--- /dev/null
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "amdgpu.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_xgmi.h"
+#include "ras_sys.h"
+#include "amdgpu_ras_mgr.h"
+#include "amdgpu_ras_cmd.h"
+#include "amdgpu_ras_process.h"
+#include "amdgpu_ras_eeprom_i2c.h"
+#include "amdgpu_ras_mp1_v13_0.h"
+#include "amdgpu_ras_nbio_v7_9.h"
+
+#define MAX_SOCKET_NUM_PER_HIVE 8
+#define MAX_AID_NUM_PER_SOCKET 4
+#define MAX_XCD_NUM_PER_AID 2
+
+/* typical ECC bad page rate is 1 bad page per 100MB VRAM */
+#define ESTIMATE_BAD_PAGE_THRESHOLD(size) ((size)/(100 * 1024 * 1024ULL))
+
+#define COUNT_BAD_PAGE_THRESHOLD(size) (((size) >> 21) << 4)
+
+/* Reserve 8 physical dram row for possible retirement.
+ * In worst cases, it will lose 8 * 2MB memory in vram domain
+ */
+#define RAS_RESERVED_VRAM_SIZE_DEFAULT (16ULL << 20)
+
+
+static void ras_mgr_init_event_mgr(struct ras_event_manager *mgr)
+{
+ struct ras_event_state *event_state;
+ int i;
+
+ memset(mgr, 0, sizeof(*mgr));
+ atomic64_set(&mgr->seqno, 0);
+
+ for (i = 0; i < ARRAY_SIZE(mgr->event_state); i++) {
+ event_state = &mgr->event_state[i];
+ event_state->last_seqno = RAS_EVENT_INVALID_ID;
+ atomic64_set(&event_state->count, 0);
+ }
+}
+
+static void amdgpu_ras_mgr_init_event_mgr(struct ras_core_context *ras_core)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)ras_core->dev;
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+ struct ras_event_manager *event_mgr;
+ struct amdgpu_hive_info *hive;
+
+ hive = amdgpu_get_xgmi_hive(adev);
+ event_mgr = hive ? &hive->event_mgr : &ras_mgr->ras_event_mgr;
+
+ /* init event manager with node 0 on xgmi system */
+ if (!amdgpu_reset_in_recovery(adev)) {
+ if (!hive || adev->gmc.xgmi.node_id == 0)
+ ras_mgr_init_event_mgr(event_mgr);
+ }
+
+ if (hive)
+ amdgpu_put_xgmi_hive(hive);
+}
+
+static int amdgpu_ras_mgr_init_aca_config(struct amdgpu_device *adev,
+ struct ras_core_config *config)
+{
+ struct ras_aca_config *aca_cfg = &config->aca_cfg;
+
+ aca_cfg->socket_num_per_hive = MAX_SOCKET_NUM_PER_HIVE;
+ aca_cfg->aid_num_per_socket = MAX_AID_NUM_PER_SOCKET;
+ aca_cfg->xcd_num_per_aid = MAX_XCD_NUM_PER_AID;
+
+ return 0;
+}
+
+static int amdgpu_ras_mgr_init_eeprom_config(struct amdgpu_device *adev,
+ struct ras_core_config *config)
+{
+ struct ras_eeprom_config *eeprom_cfg = &config->eeprom_cfg;
+
+ eeprom_cfg->eeprom_sys_fn = &amdgpu_ras_eeprom_i2c_sys_func;
+ eeprom_cfg->eeprom_i2c_adapter = adev->pm.ras_eeprom_i2c_bus;
+ if (eeprom_cfg->eeprom_i2c_adapter) {
+ const struct i2c_adapter_quirks *quirks =
+ ((struct i2c_adapter *)eeprom_cfg->eeprom_i2c_adapter)->quirks;
+
+ if (quirks) {
+ eeprom_cfg->max_i2c_read_len = quirks->max_read_len;
+ eeprom_cfg->max_i2c_write_len = quirks->max_write_len;
+ }
+ }
+
+ /*
+ * amdgpu_bad_page_threshold is used to config
+ * the threshold for the number of bad pages.
+ * -1: Threshold is set to default value
+ * Driver will issue a warning message when threshold is reached
+ * and continue runtime services.
+ * 0: Disable bad page retirement
+ * Driver will not retire bad pages
+ * which is intended for debugging purpose.
+ * -2: Threshold is determined by a formula
+ * that assumes 1 bad page per 100M of local memory.
+ * Driver will continue runtime services when threhold is reached.
+ * 0 < threshold < max number of bad page records in EEPROM,
+ * A user-defined threshold is set
+ * Driver will halt runtime services when this custom threshold is reached.
+ */
+ if (amdgpu_bad_page_threshold == NONSTOP_OVER_THRESHOLD)
+ eeprom_cfg->eeprom_record_threshold_count =
+ ESTIMATE_BAD_PAGE_THRESHOLD(adev->gmc.mc_vram_size);
+ else if (amdgpu_bad_page_threshold == WARN_NONSTOP_OVER_THRESHOLD)
+ eeprom_cfg->eeprom_record_threshold_count =
+ COUNT_BAD_PAGE_THRESHOLD(RAS_RESERVED_VRAM_SIZE_DEFAULT);
+ else
+ eeprom_cfg->eeprom_record_threshold_count = amdgpu_bad_page_threshold;
+
+ eeprom_cfg->eeprom_record_threshold_config = amdgpu_bad_page_threshold;
+
+ return 0;
+}
+
+static int amdgpu_ras_mgr_init_mp1_config(struct amdgpu_device *adev,
+ struct ras_core_config *config)
+{
+ struct ras_mp1_config *mp1_cfg = &config->mp1_cfg;
+ int ret = 0;
+
+ switch (config->mp1_ip_version) {
+ case IP_VERSION(13, 0, 6):
+ case IP_VERSION(13, 0, 14):
+ case IP_VERSION(13, 0, 12):
+ mp1_cfg->mp1_sys_fn = &amdgpu_ras_mp1_sys_func_v13_0;
+ break;
+ default:
+ RAS_DEV_ERR(adev,
+ "The mp1(0x%x) ras config is not right!\n",
+ config->mp1_ip_version);
+ ret = -EINVAL;
+ break;
+ }
+
+ return ret;
+}
+
+static int amdgpu_ras_mgr_init_nbio_config(struct amdgpu_device *adev,
+ struct ras_core_config *config)
+{
+ struct ras_nbio_config *nbio_cfg = &config->nbio_cfg;
+ int ret = 0;
+
+ switch (config->nbio_ip_version) {
+ case IP_VERSION(7, 9, 0):
+ nbio_cfg->nbio_sys_fn = &amdgpu_ras_nbio_sys_func_v7_9;
+ break;
+ default:
+ RAS_DEV_ERR(adev,
+ "The nbio(0x%x) ras config is not right!\n",
+ config->mp1_ip_version);
+ ret = -EINVAL;
+ break;
+ }
+
+ return ret;
+}
+
+static int amdgpu_ras_mgr_get_ras_psp_system_status(struct ras_core_context *ras_core,
+ struct ras_psp_sys_status *status)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)ras_core->dev;
+ struct ta_context *context = &adev->psp.ras_context.context;
+
+ status->initialized = context->initialized;
+ status->session_id = context->session_id;
+ status->psp_cmd_mutex = &adev->psp.mutex;
+
+ return 0;
+}
+
+static int amdgpu_ras_mgr_get_ras_ta_init_param(struct ras_core_context *ras_core,
+ struct ras_ta_init_param *ras_ta_param)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)ras_core->dev;
+ uint32_t nps_mode;
+
+ if (amdgpu_ras_is_poison_mode_supported(adev))
+ ras_ta_param->poison_mode_en = 1;
+
+ if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu)
+ ras_ta_param->dgpu_mode = 1;
+
+ ras_ta_param->xcc_mask = adev->gfx.xcc_mask;
+ ras_ta_param->channel_dis_num = hweight32(adev->gmc.m_half_use) * 2;
+
+ ras_ta_param->active_umc_mask = adev->umc.active_mask;
+
+ if (!amdgpu_ras_mgr_get_curr_nps_mode(adev, &nps_mode))
+ ras_ta_param->nps_mode = nps_mode;
+
+ return 0;
+}
+
+const struct ras_psp_sys_func amdgpu_ras_psp_sys_func = {
+ .get_ras_psp_system_status = amdgpu_ras_mgr_get_ras_psp_system_status,
+ .get_ras_ta_init_param = amdgpu_ras_mgr_get_ras_ta_init_param,
+};
+
+static int amdgpu_ras_mgr_init_psp_config(struct amdgpu_device *adev,
+ struct ras_core_config *config)
+{
+ struct ras_psp_config *psp_cfg = &config->psp_cfg;
+
+ psp_cfg->psp_sys_fn = &amdgpu_ras_psp_sys_func;
+
+ return 0;
+}
+
+static int amdgpu_ras_mgr_init_umc_config(struct amdgpu_device *adev,
+ struct ras_core_config *config)
+{
+ struct ras_umc_config *umc_cfg = &config->umc_cfg;
+
+ umc_cfg->umc_vram_type = adev->gmc.vram_type;
+
+ return 0;
+}
+
+static struct ras_core_context *amdgpu_ras_mgr_create_ras_core(struct amdgpu_device *adev)
+{
+ struct ras_core_config init_config;
+
+ memset(&init_config, 0, sizeof(init_config));
+
+ init_config.umc_ip_version = amdgpu_ip_version(adev, UMC_HWIP, 0);
+ init_config.mp1_ip_version = amdgpu_ip_version(adev, MP1_HWIP, 0);
+ init_config.gfx_ip_version = amdgpu_ip_version(adev, GC_HWIP, 0);
+ init_config.nbio_ip_version = amdgpu_ip_version(adev, NBIO_HWIP, 0);
+ init_config.psp_ip_version = amdgpu_ip_version(adev, MP1_HWIP, 0);
+
+ if (init_config.umc_ip_version == IP_VERSION(12, 0, 0))
+ init_config.aca_ip_version = IP_VERSION(1, 0, 0);
+
+ init_config.sys_fn = &amdgpu_ras_sys_fn;
+ init_config.ras_eeprom_supported = true;
+ init_config.poison_supported =
+ amdgpu_ras_is_poison_mode_supported(adev);
+
+ amdgpu_ras_mgr_init_aca_config(adev, &init_config);
+ amdgpu_ras_mgr_init_eeprom_config(adev, &init_config);
+ amdgpu_ras_mgr_init_mp1_config(adev, &init_config);
+ amdgpu_ras_mgr_init_nbio_config(adev, &init_config);
+ amdgpu_ras_mgr_init_psp_config(adev, &init_config);
+ amdgpu_ras_mgr_init_umc_config(adev, &init_config);
+
+ return ras_core_create(&init_config);
+}
+
+static int amdgpu_ras_mgr_sw_init(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+ struct amdgpu_ras_mgr *ras_mgr;
+ int ret = 0;
+
+ ras_mgr = kzalloc(sizeof(*ras_mgr), GFP_KERNEL);
+ if (!ras_mgr)
+ return -EINVAL;
+
+ con->ras_mgr = ras_mgr;
+ ras_mgr->adev = adev;
+
+ ras_mgr->ras_core = amdgpu_ras_mgr_create_ras_core(adev);
+ if (!ras_mgr->ras_core) {
+ RAS_DEV_ERR(adev, "Failed to create ras core!\n");
+ ret = -EINVAL;
+ goto err;
+ }
+
+ ras_mgr->ras_core->dev = adev;
+
+ amdgpu_ras_process_init(adev);
+ ras_core_sw_init(ras_mgr->ras_core);
+ amdgpu_ras_mgr_init_event_mgr(ras_mgr->ras_core);
+ return 0;
+
+err:
+ kfree(ras_mgr);
+ return ret;
+}
+
+static int amdgpu_ras_mgr_sw_fini(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+ struct amdgpu_ras_mgr *ras_mgr = (struct amdgpu_ras_mgr *)con->ras_mgr;
+
+ if (!ras_mgr)
+ return 0;
+
+ amdgpu_ras_process_fini(adev);
+ ras_core_sw_fini(ras_mgr->ras_core);
+ ras_core_destroy(ras_mgr->ras_core);
+ ras_mgr->ras_core = NULL;
+
+ kfree(con->ras_mgr);
+ con->ras_mgr = NULL;
+
+ return 0;
+}
+
+static int amdgpu_ras_mgr_hw_init(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+ int ret;
+
+ if (!ras_mgr || !ras_mgr->ras_core)
+ return -EINVAL;
+
+ ret = ras_core_hw_init(ras_mgr->ras_core);
+ if (ret) {
+ RAS_DEV_ERR(adev, "Failed to initialize ras core!\n");
+ return ret;
+ }
+
+ ras_mgr->ras_is_ready = true;
+
+ RAS_DEV_INFO(adev, "AMDGPU RAS Is Ready.\n");
+ return 0;
+}
+
+static int amdgpu_ras_mgr_hw_fini(struct amdgpu_ip_block *ip_block)
+{
+ struct amdgpu_device *adev = ip_block->adev;
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (!ras_mgr || !ras_mgr->ras_core)
+ return -EINVAL;
+
+ ras_core_hw_fini(ras_mgr->ras_core);
+
+ ras_mgr->ras_is_ready = false;
+
+ return 0;
+}
+
+struct amdgpu_ras_mgr *amdgpu_ras_mgr_get_context(struct amdgpu_device *adev)
+{
+ if (!adev || !adev->psp.ras_context.ras)
+ return NULL;
+
+ return (struct amdgpu_ras_mgr *)adev->psp.ras_context.ras->ras_mgr;
+}
+
+static const struct amd_ip_funcs ras_v1_0_ip_funcs = {
+ .name = "ras_v1_0",
+ .sw_init = amdgpu_ras_mgr_sw_init,
+ .sw_fini = amdgpu_ras_mgr_sw_fini,
+ .hw_init = amdgpu_ras_mgr_hw_init,
+ .hw_fini = amdgpu_ras_mgr_hw_fini,
+};
+
+int amdgpu_enable_unified_ras(struct amdgpu_device *adev, bool enable)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (!ras_mgr || !ras_mgr->ras_core)
+ return -EPERM;
+
+ if (amdgpu_sriov_vf(adev))
+ return -EPERM;
+
+ RAS_DEV_INFO(adev, "Enable amdgpu unified ras!");
+ return ras_core_set_status(ras_mgr->ras_core, enable);
+}
+
+bool amdgpu_unified_ras_enabled(struct amdgpu_device *adev)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (!ras_mgr || !ras_mgr->ras_core)
+ return false;
+
+ if (amdgpu_sriov_vf(adev))
+ return false;
+
+ return ras_core_is_enabled(ras_mgr->ras_core);
+}
+
+static bool amdgpu_ras_mgr_is_ready(struct amdgpu_device *adev)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (ras_mgr && ras_mgr->ras_core && ras_mgr->ras_is_ready &&
+ ras_core_is_ready(ras_mgr->ras_core))
+ return true;
+
+ return false;
+}
+
+int amdgpu_ras_mgr_handle_fatal_interrupt(struct amdgpu_device *adev, void *data)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (!amdgpu_ras_mgr_is_ready(adev))
+ return -EPERM;
+
+ return ras_core_handle_nbio_irq(ras_mgr->ras_core, data);
+}
+
+uint64_t amdgpu_ras_mgr_gen_ras_event_seqno(struct amdgpu_device *adev,
+ enum ras_seqno_type seqno_type)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+ int ret;
+ uint64_t seq_no;
+
+ if (!amdgpu_ras_mgr_is_ready(adev) ||
+ (seqno_type >= RAS_SEQNO_TYPE_COUNT_MAX))
+ return 0;
+
+ seq_no = ras_core_gen_seqno(ras_mgr->ras_core, seqno_type);
+
+ if ((seqno_type == RAS_SEQNO_TYPE_DE) ||
+ (seqno_type == RAS_SEQNO_TYPE_POISON_CONSUMPTION)) {
+ ret = ras_core_put_seqno(ras_mgr->ras_core, seqno_type, seq_no);
+ if (ret)
+ RAS_DEV_WARN(adev, "There are too many ras interrupts!");
+ }
+
+ return seq_no;
+}
+
+int amdgpu_ras_mgr_handle_controller_interrupt(struct amdgpu_device *adev, void *data)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+ struct ras_ih_info *ih_info = (struct ras_ih_info *)data;
+ uint64_t seq_no = 0;
+ int ret = 0;
+
+ if (!amdgpu_ras_mgr_is_ready(adev))
+ return -EPERM;
+
+ if (ih_info && (ih_info->block == AMDGPU_RAS_BLOCK__UMC)) {
+ if (ras_mgr->ras_core->poison_supported) {
+ seq_no = amdgpu_ras_mgr_gen_ras_event_seqno(adev, RAS_SEQNO_TYPE_DE);
+ RAS_DEV_INFO(adev,
+ "{%llu} RAS poison is created, no user action is needed.\n",
+ seq_no);
+ }
+
+ ret = amdgpu_ras_process_handle_umc_interrupt(adev, ih_info);
+ } else if (ras_mgr->ras_core->poison_supported) {
+ ret = amdgpu_ras_process_handle_unexpected_interrupt(adev, ih_info);
+ } else {
+ RAS_DEV_WARN(adev,
+ "No RAS interrupt handler for non-UMC block with poison disabled.\n");
+ }
+
+ return ret;
+}
+
+int amdgpu_ras_mgr_handle_consumer_interrupt(struct amdgpu_device *adev, void *data)
+{
+ if (!amdgpu_ras_mgr_is_ready(adev))
+ return -EPERM;
+
+ return amdgpu_ras_process_handle_consumption_interrupt(adev, data);
+}
+
+int amdgpu_ras_mgr_update_ras_ecc(struct amdgpu_device *adev)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (!amdgpu_ras_mgr_is_ready(adev))
+ return -EPERM;
+
+ return ras_core_update_ecc_info(ras_mgr->ras_core);
+}
+
+int amdgpu_ras_mgr_reset_gpu(struct amdgpu_device *adev, uint32_t flags)
+{
+ struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+ if (!amdgpu_ras_mgr_is_ready(adev))
+ return -EPERM;
+
+ con->gpu_reset_flags |= flags;
+ return amdgpu_ras_reset_gpu(adev);
+}
+
+bool amdgpu_ras_mgr_check_eeprom_safety_watermark(struct amdgpu_device *adev)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (!amdgpu_ras_mgr_is_ready(adev))
+ return false;
+
+ return ras_eeprom_check_safety_watermark(ras_mgr->ras_core);
+}
+
+int amdgpu_ras_mgr_get_curr_nps_mode(struct amdgpu_device *adev,
+ uint32_t *nps_mode)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+ uint32_t mode;
+
+ if (!amdgpu_ras_mgr_is_ready(adev))
+ return -EINVAL;
+
+ mode = ras_core_get_curr_nps_mode(ras_mgr->ras_core);
+ if (!mode || mode > AMDGPU_NPS8_PARTITION_MODE)
+ return -EINVAL;
+
+ *nps_mode = mode;
+
+ return 0;
+}
+
+bool amdgpu_ras_mgr_check_retired_addr(struct amdgpu_device *adev,
+ uint64_t addr)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (!amdgpu_ras_mgr_is_ready(adev))
+ return false;
+
+ return ras_umc_check_retired_addr(ras_mgr->ras_core, addr);
+}
+
+bool amdgpu_ras_mgr_is_rma(struct amdgpu_device *adev)
+{
+ struct amdgpu_ras_mgr *ras_mgr = amdgpu_ras_mgr_get_context(adev);
+
+ if (!ras_mgr || !ras_mgr->ras_core || !ras_mgr->ras_is_ready)
+ return false;
+
+ return ras_core_gpu_is_rma(ras_mgr->ras_core);
+}
projects / thirdparty / kernel / linux.git / commitdiff
