From: Shashank Sharma shashank.sharma@amd.com

mainline inclusion from mainline-v6.9-rc1 commit b8f67b9ddf4f8fe6dd536590712b5912ad78f99c category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/IADDFV CVE: CVE-2024-41008

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?i...

-----------------------------------------------------------

This patch changes the handling and lifecycle of vm->task_info object. The major changes are: - vm->task_info is a dynamically allocated ptr now, and its uasge is reference counted. - introducing two new helper funcs for task_info lifecycle management - amdgpu_vm_get_task_info: reference counts up task_info before returning this info - amdgpu_vm_put_task_info: reference counts down task_info - last put to task_info() frees task_info from the vm.

This patch also does logistical changes required for existing usage of vm->task_info.

V2: Do not block all the prints when task_info not found (Felix)

V3: Fixed review comments from Felix - Fix wrong indentation - No debug message for -ENOMEM - Add NULL check for task_info - Do not duplicate the debug messages (ti vs no ti) - Get first reference of task_info in vm_init(), put last in vm_fini()

V4: Fixed review comments from Felix - fix double reference increment in create_task_info - change amdgpu_vm_get_task_info_pasid - additional changes in amdgpu_gem.c while porting

Cc: Christian Koenig christian.koenig@amd.com Cc: Alex Deucher alexander.deucher@amd.com Cc: Felix Kuehling Felix.Kuehling@amd.com Reviewed-by: Felix Kuehling Felix.Kuehling@amd.com Signed-off-by: Shashank Sharma shashank.sharma@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com Conflicts: drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c drivers/gpu/drm/amd/amdgpu/amdgpu_job.c [1. Some drivers does not use struct task_info, so they are no need to modify 2. use idr instead of xarray in amdgpu_vm_get_vm_from_pasid(), because there are no functional changes] Signed-off-by: Zhang Zekun zhangzekun11@huawei.com --- drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 18 ++- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 161 ++++++++++++++------ drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 21 ++- drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 22 +-- drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c | 20 ++- drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 24 +-- drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c | 18 ++- 7 files changed, 191 insertions(+), 93 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index dcfe8a3b03ff..d45e84faed00 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -32,11 +32,9 @@ static void amdgpu_job_timedout(struct drm_sched_job *s_job) { struct amdgpu_ring *ring = to_amdgpu_ring(s_job->sched); struct amdgpu_job *job = to_amdgpu_job(s_job); - struct amdgpu_task_info ti; + struct amdgpu_task_info *ti; struct amdgpu_device *adev = ring->adev;

- memset(&ti, 0, sizeof(struct amdgpu_task_info)); - if (amdgpu_gpu_recovery && amdgpu_ring_soft_recovery(ring, job->vmid, s_job->s_fence->parent)) { DRM_ERROR("ring %s timeout, but soft recovered\n", @@ -44,12 +42,16 @@ static void amdgpu_job_timedout(struct drm_sched_job *s_job) return; }

- amdgpu_vm_get_task_info(ring->adev, job->pasid, &ti); DRM_ERROR("ring %s timeout, signaled seq=%u, emitted seq=%u\n", - job->base.sched->name, atomic_read(&ring->fence_drv.last_seq), - ring->fence_drv.sync_seq); - DRM_ERROR("Process information: process %s pid %d thread %s pid %d\n", - ti.process_name, ti.tgid, ti.task_name, ti.pid); + job->base.sched->name, atomic_read(&ring->fence_drv.last_seq), + ring->fence_drv.sync_seq); + + ti = amdgpu_vm_get_task_info_pasid(ring->adev, job->pasid); + if (ti) { + DRM_ERROR("Process information: process %s pid %d thread %s pid %d\n", + ti->process_name, ti->tgid, ti->task_name, ti->pid); + amdgpu_vm_put_task_info(ti); + }

if (amdgpu_device_should_recover_gpu(ring->adev)) { amdgpu_device_gpu_recover(ring->adev, job); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index ad3de971a087..a05eb8f393dc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -1525,16 +1525,20 @@ static int amdgpu_vm_update_ptes(struct amdgpu_vm_update_params *params, uint64_t upd_end = min(entry_end, frag_end); unsigned nptes = (upd_end - frag_start) >> shift; uint64_t upd_flags = flags | AMDGPU_PTE_FRAG(frag); + struct amdgpu_task_info *ti = amdgpu_vm_get_task_info_vm(vm);

/* This can happen when we set higher level PDs to * silent to stop fault floods. */ nptes = max(nptes, 1u);

- trace_amdgpu_vm_update_ptes(params, frag_start, upd_end, - nptes, dst, incr, upd_flags, - vm->task_info.pid, - vm->immediate.fence_context); + if (ti) { + trace_amdgpu_vm_update_ptes(params, frag_start, upd_end, + nptes, dst, incr, upd_flags, + ti->pid, + vm->immediate.fence_context); + amdgpu_vm_put_task_info(ti); + } amdgpu_vm_update_flags(params, pt, cursor.level, pe_start, dst, nptes, incr, upd_flags); @@ -2819,6 +2823,108 @@ long amdgpu_vm_wait_idle(struct amdgpu_vm *vm, long timeout) return dma_fence_wait_timeout(vm->last_unlocked, true, timeout); }

+static void amdgpu_vm_destroy_task_info(struct kref *kref) +{ + struct amdgpu_task_info *ti = container_of(kref, struct amdgpu_task_info, refcount); + + kfree(ti); +} + +static inline struct amdgpu_vm * +amdgpu_vm_get_vm_from_pasid(struct amdgpu_device *adev, u32 pasid) +{ + struct amdgpu_vm *vm; + unsigned long flags; + + spin_lock_irqsave(&adev->vm_manager.pasid_lock, flags); + vm = idr_find(&adev->vm_manager.pasid_idr, pasid); + spin_unlock_irqrestore(&adev->vm_manager.pasid_lock, flags); + + return vm; +} + +/** + * amdgpu_vm_put_task_info - reference down the vm task_info ptr + * + * @task_info: task_info struct under discussion. + * + * frees the vm task_info ptr at the last put + */ +void amdgpu_vm_put_task_info(struct amdgpu_task_info *task_info) +{ + kref_put(&task_info->refcount, amdgpu_vm_destroy_task_info); +} + +/** + * amdgpu_vm_get_task_info_vm - Extracts task info for a vm. + * + * @vm: VM to get info from + * + * Returns the reference counted task_info structure, which must be + * referenced down with amdgpu_vm_put_task_info. + */ +struct amdgpu_task_info * +amdgpu_vm_get_task_info_vm(struct amdgpu_vm *vm) +{ + struct amdgpu_task_info *ti = NULL; + + if (vm) { + ti = vm->task_info; + kref_get(&vm->task_info->refcount); + } + + return ti; +} + +/** + * amdgpu_vm_get_task_info_pasid - Extracts task info for a PASID. + * + * @adev: drm device pointer + * @pasid: PASID identifier for VM + * + * Returns the reference counted task_info structure, which must be + * referenced down with amdgpu_vm_put_task_info. + */ +struct amdgpu_task_info * +amdgpu_vm_get_task_info_pasid(struct amdgpu_device *adev, u32 pasid) +{ + return amdgpu_vm_get_task_info_vm( + amdgpu_vm_get_vm_from_pasid(adev, pasid)); +} + +static int amdgpu_vm_create_task_info(struct amdgpu_vm *vm) +{ + vm->task_info = kzalloc(sizeof(struct amdgpu_task_info), GFP_KERNEL); + if (!vm->task_info) + return -ENOMEM; + + kref_init(&vm->task_info->refcount); + return 0; +} + +/** + * amdgpu_vm_set_task_info - Sets VMs task info. + * + * @vm: vm for which to set the info + */ +void amdgpu_vm_set_task_info(struct amdgpu_vm *vm) +{ + if (!vm->task_info) + return; + + if (vm->task_info->pid == current->pid) + return; + + vm->task_info->pid = current->pid; + get_task_comm(vm->task_info->task_name, current); + + if (current->group_leader->mm != current->mm) + return; + + vm->task_info->tgid = current->group_leader->pid; + get_task_comm(vm->task_info->process_name, current->group_leader); +} + /** * amdgpu_vm_init - initialize a vm instance * @@ -2914,6 +3020,10 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, if (r) goto error_unreserve;

+ r = amdgpu_vm_create_task_info(vm); + if (r) + DRM_DEBUG("Failed to create task info for VM\n"); + amdgpu_bo_unreserve(vm->root.base.bo);

if (pasid) { @@ -3141,6 +3251,7 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)

root = amdgpu_bo_ref(vm->root.base.bo); amdgpu_bo_reserve(root, true); + amdgpu_vm_put_task_info(vm->task_info); if (vm->pasid) { unsigned long flags;

@@ -3309,48 +3420,6 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) return 0; }

-/** - * amdgpu_vm_get_task_info - Extracts task info for a PASID. - * - * @adev: drm device pointer - * @pasid: PASID identifier for VM - * @task_info: task_info to fill. - */ -void amdgpu_vm_get_task_info(struct amdgpu_device *adev, u32 pasid, - struct amdgpu_task_info *task_info) -{ - struct amdgpu_vm *vm; - unsigned long flags; - - spin_lock_irqsave(&adev->vm_manager.pasid_lock, flags); - - vm = idr_find(&adev->vm_manager.pasid_idr, pasid); - if (vm) - *task_info = vm->task_info; - - spin_unlock_irqrestore(&adev->vm_manager.pasid_lock, flags); -} - -/** - * amdgpu_vm_set_task_info - Sets VMs task info. - * - * @vm: vm for which to set the info - */ -void amdgpu_vm_set_task_info(struct amdgpu_vm *vm) -{ - if (vm->task_info.pid) - return; - - vm->task_info.pid = current->pid; - get_task_comm(vm->task_info.task_name, current); - - if (current->group_leader->mm != current->mm) - return; - - vm->task_info.tgid = current->group_leader->pid; - get_task_comm(vm->task_info.process_name, current->group_leader); -} - /** * amdgpu_vm_handle_fault - graceful handling of VM faults. * @adev: amdgpu device pointer diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index c4218800e043..0e663464f1d6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -180,10 +180,11 @@ struct amdgpu_vm_pte_funcs { };

struct amdgpu_task_info { - char process_name[TASK_COMM_LEN]; - char task_name[TASK_COMM_LEN]; - pid_t pid; - pid_t tgid; + char process_name[TASK_COMM_LEN]; + char task_name[TASK_COMM_LEN]; + pid_t pid; + pid_t tgid; + struct kref refcount; };

/** @@ -311,7 +312,7 @@ struct amdgpu_vm { uint64_t pd_phys_addr;

/* Some basic info about the task */ - struct amdgpu_task_info task_info; + struct amdgpu_task_info *task_info;

/* Store positions of group of BOs */ struct ttm_lru_bulk_move lru_bulk_move; @@ -431,8 +432,14 @@ bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring, struct amdgpu_job *job); void amdgpu_vm_check_compute_bug(struct amdgpu_device *adev);

-void amdgpu_vm_get_task_info(struct amdgpu_device *adev, u32 pasid, - struct amdgpu_task_info *task_info); +struct amdgpu_task_info * +amdgpu_vm_get_task_info_pasid(struct amdgpu_device *adev, u32 pasid); + +struct amdgpu_task_info * +amdgpu_vm_get_task_info_vm(struct amdgpu_vm *vm); + +void amdgpu_vm_put_task_info(struct amdgpu_task_info *task_info); + bool amdgpu_vm_handle_fault(struct amdgpu_device *adev, u32 pasid, uint64_t addr);

diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c index 2aa9242c58ab..67d7ca92d95a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c @@ -114,18 +114,22 @@ static int gmc_v10_0_process_interrupt(struct amdgpu_device *adev, }

if (printk_ratelimit()) { - struct amdgpu_task_info task_info; - - memset(&task_info, 0, sizeof(struct amdgpu_task_info)); - amdgpu_vm_get_task_info(adev, entry->pasid, &task_info); + struct amdgpu_task_info *task_info;

dev_err(adev->dev, - "[%s] page fault (src_id:%u ring:%u vmid:%u pasid:%u, " - "for process %s pid %d thread %s pid %d)\n", + "[%s] page fault (src_id:%u ring:%u vmid:%u pasid:%u)\n", entry->vmid_src ? "mmhub" : "gfxhub", - entry->src_id, entry->ring_id, entry->vmid, - entry->pasid, task_info.process_name, task_info.tgid, - task_info.task_name, task_info.pid); + 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_err(adev->dev, + " in 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); + } + dev_err(adev->dev, " in page starting at address 0x%016llx from client %d\n", addr, entry->client_id); if (!amdgpu_sriov_vf(adev)) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 873bc33912e2..a7d80bdec5c0 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -1469,18 +1469,24 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev, gmc_v8_0_set_fault_enable_default(adev, false);

if (printk_ratelimit()) { - struct amdgpu_task_info task_info; + struct amdgpu_task_info *task_info;

- memset(&task_info, 0, sizeof(struct amdgpu_task_info)); - amdgpu_vm_get_task_info(adev, entry->pasid, &task_info); + dev_err(adev->dev, "GPU fault detected: %d 0x%08x\n", + entry->src_id, entry->src_data[0]); + + task_info = amdgpu_vm_get_task_info_pasid(adev, entry->pasid); + if (task_info) { + dev_err(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); + }

- dev_err(adev->dev, "GPU fault detected: %d 0x%08x for process %s pid %d thread %s pid %d\n", - entry->src_id, entry->src_data[0], task_info.process_name, - task_info.tgid, task_info.task_name, task_info.pid); dev_err(adev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n", - addr); + addr); dev_err(adev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", status); + gmc_v8_0_vm_decode_fault(adev, status, addr, mc_client, entry->pasid); } diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index 945cbdbc2f99..41f0075c70a9 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -561,18 +561,22 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, }

if (printk_ratelimit()) { - struct amdgpu_task_info task_info; - - memset(&task_info, 0, sizeof(struct amdgpu_task_info)); - amdgpu_vm_get_task_info(adev, entry->pasid, &task_info); + struct amdgpu_task_info *task_info;

dev_err(adev->dev, - "[%s] %s page fault (src_id:%u ring:%u vmid:%u " - "pasid:%u, for process %s pid %d thread %s pid %d)\n", - hub_name, retry_fault ? "retry" : "no-retry", - entry->src_id, entry->ring_id, entry->vmid, - entry->pasid, task_info.process_name, task_info.tgid, - task_info.task_name, task_info.pid); + "[%s] %s page fault (src_id:%u ring:%u vmid:%u pasid:%u)\n", hub_name, + retry_fault ? "retry" : "no-retry", + 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_err(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); + } + dev_err(adev->dev, " in page starting at address 0x%016llx from client %d\n", addr, entry->client_id); if (!amdgpu_sriov_vf(adev)) { diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c index bd4caa36ab2e..a79a5da24392 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c @@ -231,7 +231,7 @@ void kfd_smi_event_update_thermal_throttling(struct kfd_dev *dev, void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid) { struct amdgpu_device *adev = (struct amdgpu_device *)dev->kgd; - struct amdgpu_task_info task_info; + struct amdgpu_task_info *task_info; /* VmFault msg = (hex)uint32_pid(8) + :(1) + task name(16) = 25 */ /* 1 byte event + 1 byte space + 25 bytes msg + 1 byte \n + * 1 byte \0 = 29 @@ -242,14 +242,20 @@ void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid) if (list_empty(&dev->smi_clients)) return;

- memset(&task_info, 0, sizeof(struct amdgpu_task_info)); - amdgpu_vm_get_task_info(adev, pasid, &task_info); + task_info = amdgpu_vm_get_task_info_pasid(adev, pasid); + if (!task_info) + return; + /* Report VM faults from user applications, not retry from kernel */ - if (!task_info.pid) + if (!task_info->pid) { + amdgpu_vm_put_task_info(task_info); return; + }

- len = snprintf(fifo_in, sizeof(fifo_in), "%x %x:%s\n", KFD_SMI_EVENT_VMFAULT, - task_info.pid, task_info.task_name); + len = snprintf(fifo_in, sizeof(fifo_in), "%x %x:%s\n", + KFD_SMI_EVENT_VMFAULT, task_info->pid, + task_info->task_name); + amdgpu_vm_put_task_info(task_info);

add_event_to_kfifo(dev, KFD_SMI_EVENT_VMFAULT, fifo_in, len); }