*/
#include <linux/power_supply.h>
#include <linux/kthread.h>
+#include <linux/module.h>
#include <linux/console.h>
#include <linux/slab.h>
-#include <drm/drmP.h>
+
#include <drm/drm_atomic_helper.h>
#include <drm/drm_probe_helper.h>
#include <drm/amdgpu_drm.h>
#endif
#include "vi.h"
#include "soc15.h"
+#include "nv.h"
#include "bif/bif_4_1_d.h"
#include <linux/pci.h>
#include <linux/firmware.h>
#include "amdgpu_xgmi.h"
#include "amdgpu_ras.h"
+#include "amdgpu_pmu.h"
MODULE_FIRMWARE("amdgpu/vega10_gpu_info.bin");
MODULE_FIRMWARE("amdgpu/vega12_gpu_info.bin");
MODULE_FIRMWARE("amdgpu/raven_gpu_info.bin");
MODULE_FIRMWARE("amdgpu/picasso_gpu_info.bin");
MODULE_FIRMWARE("amdgpu/raven2_gpu_info.bin");
+MODULE_FIRMWARE("amdgpu/navi10_gpu_info.bin");
#define AMDGPU_RESUME_MS 2000
"VEGA12",
"VEGA20",
"RAVEN",
+ "NAVI10",
"LAST",
};
+/**
+ * DOC: pcie_replay_count
+ *
+ * The amdgpu driver provides a sysfs API for reporting the total number
+ * of PCIe replays (NAKs)
+ * The file pcie_replay_count is used for this and returns the total
+ * number of replays as a sum of the NAKs generated and NAKs received
+ */
+
+static ssize_t amdgpu_device_get_pcie_replay_count(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct drm_device *ddev = dev_get_drvdata(dev);
+ struct amdgpu_device *adev = ddev->dev_private;
+ uint64_t cnt = amdgpu_asic_get_pcie_replay_count(adev);
+
+ return snprintf(buf, PAGE_SIZE, "%llu\n", cnt);
+}
+
+static DEVICE_ATTR(pcie_replay_count, S_IRUGO,
+ amdgpu_device_get_pcie_replay_count, NULL);
+
static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev);
/**
} else {
tmp = RREG32(reg);
tmp &= ~and_mask;
- tmp |= or_mask;
+ if (adev->family >= AMDGPU_FAMILY_AI)
+ tmp |= (or_mask & and_mask);
+ else
+ tmp |= or_mask;
}
WREG32(reg, tmp);
}
* Validates certain module parameters and updates
* the associated values used by the driver (all asics).
*/
-static void amdgpu_device_check_arguments(struct amdgpu_device *adev)
+static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
{
+ int ret = 0;
+
if (amdgpu_sched_jobs < 4) {
dev_warn(adev->dev, "sched jobs (%d) must be at least 4\n",
amdgpu_sched_jobs);
amdgpu_device_check_block_size(adev);
- if (amdgpu_vram_page_split != -1 && (amdgpu_vram_page_split < 16 ||
- !is_power_of_2(amdgpu_vram_page_split))) {
- dev_warn(adev->dev, "invalid VRAM page split (%d)\n",
- amdgpu_vram_page_split);
- amdgpu_vram_page_split = 1024;
- }
-
- if (amdgpu_lockup_timeout == 0) {
- dev_warn(adev->dev, "lockup_timeout msut be > 0, adjusting to 10000\n");
- amdgpu_lockup_timeout = 10000;
+ ret = amdgpu_device_get_job_timeout_settings(adev);
+ if (ret) {
+ dev_err(adev->dev, "invalid lockup_timeout parameter syntax\n");
+ return ret;
}
adev->firmware.load_type = amdgpu_ucode_get_load_type(adev, amdgpu_fw_load_type);
+
+ return ret;
}
/**
else
chip_name = "raven";
break;
+ case CHIP_NAVI10:
+ chip_name = "navi10";
+ break;
}
snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_gpu_info.bin", chip_name);
adev->gfx.cu_info.max_scratch_slots_per_cu =
le32_to_cpu(gpu_info_fw->gc_max_scratch_slots_per_cu);
adev->gfx.cu_info.lds_size = le32_to_cpu(gpu_info_fw->gc_lds_size);
+ if (hdr->version_minor >= 1) {
+ const struct gpu_info_firmware_v1_1 *gpu_info_fw =
+ (const struct gpu_info_firmware_v1_1 *)(adev->firmware.gpu_info_fw->data +
+ le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+ adev->gfx.config.num_sc_per_sh =
+ le32_to_cpu(gpu_info_fw->num_sc_per_sh);
+ adev->gfx.config.num_packer_per_sc =
+ le32_to_cpu(gpu_info_fw->num_packer_per_sc);
+ }
+#ifdef CONFIG_DRM_AMD_DC_DCN2_0
+ if (hdr->version_minor == 2) {
+ const struct gpu_info_firmware_v1_2 *gpu_info_fw =
+ (const struct gpu_info_firmware_v1_2 *)(adev->firmware.gpu_info_fw->data +
+ le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+ adev->dm.soc_bounding_box = &gpu_info_fw->soc_bounding_box;
+ }
+#endif
break;
}
default:
if (r)
return r;
break;
+ case CHIP_NAVI10:
+ adev->family = AMDGPU_FAMILY_NV;
+
+ r = nv_set_ip_blocks(adev);
+ if (r)
+ return r;
+ break;
default:
/* FIXME: not supported yet */
return -EINVAL;
r = amdgpu_virt_request_full_gpu(adev, true);
if (r)
return -EAGAIN;
+
+ /* query the reg access mode at the very beginning */
+ amdgpu_virt_init_reg_access_mode(adev);
}
adev->pm.pp_feature = amdgpu_pp_feature_mask;
adev->ip_blocks[i].status.valid = true;
}
}
+ /* get the vbios after the asic_funcs are set up */
+ if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON) {
+ /* Read BIOS */
+ if (!amdgpu_get_bios(adev))
+ return -EINVAL;
+
+ r = amdgpu_atombios_init(adev);
+ if (r) {
+ dev_err(adev->dev, "amdgpu_atombios_init failed\n");
+ amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, 0, 0);
+ return r;
+ }
+ }
}
adev->cg_flags &= amdgpu_cg_mask;
if (adev->ip_blocks[i].status.hw)
continue;
if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
+ (amdgpu_sriov_vf(adev) && (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP)) ||
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) {
r = adev->ip_blocks[i].version->funcs->hw_init(adev);
if (r) {
adev->ip_blocks[i].status.hw = true;
/* right after GMC hw init, we create CSA */
- if (amdgpu_sriov_vf(adev)) {
+ if (amdgpu_mcbp || amdgpu_sriov_vf(adev)) {
r = amdgpu_allocate_static_csa(adev, &adev->virt.csa_obj,
AMDGPU_GEM_DOMAIN_VRAM,
AMDGPU_CSA_SIZE);
return 0;
}
+static int amdgpu_device_enable_mgpu_fan_boost(void)
+{
+ struct amdgpu_gpu_instance *gpu_ins;
+ struct amdgpu_device *adev;
+ int i, ret = 0;
+
+ mutex_lock(&mgpu_info.mutex);
+
+ /*
+ * MGPU fan boost feature should be enabled
+ * only when there are two or more dGPUs in
+ * the system
+ */
+ if (mgpu_info.num_dgpu < 2)
+ goto out;
+
+ for (i = 0; i < mgpu_info.num_dgpu; i++) {
+ gpu_ins = &(mgpu_info.gpu_ins[i]);
+ adev = gpu_ins->adev;
+ if (!(adev->flags & AMD_IS_APU) &&
+ !gpu_ins->mgpu_fan_enabled &&
+ adev->powerplay.pp_funcs &&
+ adev->powerplay.pp_funcs->enable_mgpu_fan_boost) {
+ ret = amdgpu_dpm_enable_mgpu_fan_boost(adev);
+ if (ret)
+ break;
+
+ gpu_ins->mgpu_fan_enabled = 1;
+ }
+ }
+
+out:
+ mutex_unlock(&mgpu_info.mutex);
+
+ return ret;
+}
+
/**
* amdgpu_device_ip_late_init - run late init for hardware IPs
*
amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
- queue_delayed_work(system_wq, &adev->late_init_work,
- msecs_to_jiffies(AMDGPU_RESUME_MS));
-
amdgpu_device_fill_reset_magic(adev);
+ r = amdgpu_device_enable_mgpu_fan_boost();
+ if (r)
+ DRM_ERROR("enable mgpu fan boost failed (%d).\n", r);
+
+ /* set to low pstate by default */
+ amdgpu_xgmi_set_pstate(adev, 0);
+
return 0;
}
return 0;
}
-static int amdgpu_device_enable_mgpu_fan_boost(void)
-{
- struct amdgpu_gpu_instance *gpu_ins;
- struct amdgpu_device *adev;
- int i, ret = 0;
-
- mutex_lock(&mgpu_info.mutex);
-
- /*
- * MGPU fan boost feature should be enabled
- * only when there are two or more dGPUs in
- * the system
- */
- if (mgpu_info.num_dgpu < 2)
- goto out;
-
- for (i = 0; i < mgpu_info.num_dgpu; i++) {
- gpu_ins = &(mgpu_info.gpu_ins[i]);
- adev = gpu_ins->adev;
- if (!(adev->flags & AMD_IS_APU) &&
- !gpu_ins->mgpu_fan_enabled &&
- adev->powerplay.pp_funcs &&
- adev->powerplay.pp_funcs->enable_mgpu_fan_boost) {
- ret = amdgpu_dpm_enable_mgpu_fan_boost(adev);
- if (ret)
- break;
-
- gpu_ins->mgpu_fan_enabled = 1;
- }
- }
-
-out:
- mutex_unlock(&mgpu_info.mutex);
-
- return ret;
-}
-
/**
- * amdgpu_device_ip_late_init_func_handler - work handler for ib test
+ * amdgpu_device_delayed_init_work_handler - work handler for IB tests
*
* @work: work_struct.
*/
-static void amdgpu_device_ip_late_init_func_handler(struct work_struct *work)
+static void amdgpu_device_delayed_init_work_handler(struct work_struct *work)
{
struct amdgpu_device *adev =
- container_of(work, struct amdgpu_device, late_init_work.work);
+ container_of(work, struct amdgpu_device, delayed_init_work.work);
int r;
r = amdgpu_ib_ring_tests(adev);
if (r)
DRM_ERROR("ib ring test failed (%d).\n", r);
-
- r = amdgpu_device_enable_mgpu_fan_boost();
- if (r)
- DRM_ERROR("enable mgpu fan boost failed (%d).\n", r);
-
- /*set to low pstate by default */
- amdgpu_xgmi_set_pstate(adev, 0);
-
}
static void amdgpu_device_delay_enable_gfx_off(struct work_struct *work)
case CHIP_VEGA20:
#if defined(CONFIG_DRM_AMD_DC_DCN1_0)
case CHIP_RAVEN:
+#endif
+#if defined(CONFIG_DRM_AMD_DC_DCN2_0)
+ case CHIP_NAVI10:
#endif
return amdgpu_dc != 0;
#endif
hash_init(adev->mn_hash);
mutex_init(&adev->lock_reset);
mutex_init(&adev->virt.dpm_mutex);
+ mutex_init(&adev->psp.mutex);
- amdgpu_device_check_arguments(adev);
+ r = amdgpu_device_check_arguments(adev);
+ if (r)
+ return r;
spin_lock_init(&adev->mmio_idx_lock);
spin_lock_init(&adev->smc_idx_lock);
INIT_LIST_HEAD(&adev->ring_lru_list);
spin_lock_init(&adev->ring_lru_list_lock);
- INIT_DELAYED_WORK(&adev->late_init_work,
- amdgpu_device_ip_late_init_func_handler);
+ INIT_DELAYED_WORK(&adev->delayed_init_work,
+ amdgpu_device_delayed_init_work_handler);
INIT_DELAYED_WORK(&adev->gfx.gfx_off_delay_work,
amdgpu_device_delay_enable_gfx_off);
if (adev->rio_mem == NULL)
DRM_INFO("PCI I/O BAR is not found.\n");
+ /* enable PCIE atomic ops */
+ r = pci_enable_atomic_ops_to_root(adev->pdev,
+ PCI_EXP_DEVCAP2_ATOMIC_COMP32 |
+ PCI_EXP_DEVCAP2_ATOMIC_COMP64);
+ if (r) {
+ adev->have_atomics_support = false;
+ DRM_INFO("PCIE atomic ops is not supported\n");
+ } else {
+ adev->have_atomics_support = true;
+ }
+
amdgpu_device_get_pcie_info(adev);
+ if (amdgpu_mcbp)
+ DRM_INFO("MCBP is enabled\n");
+
+ if (amdgpu_mes && adev->asic_type >= CHIP_NAVI10)
+ adev->enable_mes = true;
+
+ if (amdgpu_discovery && adev->asic_type >= CHIP_NAVI10) {
+ r = amdgpu_discovery_init(adev);
+ if (r) {
+ dev_err(adev->dev, "amdgpu_discovery_init failed\n");
+ return r;
+ }
+ }
+
/* early init functions */
r = amdgpu_device_ip_early_init(adev);
if (r)
goto fence_driver_init;
}
- /* Read BIOS */
- if (!amdgpu_get_bios(adev)) {
- r = -EINVAL;
- goto failed;
- }
-
- r = amdgpu_atombios_init(adev);
- if (r) {
- dev_err(adev->dev, "amdgpu_atombios_init failed\n");
- amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, 0, 0);
- goto failed;
- }
-
/* detect if we are with an SRIOV vbios */
amdgpu_device_detect_sriov_bios(adev);
amdgpu_fbdev_init(adev);
+ if (amdgpu_sriov_vf(adev) && amdgim_is_hwperf(adev))
+ amdgpu_pm_virt_sysfs_init(adev);
+
r = amdgpu_pm_sysfs_init(adev);
if (r)
DRM_ERROR("registering pm debugfs failed (%d).\n", r);
+ r = amdgpu_ucode_sysfs_init(adev);
+ if (r)
+ DRM_ERROR("Creating firmware sysfs failed (%d).\n", r);
+
r = amdgpu_debugfs_gem_init(adev);
if (r)
DRM_ERROR("registering gem debugfs failed (%d).\n", r);
}
/* must succeed. */
- amdgpu_ras_post_init(adev);
+ amdgpu_ras_resume(adev);
+
+ queue_delayed_work(system_wq, &adev->delayed_init_work,
+ msecs_to_jiffies(AMDGPU_RESUME_MS));
+
+ r = device_create_file(adev->dev, &dev_attr_pcie_replay_count);
+ if (r) {
+ dev_err(adev->dev, "Could not create pcie_replay_count");
+ return r;
+ }
+
+ if (IS_ENABLED(CONFIG_PERF_EVENTS))
+ r = amdgpu_pmu_init(adev);
+ if (r)
+ dev_err(adev->dev, "amdgpu_pmu_init failed\n");
return 0;
adev->firmware.gpu_info_fw = NULL;
}
adev->accel_working = false;
- cancel_delayed_work_sync(&adev->late_init_work);
+ cancel_delayed_work_sync(&adev->delayed_init_work);
/* free i2c buses */
if (!amdgpu_device_has_dc_support(adev))
amdgpu_i2c_fini(adev);
iounmap(adev->rmmio);
adev->rmmio = NULL;
amdgpu_device_doorbell_fini(adev);
+ if (amdgpu_sriov_vf(adev) && amdgim_is_hwperf(adev))
+ amdgpu_pm_virt_sysfs_fini(adev);
+
amdgpu_debugfs_regs_cleanup(adev);
+ device_remove_file(adev->dev, &dev_attr_pcie_replay_count);
+ amdgpu_ucode_sysfs_fini(adev);
+ if (IS_ENABLED(CONFIG_PERF_EVENTS))
+ amdgpu_pmu_fini(adev);
+ amdgpu_debugfs_preempt_cleanup(adev);
+ if (amdgpu_discovery && adev->asic_type >= CHIP_NAVI10)
+ amdgpu_discovery_fini(adev);
}
if (fbcon)
amdgpu_fbdev_set_suspend(adev, 1);
- cancel_delayed_work_sync(&adev->late_init_work);
+ cancel_delayed_work_sync(&adev->delayed_init_work);
if (!amdgpu_device_has_dc_support(adev)) {
/* turn off display hw */
amdgpu_amdkfd_suspend(adev);
+ amdgpu_ras_suspend(adev);
+
r = amdgpu_device_ip_suspend_phase1(adev);
/* evict vram memory */
if (r)
return r;
+ queue_delayed_work(system_wq, &adev->delayed_init_work,
+ msecs_to_jiffies(AMDGPU_RESUME_MS));
+
if (!amdgpu_device_has_dc_support(adev)) {
/* pin cursors */
list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
return r;
/* Make sure IB tests flushed */
- flush_delayed_work(&adev->late_init_work);
+ flush_delayed_work(&adev->delayed_init_work);
/* blat the mode back in */
if (fbcon) {
drm_kms_helper_poll_enable(dev);
+ amdgpu_ras_resume(adev);
+
/*
* Most of the connector probing functions try to acquire runtime pm
* refs to ensure that the GPU is powered on when connector polling is
if (!ring || !ring->sched.thread)
continue;
- drm_sched_stop(&ring->sched);
-
/* after all hw jobs are reset, hw fence is meaningless, so force_completion */
amdgpu_fence_driver_force_completion(ring);
}
if(job)
drm_sched_increase_karma(&job->base);
-
-
+ /* Don't suspend on bare metal if we are not going to HW reset the ASIC */
if (!amdgpu_sriov_vf(adev)) {
if (!need_full_reset)
if (vram_lost)
amdgpu_device_fill_reset_magic(tmp_adev);
+ /*
+ * Add this ASIC as tracked as reset was already
+ * complete successfully.
+ */
+ amdgpu_register_gpu_instance(tmp_adev);
+
+ r = amdgpu_device_ip_late_init(tmp_adev);
+ if (r)
+ goto out;
+
+ /* must succeed. */
+ amdgpu_ras_resume(tmp_adev);
+
/* Update PSP FW topology after reset */
if (hive && tmp_adev->gmc.xgmi.num_physical_nodes > 1)
r = amdgpu_xgmi_update_topology(hive, tmp_adev);
return r;
}
-static void amdgpu_device_post_asic_reset(struct amdgpu_device *adev,
- struct amdgpu_job *job)
+static bool amdgpu_device_lock_adev(struct amdgpu_device *adev, bool trylock)
{
- int i;
+ if (trylock) {
+ if (!mutex_trylock(&adev->lock_reset))
+ return false;
+ } else
+ mutex_lock(&adev->lock_reset);
- for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
- struct amdgpu_ring *ring = adev->rings[i];
-
- if (!ring || !ring->sched.thread)
- continue;
-
- if (!adev->asic_reset_res)
- drm_sched_resubmit_jobs(&ring->sched);
-
- drm_sched_start(&ring->sched, !adev->asic_reset_res);
- }
-
- if (!amdgpu_device_has_dc_support(adev)) {
- drm_helper_resume_force_mode(adev->ddev);
- }
-
- adev->asic_reset_res = 0;
-}
-
-static void amdgpu_device_lock_adev(struct amdgpu_device *adev)
-{
- mutex_lock(&adev->lock_reset);
atomic_inc(&adev->gpu_reset_counter);
adev->in_gpu_reset = 1;
/* Block kfd: SRIOV would do it separately */
if (!amdgpu_sriov_vf(adev))
amdgpu_amdkfd_pre_reset(adev);
+
+ return true;
}
static void amdgpu_device_unlock_adev(struct amdgpu_device *adev)
int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
struct amdgpu_job *job)
{
- int r;
+ struct list_head device_list, *device_list_handle = NULL;
+ bool need_full_reset, job_signaled;
struct amdgpu_hive_info *hive = NULL;
- bool need_full_reset = false;
struct amdgpu_device *tmp_adev = NULL;
- struct list_head device_list, *device_list_handle = NULL;
+ int i, r = 0;
+ need_full_reset = job_signaled = false;
INIT_LIST_HEAD(&device_list);
dev_info(adev->dev, "GPU reset begin!\n");
+ cancel_delayed_work_sync(&adev->delayed_init_work);
+
+ hive = amdgpu_get_xgmi_hive(adev, false);
+
/*
- * In case of XGMI hive disallow concurrent resets to be triggered
- * by different nodes. No point also since the one node already executing
- * reset will also reset all the other nodes in the hive.
+ * Here we trylock to avoid chain of resets executing from
+ * either trigger by jobs on different adevs in XGMI hive or jobs on
+ * different schedulers for same device while this TO handler is running.
+ * We always reset all schedulers for device and all devices for XGMI
+ * hive so that should take care of them too.
*/
- hive = amdgpu_get_xgmi_hive(adev, 0);
- if (hive && adev->gmc.xgmi.num_physical_nodes > 1 &&
- !mutex_trylock(&hive->reset_lock))
+
+ if (hive && !mutex_trylock(&hive->reset_lock)) {
+ DRM_INFO("Bailing on TDR for s_job:%llx, hive: %llx as another already in progress",
+ job->base.id, hive->hive_id);
return 0;
+ }
/* Start with adev pre asic reset first for soft reset check.*/
- amdgpu_device_lock_adev(adev);
- r = amdgpu_device_pre_asic_reset(adev,
- job,
- &need_full_reset);
- if (r) {
- /*TODO Should we stop ?*/
- DRM_ERROR("GPU pre asic reset failed with err, %d for drm dev, %s ",
- r, adev->ddev->unique);
- adev->asic_reset_res = r;
+ if (!amdgpu_device_lock_adev(adev, !hive)) {
+ DRM_INFO("Bailing on TDR for s_job:%llx, as another already in progress",
+ job->base.id);
+ return 0;
}
/* Build list of devices to reset */
- if (need_full_reset && adev->gmc.xgmi.num_physical_nodes > 1) {
+ if (adev->gmc.xgmi.num_physical_nodes > 1) {
if (!hive) {
amdgpu_device_unlock_adev(adev);
return -ENODEV;
device_list_handle = &device_list;
}
+ /*
+ * Mark these ASICs to be reseted as untracked first
+ * And add them back after reset completed
+ */
+ list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head)
+ amdgpu_unregister_gpu_instance(tmp_adev);
+
+ /* block all schedulers and reset given job's ring */
+ list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) {
+ /* disable ras on ALL IPs */
+ if (amdgpu_device_ip_need_full_reset(tmp_adev))
+ amdgpu_ras_suspend(tmp_adev);
+
+ for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
+ struct amdgpu_ring *ring = tmp_adev->rings[i];
+
+ if (!ring || !ring->sched.thread)
+ continue;
+
+ drm_sched_stop(&ring->sched, &job->base);
+ }
+ }
+
+
+ /*
+ * Must check guilty signal here since after this point all old
+ * HW fences are force signaled.
+ *
+ * job->base holds a reference to parent fence
+ */
+ if (job && job->base.s_fence->parent &&
+ dma_fence_is_signaled(job->base.s_fence->parent))
+ job_signaled = true;
+
+ if (!amdgpu_device_ip_need_full_reset(adev))
+ device_list_handle = &device_list;
+
+ if (job_signaled) {
+ dev_info(adev->dev, "Guilty job already signaled, skipping HW reset");
+ goto skip_hw_reset;
+ }
+
+
+ /* Guilty job will be freed after this*/
+ r = amdgpu_device_pre_asic_reset(adev,
+ job,
+ &need_full_reset);
+ if (r) {
+ /*TODO Should we stop ?*/
+ DRM_ERROR("GPU pre asic reset failed with err, %d for drm dev, %s ",
+ r, adev->ddev->unique);
+ adev->asic_reset_res = r;
+ }
+
retry: /* Rest of adevs pre asic reset from XGMI hive. */
list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) {
if (tmp_adev == adev)
continue;
- amdgpu_device_lock_adev(tmp_adev);
+ amdgpu_device_lock_adev(tmp_adev, false);
r = amdgpu_device_pre_asic_reset(tmp_adev,
NULL,
&need_full_reset);
goto retry;
}
+skip_hw_reset:
+
/* Post ASIC reset for all devs .*/
list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) {
- amdgpu_device_post_asic_reset(tmp_adev, tmp_adev == adev ? job : NULL);
+ for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
+ struct amdgpu_ring *ring = tmp_adev->rings[i];
+
+ if (!ring || !ring->sched.thread)
+ continue;
+
+ /* No point to resubmit jobs if we didn't HW reset*/
+ if (!tmp_adev->asic_reset_res && !job_signaled)
+ drm_sched_resubmit_jobs(&ring->sched);
+
+ drm_sched_start(&ring->sched, !tmp_adev->asic_reset_res);
+ }
+
+ if (!amdgpu_device_has_dc_support(tmp_adev) && !job_signaled) {
+ drm_helper_resume_force_mode(tmp_adev->ddev);
+ }
+
+ tmp_adev->asic_reset_res = 0;
if (r) {
/* bad news, how to tell it to userspace ? */
amdgpu_device_unlock_adev(tmp_adev);
}
- if (hive && adev->gmc.xgmi.num_physical_nodes > 1)
+ if (hive)
mutex_unlock(&hive->reset_lock);
if (r)
return r;
}
-static void amdgpu_device_get_min_pci_speed_width(struct amdgpu_device *adev,
- enum pci_bus_speed *speed,
- enum pcie_link_width *width)
-{
- struct pci_dev *pdev = adev->pdev;
- enum pci_bus_speed cur_speed;
- enum pcie_link_width cur_width;
- u32 ret = 1;
-
- *speed = PCI_SPEED_UNKNOWN;
- *width = PCIE_LNK_WIDTH_UNKNOWN;
-
- while (pdev) {
- cur_speed = pcie_get_speed_cap(pdev);
- cur_width = pcie_get_width_cap(pdev);
- ret = pcie_bandwidth_available(adev->pdev, NULL,
- NULL, &cur_width);
- if (!ret)
- cur_width = PCIE_LNK_WIDTH_RESRV;
-
- if (cur_speed != PCI_SPEED_UNKNOWN) {
- if (*speed == PCI_SPEED_UNKNOWN)
- *speed = cur_speed;
- else if (cur_speed < *speed)
- *speed = cur_speed;
- }
-
- if (cur_width != PCIE_LNK_WIDTH_UNKNOWN) {
- if (*width == PCIE_LNK_WIDTH_UNKNOWN)
- *width = cur_width;
- else if (cur_width < *width)
- *width = cur_width;
- }
- pdev = pci_upstream_bridge(pdev);
- }
-}
-
/**
* amdgpu_device_get_pcie_info - fence pcie info about the PCIE slot
*
if (adev->pm.pcie_gen_mask && adev->pm.pcie_mlw_mask)
return;
- amdgpu_device_get_min_pci_speed_width(adev, &platform_speed_cap,
- &platform_link_width);
+ pcie_bandwidth_available(adev->pdev, NULL,
+ &platform_speed_cap, &platform_link_width);
if (adev->pm.pcie_gen_mask == 0) {
/* asic caps */
git.samba.org / sfrench / cifs-2.6.git / blobdiff