GCVE Workshop - 22 September 2026 (14:00-18:00), Luxembourg Before The Vulnopticon Conference - Registration

GSD-2024-26628

Vulnerability from gsd - Updated: 2024-02-20 06:02
Details
In the Linux kernel, the following vulnerability has been resolved: drm/amdkfd: Fix lock dependency warning ====================================================== WARNING: possible circular locking dependency detected 6.5.0-kfd-fkuehlin #276 Not tainted ------------------------------------------------------ kworker/8:2/2676 is trying to acquire lock: ffff9435aae95c88 ((work_completion)(&svm_bo->eviction_work)){+.+.}-{0:0}, at: __flush_work+0x52/0x550 but task is already holding lock: ffff9435cd8e1720 (&svms->lock){+.+.}-{3:3}, at: svm_range_deferred_list_work+0xe8/0x340 [amdgpu] which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #2 (&svms->lock){+.+.}-{3:3}: __mutex_lock+0x97/0xd30 kfd_ioctl_alloc_memory_of_gpu+0x6d/0x3c0 [amdgpu] kfd_ioctl+0x1b2/0x5d0 [amdgpu] __x64_sys_ioctl+0x86/0xc0 do_syscall_64+0x39/0x80 entry_SYSCALL_64_after_hwframe+0x63/0xcd -> #1 (&mm->mmap_lock){++++}-{3:3}: down_read+0x42/0x160 svm_range_evict_svm_bo_worker+0x8b/0x340 [amdgpu] process_one_work+0x27a/0x540 worker_thread+0x53/0x3e0 kthread+0xeb/0x120 ret_from_fork+0x31/0x50 ret_from_fork_asm+0x11/0x20 -> #0 ((work_completion)(&svm_bo->eviction_work)){+.+.}-{0:0}: __lock_acquire+0x1426/0x2200 lock_acquire+0xc1/0x2b0 __flush_work+0x80/0x550 __cancel_work_timer+0x109/0x190 svm_range_bo_release+0xdc/0x1c0 [amdgpu] svm_range_free+0x175/0x180 [amdgpu] svm_range_deferred_list_work+0x15d/0x340 [amdgpu] process_one_work+0x27a/0x540 worker_thread+0x53/0x3e0 kthread+0xeb/0x120 ret_from_fork+0x31/0x50 ret_from_fork_asm+0x11/0x20 other info that might help us debug this: Chain exists of: (work_completion)(&svm_bo->eviction_work) --> &mm->mmap_lock --> &svms->lock Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(&svms->lock); lock(&mm->mmap_lock); lock(&svms->lock); lock((work_completion)(&svm_bo->eviction_work)); I believe this cannot really lead to a deadlock in practice, because svm_range_evict_svm_bo_worker only takes the mmap_read_lock if the BO refcount is non-0. That means it's impossible that svm_range_bo_release is running concurrently. However, there is no good way to annotate this. To avoid the problem, take a BO reference in svm_range_schedule_evict_svm_bo instead of in the worker. That way it's impossible for a BO to get freed while eviction work is pending and the cancel_work_sync call in svm_range_bo_release can be eliminated. v2: Use svm_bo_ref_unless_zero and explained why that's safe. Also removed redundant checks that are already done in amdkfd_fence_enable_signaling.
Aliases
To clipboard 

{
  "gsd": {
    "metadata": {
      "exploitCode": "unknown",
      "remediation": "unknown",
      "reportConfidence": "confirmed",
      "type": "vulnerability"
    },
    "osvSchema": {
      "aliases": [
        "CVE-2024-26628"
      ],
      "details": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdkfd: Fix lock dependency warning\n\n======================================================\nWARNING: possible circular locking dependency detected\n6.5.0-kfd-fkuehlin #276 Not tainted\n------------------------------------------------------\nkworker/8:2/2676 is trying to acquire lock:\nffff9435aae95c88 ((work_completion)(\u0026svm_bo-\u003eeviction_work)){+.+.}-{0:0}, at: __flush_work+0x52/0x550\n\nbut task is already holding lock:\nffff9435cd8e1720 (\u0026svms-\u003elock){+.+.}-{3:3}, at: svm_range_deferred_list_work+0xe8/0x340 [amdgpu]\n\nwhich lock already depends on the new lock.\n\nthe existing dependency chain (in reverse order) is:\n\n-\u003e #2 (\u0026svms-\u003elock){+.+.}-{3:3}:\n       __mutex_lock+0x97/0xd30\n       kfd_ioctl_alloc_memory_of_gpu+0x6d/0x3c0 [amdgpu]\n       kfd_ioctl+0x1b2/0x5d0 [amdgpu]\n       __x64_sys_ioctl+0x86/0xc0\n       do_syscall_64+0x39/0x80\n       entry_SYSCALL_64_after_hwframe+0x63/0xcd\n\n-\u003e #1 (\u0026mm-\u003emmap_lock){++++}-{3:3}:\n       down_read+0x42/0x160\n       svm_range_evict_svm_bo_worker+0x8b/0x340 [amdgpu]\n       process_one_work+0x27a/0x540\n       worker_thread+0x53/0x3e0\n       kthread+0xeb/0x120\n       ret_from_fork+0x31/0x50\n       ret_from_fork_asm+0x11/0x20\n\n-\u003e #0 ((work_completion)(\u0026svm_bo-\u003eeviction_work)){+.+.}-{0:0}:\n       __lock_acquire+0x1426/0x2200\n       lock_acquire+0xc1/0x2b0\n       __flush_work+0x80/0x550\n       __cancel_work_timer+0x109/0x190\n       svm_range_bo_release+0xdc/0x1c0 [amdgpu]\n       svm_range_free+0x175/0x180 [amdgpu]\n       svm_range_deferred_list_work+0x15d/0x340 [amdgpu]\n       process_one_work+0x27a/0x540\n       worker_thread+0x53/0x3e0\n       kthread+0xeb/0x120\n       ret_from_fork+0x31/0x50\n       ret_from_fork_asm+0x11/0x20\n\nother info that might help us debug this:\n\nChain exists of:\n  (work_completion)(\u0026svm_bo-\u003eeviction_work) --\u003e \u0026mm-\u003emmap_lock --\u003e \u0026svms-\u003elock\n\n Possible unsafe locking scenario:\n\n       CPU0                    CPU1\n       ----                    ----\n  lock(\u0026svms-\u003elock);\n                               lock(\u0026mm-\u003emmap_lock);\n                               lock(\u0026svms-\u003elock);\n  lock((work_completion)(\u0026svm_bo-\u003eeviction_work));\n\nI believe this cannot really lead to a deadlock in practice, because\nsvm_range_evict_svm_bo_worker only takes the mmap_read_lock if the BO\nrefcount is non-0. That means it\u0027s impossible that svm_range_bo_release\nis running concurrently. However, there is no good way to annotate this.\n\nTo avoid the problem, take a BO reference in\nsvm_range_schedule_evict_svm_bo instead of in the worker. That way it\u0027s\nimpossible for a BO to get freed while eviction work is pending and the\ncancel_work_sync call in svm_range_bo_release can be eliminated.\n\nv2: Use svm_bo_ref_unless_zero and explained why that\u0027s safe. Also\nremoved redundant checks that are already done in\namdkfd_fence_enable_signaling.",
      "id": "GSD-2024-26628",
      "modified": "2024-02-20T06:02:29.177873Z",
      "schema_version": "1.4.0"
    }
  },
  "namespaces": {
    "cve.org": {
      "CVE_data_meta": {
        "ASSIGNER": "cve@kernel.org",
        "ID": "CVE-2024-26628",
        "STATE": "REJECT"
      },
      "data_format": "MITRE",
      "data_type": "CVE",
      "data_version": "4.0",
      "description": {
        "description_data": [
          {
            "lang": "eng",
            "value": "** REJECT ** This CVE ID has been rejected or withdrawn by its CVE Numbering Authority."
          }
        ]
      }
    },
    "nvd.nist.gov": {
      "cve": {
        "descriptions": [
          {
            "lang": "en",
            "value": "Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority."
          }
        ],
        "id": "CVE-2024-26628",
        "lastModified": "2024-03-20T17:15:07.367",
        "metrics": {},
        "published": "2024-03-06T07:15:13.173",
        "references": [],
        "sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
        "vulnStatus": "Rejected"
      }
    }
  }
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.





Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.


Detection rules are retrieved from Rulezet.


Related by attack behaviour

Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.