt0; - uint64_t err_addr, soc_pa, retired_page, column; + uint64_t mc_umc_status = 0, mc_umc_addrt0, err_addr; - if (mca_addr == UMC_INVALID_ADDR) { - mc_umc_status_addr = - SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0); - mc_umc_addrt0 = - SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_ADDRT0); + mc_umc_status_addr = + SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0); + mc_umc_addrt0 = + SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_ADDRT0); - mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4); + mc_umc_status = RREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4); - if (mc_umc_status == 0) - return; + if (mc_umc_status == 0) + return; - if (!err_data->err_addr) { - /* clear umc status */ - WREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL); - return; - } + if (!err_data->err_addr) { + /* clear umc status */ + WREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL); + return; } - channel_index = - adev->umc.channel_idx_tbl[umc_inst * adev->umc.channel_inst_num + ch_inst]; - /* calculate error address if ue error is detected */ - if ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 && - REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1) || - mca_addr != UMC_INVALID_ADDR) { - if (mca_addr == UMC_INVALID_ADDR) { - err_addr = RREG64_PCIE((mc_umc_addrt0 + umc_reg_offset) * 4); - err_addr = - REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr); - } else { - err_addr = mca_addr; - } + if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 && + REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UECC) == 1) { + err_addr = RREG64_PCIE((mc_umc_addrt0 + umc_reg_offset) * 4); + err_addr = + REG_GET_FIELD(err_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr); - /* translate umc channel address to soc pa, 3 parts are included */ - soc_pa = ADDR_OF_8KB_BLOCK(err_addr) | - ADDR_OF_256B_BLOCK(channel_index) | - OFFSET_IN_256B_BLOCK(err_addr); - - /* The umc channel bits are not original values, they are hashed */ - SET_CHANNEL_HASH(channel_index, soc_pa); - - /* clear [C4 C3 C2] in soc physical address */ - soc_pa &= ~(0x7ULL << UMC_V6_7_PA_C2_BIT); - - /* loop for all possibilities of [C4 C3 C2] */ - for (column = 0; column < UMC_V6_7_NA_MAP_PA_NUM; column++) { - retired_page = soc_pa | (column << UMC_V6_7_PA_C2_BIT); - dev_info(adev->dev, "Error Address(PA): 0x%llx\n", retired_page); - amdgpu_umc_fill_error_record(err_data, err_addr, - retired_page, channel_index, umc_inst); - - /* shift R14 bit */ - retired_page ^= (0x1ULL << UMC_V6_7_PA_R14_BIT); - dev_info(adev->dev, "Error Address(PA): 0x%llx\n", retired_page); - amdgpu_umc_fill_error_record(err_data, err_addr, - retired_page, channel_index, umc_inst); - } + umc_v6_7_convert_error_address(adev, err_data, err_addr, + ch_inst, umc_inst); } /* clear umc status */ - if (mca_addr == UMC_INVALID_ADDR) - WREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL); + WREG64_PCIE((mc_umc_status_addr + umc_reg_offset) * 4, 0x0ULL); } static void umc_v6_7_query_ras_error_address(struct amdgpu_device *adev, @@ -538,7 +512,7 @@ static void umc_v6_7_query_ras_error_address(struct amdgpu_device *adev, umc_v6_7_query_error_address(adev, err_data, umc_reg_offset, ch_inst, - umc_inst, UMC_INVALID_ADDR); + umc_inst); } } @@ -579,5 +553,5 @@ struct amdgpu_umc_ras umc_v6_7_ras = { .query_ras_poison_mode = umc_v6_7_query_ras_poison_mode, .ecc_info_query_ras_error_count = umc_v6_7_ecc_info_query_ras_error_count, .ecc_info_query_ras_error_address = umc_v6_7_ecc_info_query_ras_error_address, - .convert_ras_error_address = umc_v6_7_query_error_address, + .convert_ras_error_address = umc_v6_7_convert_error_address, }; -- cgit v1.2.3 From 38dbbfa57c08b29ef8cf1d3fb3ad639ae819754e Mon Sep 17 00:00:00 2001 From: Tao Zhou Date: Thu, 29 Sep 2022 14:59:11 +0800 Subject: drm/amdgpu: fix coding style issue for mca notifier Fix some issues found by checkpatch script. Signed-off-by: Tao Zhou Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index c2f9970e851c..2dad7aa9a03b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -2877,9 +2877,9 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb, err_data.err_addr = kcalloc(adev->umc.max_ras_err_cnt_per_query, sizeof(struct eeprom_table_record), GFP_KERNEL); - if(!err_data.err_addr) { - dev_warn(adev->dev, "Failed to alloc memory for " - "umc error address record in mca notifier!\n"); + if (!err_data.err_addr) { + dev_warn(adev->dev, + "Failed to alloc memory for umc error record in mca notifier!\n"); return NOTIFY_DONE; } -- cgit v1.2.3 From b98a1648d6616d288e888c6dc6dcd4fa543585b3 Mon Sep 17 00:00:00 2001 From: Victor Zhao Date: Thu, 13 Oct 2022 11:06:33 +0800 Subject: Revert "drm/amdgpu: let mode2 reset fallback to default when failure" This reverts commit dac6b80818ac2353631c5a33d140d8d5508e2957. This commit reverted the AMDGPU_SKIP_MODE2_RESET as it conflicts with the original design of reset handler. Will redesign it. Fixes: dac6b80818ac23 ("drm/amdgpu: let mode2 reset fallback to default when failure") Signed-off-by: Victor Zhao Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 1 - drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 7 +------ drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 1 - drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 1 - drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c | 6 ------ drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h | 3 +-- drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c | 1 - drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c | 1 - drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c | 1 - 9 files changed, 2 insertions(+), 20 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 03bbfaa51cbc..0561812aa0a4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -134,7 +134,6 @@ static void amdgpu_amdkfd_reset_work(struct work_struct *work) reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); - clear_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context.flags); amdgpu_device_gpu_recover(adev, NULL, &reset_context); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index ab8f970b2849..bb73fb420ffc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -5210,7 +5210,6 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, reset_context->job = job; reset_context->hive = hive; - /* * Build list of devices to reset. * In case we are in XGMI hive mode, resort the device list @@ -5337,11 +5336,8 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ amdgpu_ras_resume(adev); } else { r = amdgpu_do_asic_reset(device_list_handle, reset_context); - if (r && r == -EAGAIN) { - set_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context->flags); - adev->asic_reset_res = 0; + if (r && r == -EAGAIN) goto retry; - } if (!r && gpu_reset_for_dev_remove) goto recover_end; @@ -5777,7 +5773,6 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev) reset_context.reset_req_dev = adev; set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags); - set_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context.flags); adev->no_hw_access = true; r = amdgpu_device_pre_asic_reset(adev, &reset_context); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index 46c99331d7f1..cd968e781077 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -72,7 +72,6 @@ static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job) reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); - clear_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context.flags); r = amdgpu_device_gpu_recover(ring->adev, job, &reset_context); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 2dad7aa9a03b..75f1402101f4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1950,7 +1950,6 @@ static void amdgpu_ras_do_recovery(struct work_struct *work) reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); - clear_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context.flags); amdgpu_device_gpu_recover(ras->adev, NULL, &reset_context); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c index 831fb222139c..f778466bb9db 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c @@ -74,9 +74,6 @@ int amdgpu_reset_prepare_hwcontext(struct amdgpu_device *adev, { struct amdgpu_reset_handler *reset_handler = NULL; - if (test_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context->flags)) - return -ENOSYS; - if (adev->reset_cntl && adev->reset_cntl->get_reset_handler) reset_handler = adev->reset_cntl->get_reset_handler( adev->reset_cntl, reset_context); @@ -93,9 +90,6 @@ int amdgpu_reset_perform_reset(struct amdgpu_device *adev, int ret; struct amdgpu_reset_handler *reset_handler = NULL; - if (test_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context->flags)) - return -ENOSYS; - if (adev->reset_cntl) reset_handler = adev->reset_cntl->get_reset_handler( adev->reset_cntl, reset_context); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h index f5318fedf2f0..f4a501ff87d9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h @@ -30,8 +30,7 @@ enum AMDGPU_RESET_FLAGS { AMDGPU_NEED_FULL_RESET = 0, AMDGPU_SKIP_HW_RESET = 1, - AMDGPU_SKIP_MODE2_RESET = 2, - AMDGPU_RESET_FOR_DEVICE_REMOVE = 3, + AMDGPU_RESET_FOR_DEVICE_REMOVE = 2, }; struct amdgpu_reset_context { diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c index a2f04b249132..12906ba74462 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c @@ -290,7 +290,6 @@ flr_done: reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); - clear_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context.flags); amdgpu_device_gpu_recover(adev, NULL, &reset_context); } diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c index a977f0027928..e07757eea7ad 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c @@ -317,7 +317,6 @@ flr_done: reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); - clear_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context.flags); amdgpu_device_gpu_recover(adev, NULL, &reset_context); } diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c index fd14fa9b9cd7..288c414babdf 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c @@ -529,7 +529,6 @@ static void xgpu_vi_mailbox_flr_work(struct work_struct *work) reset_context.method = AMD_RESET_METHOD_NONE; reset_context.reset_req_dev = adev; clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); - clear_bit(AMDGPU_SKIP_MODE2_RESET, &reset_context.flags); amdgpu_device_gpu_recover(adev, NULL, &reset_context); } -- cgit v1.2.3 From 073285efde229ae82d3b853c7f4bcca81f97a55f Mon Sep 17 00:00:00 2001 From: YiPeng Chai Date: Tue, 27 Sep 2022 14:06:42 +0800 Subject: drm/amdgpu: Enable ras support for mp0 v13_0_0 and v13_0_10 V1: Enable ras support for CHIP_IP_DISCOVERY asic type. V2: 1. Change commit comment. 2. Enable ras support for mp0 v13_0_0 and v13_0_10. Signed-off-by: YiPeng Chai Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 75f1402101f4..4a8f73cc4cb5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -2267,6 +2267,16 @@ static int amdgpu_ras_recovery_fini(struct amdgpu_device *adev) static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev) { + if (adev->asic_type == CHIP_IP_DISCOVERY) { + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(13, 0, 0): + case IP_VERSION(13, 0, 10): + return true; + default: + return false; + } + } + return adev->asic_type == CHIP_VEGA10 || adev->asic_type == CHIP_VEGA20 || adev->asic_type == CHIP_ARCTURUS || -- cgit v1.2.3 From 82835055c62fa49b50dc00736743e8f99ed93638 Mon Sep 17 00:00:00 2001 From: YiPeng Chai Date: Wed, 28 Sep 2022 15:52:02 +0800 Subject: drm/amdgpu: Add sriov vf ras support in amdgpu_ras_asic_supported V2: Add sriov vf ras support in amdgpu_ras_asic_supported. Signed-off-by: YiPeng Chai Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 4a8f73cc4cb5..a4b47e1bd111 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -2267,6 +2267,15 @@ static int amdgpu_ras_recovery_fini(struct amdgpu_device *adev) static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev) { + if (amdgpu_sriov_vf(adev)) { + switch (adev->ip_versions[MP0_HWIP][0]) { + case IP_VERSION(13, 0, 2): + return true; + default: + return false; + } + } + if (adev->asic_type == CHIP_IP_DISCOVERY) { switch (adev->ip_versions[MP0_HWIP][0]) { case IP_VERSION(13, 0, 0): @@ -2320,11 +2329,6 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev) !amdgpu_ras_asic_supported(adev)) return; - /* If driver run on sriov guest side, only enable ras for aldebaran */ - if (amdgpu_sriov_vf(adev) && - adev->ip_versions[MP1_HWIP][0] != IP_VERSION(13, 0, 2)) - return; - if (!adev->gmc.xgmi.connected_to_cpu) { if (amdgpu_atomfirmware_mem_ecc_supported(adev)) { dev_info(adev->dev, "MEM ECC is active.\n"); -- cgit v1.2.3 From 6c0ca748205dc815505c6de79ecf565953390b66 Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Fri, 14 Oct 2022 15:17:43 +0800 Subject: drm/amdgpu: move convert_error_address out of umc_ras RAS error address translation algorithm is common across dGPU and A + A platform as along as the SOC integrates the same generation of UMC IP. UMC RAS is managed by x86 MCA on A + A platform, umc_ras in GPU driver is not initialized at all on A + A platform. In such case, any umc_ras callback implemented for dGPU config shouldn't be invoked from A + A specific callback. The change moves convert_error_address out of dGPU umc_ras structure and makes it share between A + A and dGPU config. Signed-off-by: Hawking Zhang Reviewed-by: Stanley Yang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 16 ++++++++++++---- drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h | 3 --- drivers/gpu/drm/amd/amdgpu/umc_v6_7.c | 7 +++---- drivers/gpu/drm/amd/amdgpu/umc_v6_7.h | 4 +++- 4 files changed, 18 insertions(+), 12 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index a4b47e1bd111..21a47f2bb87b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -36,6 +36,7 @@ #include "ivsrcid/nbio/irqsrcs_nbif_7_4.h" #include "atom.h" #include "amdgpu_reset.h" +#include "umc_v6_7.h" #ifdef CONFIG_X86_MCE_AMD #include @@ -2899,10 +2900,17 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb, /* * Translate UMC channel address to Physical address */ - if (adev->umc.ras && - adev->umc.ras->convert_ras_error_address) - adev->umc.ras->convert_ras_error_address(adev, - &err_data, m->addr, ch_inst, umc_inst); + switch (adev->ip_versions[UMC_HWIP][0]) { + case IP_VERSION(6, 7, 0): + umc_v6_7_convert_error_address(adev, + &err_data, m->addr, ch_inst, umc_inst); + break; + default: + dev_warn(adev->dev, + "UMC address to Physical address translation is not supported\n"); + kfree(err_data.err_addr); + return NOTIFY_DONE; + } if (amdgpu_bad_page_threshold != 0) { amdgpu_ras_add_bad_pages(adev, err_data.err_addr, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h index e46439274f3a..3629d8f292ef 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h @@ -51,9 +51,6 @@ struct amdgpu_umc_ras { struct amdgpu_ras_block_object ras_block; void (*err_cnt_init)(struct amdgpu_device *adev); bool (*query_ras_poison_mode)(struct amdgpu_device *adev); - void (*convert_ras_error_address)(struct amdgpu_device *adev, - struct ras_err_data *err_data, uint64_t err_addr, - uint32_t ch_inst, uint32_t umc_inst); void (*ecc_info_query_ras_error_count)(struct amdgpu_device *adev, void *ras_error_status); void (*ecc_info_query_ras_error_address)(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c index 5d5d031c9e7d..72fd963f178b 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c +++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.c @@ -187,9 +187,9 @@ static void umc_v6_7_ecc_info_query_ras_error_count(struct amdgpu_device *adev, } } -static void umc_v6_7_convert_error_address(struct amdgpu_device *adev, - struct ras_err_data *err_data, uint64_t err_addr, - uint32_t ch_inst, uint32_t umc_inst) +void umc_v6_7_convert_error_address(struct amdgpu_device *adev, + struct ras_err_data *err_data, uint64_t err_addr, + uint32_t ch_inst, uint32_t umc_inst) { uint32_t channel_index; uint64_t soc_pa, retired_page, column; @@ -553,5 +553,4 @@ struct amdgpu_umc_ras umc_v6_7_ras = { .query_ras_poison_mode = umc_v6_7_query_ras_poison_mode, .ecc_info_query_ras_error_count = umc_v6_7_ecc_info_query_ras_error_count, .ecc_info_query_ras_error_address = umc_v6_7_ecc_info_query_ras_error_address, - .convert_ras_error_address = umc_v6_7_convert_error_address, }; diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.h b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.h index fe41ed2f5945..105245d5b6e5 100644 --- a/drivers/gpu/drm/amd/amdgpu/umc_v6_7.h +++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_7.h @@ -71,5 +71,7 @@ extern const uint32_t umc_v6_7_channel_idx_tbl_second[UMC_V6_7_UMC_INSTANCE_NUM][UMC_V6_7_CHANNEL_INSTANCE_NUM]; extern const uint32_t umc_v6_7_channel_idx_tbl_first[UMC_V6_7_UMC_INSTANCE_NUM][UMC_V6_7_CHANNEL_INSTANCE_NUM]; - +void umc_v6_7_convert_error_address(struct amdgpu_device *adev, + struct ras_err_data *err_data, uint64_t err_addr, + uint32_t ch_inst, uint32_t umc_inst); #endif -- cgit v1.2.3 From 24b822928b5139b85ee9a818a65e343b7e3bb4fe Mon Sep 17 00:00:00 2001 From: Tao Zhou Date: Mon, 17 Oct 2022 18:26:17 +0800 Subject: drm/amdgpu: use page retirement API in MCA notifier Make the code more readable. Signed-off-by: Tao Zhou Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 36 +++------------------------------ 1 file changed, 3 insertions(+), 33 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 21a47f2bb87b..28463b47ce33 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -36,7 +36,6 @@ #include "ivsrcid/nbio/irqsrcs_nbif_7_4.h" #include "atom.h" #include "amdgpu_reset.h" -#include "umc_v6_7.h" #ifdef CONFIG_X86_MCE_AMD #include @@ -2849,7 +2848,6 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb, struct amdgpu_device *adev = NULL; uint32_t gpu_id = 0; uint32_t umc_inst = 0, ch_inst = 0; - struct ras_err_data err_data = {0, 0, 0, NULL}; /* * If the error was generated in UMC_V2, which belongs to GPU UMCs, @@ -2888,38 +2886,10 @@ static int amdgpu_bad_page_notifier(struct notifier_block *nb, dev_info(adev->dev, "Uncorrectable error detected in UMC inst: %d, chan_idx: %d", umc_inst, ch_inst); - err_data.err_addr = - kcalloc(adev->umc.max_ras_err_cnt_per_query, - sizeof(struct eeprom_table_record), GFP_KERNEL); - if (!err_data.err_addr) { - dev_warn(adev->dev, - "Failed to alloc memory for umc error record in mca notifier!\n"); - return NOTIFY_DONE; - } - - /* - * Translate UMC channel address to Physical address - */ - switch (adev->ip_versions[UMC_HWIP][0]) { - case IP_VERSION(6, 7, 0): - umc_v6_7_convert_error_address(adev, - &err_data, m->addr, ch_inst, umc_inst); - break; - default: - dev_warn(adev->dev, - "UMC address to Physical address translation is not supported\n"); - kfree(err_data.err_addr); + if (!amdgpu_umc_page_retirement_mca(adev, m->addr, ch_inst, umc_inst)) + return NOTIFY_OK; + else return NOTIFY_DONE; - } - - if (amdgpu_bad_page_threshold != 0) { - amdgpu_ras_add_bad_pages(adev, err_data.err_addr, - err_data.err_addr_cnt); - amdgpu_ras_save_bad_pages(adev); - } - - kfree(err_data.err_addr); - return NOTIFY_OK; } static struct notifier_block amdgpu_bad_page_nb = { -- cgit v1.2.3 From 1ed0e176902483e67cd02530d387a7551b0e99a4 Mon Sep 17 00:00:00 2001 From: Tao Zhou Date: Tue, 18 Oct 2022 10:31:09 +0800 Subject: drm/amdgpu: remove ras_error_status parameter for UMC poison handler Make the code simpler. Signed-off-by: Tao Zhou Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 4 +--- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 3 +-- drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c | 13 +++++-------- drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h | 4 +--- 4 files changed, 8 insertions(+), 16 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 0561812aa0a4..37db39ba8718 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -753,9 +753,7 @@ bool amdgpu_amdkfd_have_atomics_support(struct amdgpu_device *adev) void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev, bool reset) { - struct ras_err_data err_data = {0, 0, 0, NULL}; - - amdgpu_umc_poison_handler(adev, &err_data, reset); + amdgpu_umc_poison_handler(adev, reset); } bool amdgpu_amdkfd_ras_query_utcl2_poison_status(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 28463b47ce33..693bce07eb46 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -1561,7 +1561,6 @@ static void amdgpu_ras_interrupt_poison_consumption_handler(struct ras_manager * { bool poison_stat = false; struct amdgpu_device *adev = obj->adev; - struct ras_err_data err_data = {0, 0, 0, NULL}; struct amdgpu_ras_block_object *block_obj = amdgpu_ras_get_ras_block(adev, obj->head.block, 0); @@ -1584,7 +1583,7 @@ static void amdgpu_ras_interrupt_poison_consumption_handler(struct ras_manager * } if (!adev->gmc.xgmi.connected_to_cpu) - amdgpu_umc_poison_handler(adev, &err_data, false); + amdgpu_umc_poison_handler(adev, false); if (block_obj->hw_ops->handle_poison_consumption) poison_stat = block_obj->hw_ops->handle_poison_consumption(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c index 758942150c09..f76c19fc0392 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c @@ -165,25 +165,22 @@ static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev, return AMDGPU_RAS_SUCCESS; } -int amdgpu_umc_poison_handler(struct amdgpu_device *adev, - void *ras_error_status, - bool reset) +int amdgpu_umc_poison_handler(struct amdgpu_device *adev, bool reset) { int ret = AMDGPU_RAS_SUCCESS; if (!adev->gmc.xgmi.connected_to_cpu) { - struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status; + struct ras_err_data err_data = {0, 0, 0, NULL}; struct ras_common_if head = { .block = AMDGPU_RAS_BLOCK__UMC, }; struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head); - ret = - amdgpu_umc_do_page_retirement(adev, ras_error_status, NULL, reset); + ret = amdgpu_umc_do_page_retirement(adev, &err_data, NULL, reset); if (ret == AMDGPU_RAS_SUCCESS && obj) { - obj->err_data.ue_count += err_data->ue_count; - obj->err_data.ce_count += err_data->ce_count; + obj->err_data.ue_count += err_data.ue_count; + obj->err_data.ce_count += err_data.ce_count; } } else if (reset) { /* MCA poison handler is only responsible for GPU reset, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h index 659a10de29c9..a6951160f13a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h @@ -83,9 +83,7 @@ struct amdgpu_umc { }; int amdgpu_umc_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block); -int amdgpu_umc_poison_handler(struct amdgpu_device *adev, - void *ras_error_status, - bool reset); +int amdgpu_umc_poison_handler(struct amdgpu_device *adev, bool reset); int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry); -- cgit v1.2.3