Vulnerability-Lookup

GHSA-RF2R-54VM-QV57

Vulnerability from github – Published: 2026-08-15 06:32 – Updated: 2026-08-17 06:33

Details

In the Linux kernel, the following vulnerability has been resolved:

drm/amdgpu: fix KASAN slab-out-of-bounds in amdgpu_coredump ring dump

The ring content dump in amdgpu_coredump() uses two separate loops over adev->rings[]: the first counts rings with unsignalled fences to size the allocation, and the second copies ring data into the allocated buffers.

Both loops use the same condition to skip rings:

atomic_read(&ring->fence_drv.last_seq) == ring->fence_drv.sync_seq

Because last_seq is an atomic that is updated concurrently by the fence signalling path, additional rings may appear unsignalled in the second loop that were signalled during the first. When this happens, idx exceeds the allocated ring_count and the store to coredump->rings[idx] writes past the end of the kcalloc-ed buffer.

This was found during IGT stressful test amd_queue_reset which triggers random GPU resets. The OVERSIZE subtest (CMD_STREAM_EXEC_INVALID_PACKET_LENGTH_OVERSIZE on GFX ring) provokes a ring timeout and subsequent coredump, which hits the race between the counting and copying loops. The failure is non-deterministic and depends on fence signalling timing during the reset.

KASAN log:

BUG: KASAN: slab-out-of-bounds in amdgpu_coredump+0x1274/0x12f0 [amdgpu] Write of size 4 at addr ffff888106154258 by task kworker/u128:5/23625 CPU: 16 UID: 0 PID: 23625 Comm: kworker/u128:5 Not tainted 6.19.0+ #35 Workqueue: amdgpu-reset-dev drm_sched_job_timedout [gpu_sched] Call Trace: dump_stack_lvl+0xa5/0x110 print_report+0xd1/0x660 kasan_report+0xf3/0x130 __asan_report_store4_noabort+0x17/0x30 amdgpu_coredump+0x1274/0x12f0 [amdgpu] amdgpu_job_timedout+0xef0/0x16c0 [amdgpu] drm_sched_job_timedout+0x194/0x5c0 [gpu_sched] process_one_work+0x84b/0x1990 worker_thread+0x6b8/0x11b0

Allocated by task 23625: kasan_save_stack+0x39/0x70 __kasan_kmalloc+0xc3/0xd0 __kmalloc_noprof+0x2ec/0x910 amdgpu_coredump+0x5c5/0x12f0 [amdgpu] amdgpu_job_timedout+0xef0/0x16c0 [amdgpu]

The buggy address belongs to the object at ffff888106154200 which belongs to the cache kmalloc-rnd-09-96 of size 96 The buggy address is located 16 bytes to the right of allocated 72-byte region [ffff888106154200, ffff888106154248)

72 bytes = 3 * sizeof(struct amdgpu_coredump_ring), so ring_count was 3 but idx reached 3+, writing ring_index (at struct offset 16) 16 bytes past the allocation.

Fix by adding an idx < ring_count guard to the copy loop so it cannot exceed the allocated count even when the fence state changes between the two passes.

Severity

7.8 (High)


                  
                    CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

Show details on source website

JSON

To clipboard

{
  "affected": [],
  "aliases": [
    "CVE-2026-74357"
  ],
  "database_specific": {
    "cwe_ids": [],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2026-08-15T06:22:37Z",
    "severity": "HIGH"
  },
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu: fix KASAN slab-out-of-bounds in amdgpu_coredump ring dump\n\nThe ring content dump in amdgpu_coredump() uses two separate loops over\nadev-\u003erings[]: the first counts rings with unsignalled fences to size\nthe allocation, and the second copies ring data into the allocated\nbuffers.\n\nBoth loops use the same condition to skip rings:\n\n    atomic_read(\u0026ring-\u003efence_drv.last_seq) == ring-\u003efence_drv.sync_seq\n\nBecause last_seq is an atomic that is updated concurrently by the fence\nsignalling path, additional rings may appear unsignalled in the second\nloop that were signalled during the first. When this happens, idx\nexceeds the allocated ring_count and the store to coredump-\u003erings[idx]\nwrites past the end of the kcalloc-ed buffer.\n\nThis was found during IGT stressful test amd_queue_reset which\ntriggers random GPU resets. The OVERSIZE subtest\n(CMD_STREAM_EXEC_INVALID_PACKET_LENGTH_OVERSIZE on GFX ring) provokes\na ring timeout and subsequent coredump, which hits the race between\nthe counting and copying loops. The failure is non-deterministic and\ndepends on fence signalling timing during the reset.\n\nKASAN log:\n\n  BUG: KASAN: slab-out-of-bounds in amdgpu_coredump+0x1274/0x12f0 [amdgpu]\n  Write of size 4 at addr ffff888106154258 by task kworker/u128:5/23625\n  CPU: 16 UID: 0 PID: 23625 Comm: kworker/u128:5 Not tainted 6.19.0+ #35\n  Workqueue: amdgpu-reset-dev drm_sched_job_timedout [gpu_sched]\n  Call Trace:\n   \u003cTASK\u003e\n   dump_stack_lvl+0xa5/0x110\n   print_report+0xd1/0x660\n   kasan_report+0xf3/0x130\n   __asan_report_store4_noabort+0x17/0x30\n   amdgpu_coredump+0x1274/0x12f0 [amdgpu]\n   amdgpu_job_timedout+0xef0/0x16c0 [amdgpu]\n   drm_sched_job_timedout+0x194/0x5c0 [gpu_sched]\n   process_one_work+0x84b/0x1990\n   worker_thread+0x6b8/0x11b0\n   \u003c/TASK\u003e\n\n  Allocated by task 23625:\n   kasan_save_stack+0x39/0x70\n   __kasan_kmalloc+0xc3/0xd0\n   __kmalloc_noprof+0x2ec/0x910\n   amdgpu_coredump+0x5c5/0x12f0 [amdgpu]\n   amdgpu_job_timedout+0xef0/0x16c0 [amdgpu]\n\n  The buggy address belongs to the object at ffff888106154200\n   which belongs to the cache kmalloc-rnd-09-96 of size 96\n  The buggy address is located 16 bytes to the right of\n   allocated 72-byte region [ffff888106154200, ffff888106154248)\n\n72 bytes = 3 * sizeof(struct amdgpu_coredump_ring), so ring_count was 3\nbut idx reached 3+, writing ring_index (at struct offset 16) 16 bytes\npast the allocation.\n\nFix by adding an idx \u003c ring_count guard to the copy loop so it cannot\nexceed the allocated count even when the fence state changes between\nthe two passes.",
  "id": "GHSA-rf2r-54vm-qv57",
  "modified": "2026-08-17T06:33:40Z",
  "published": "2026-08-15T06:32:31Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-74357"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/08ac3a7879d300302a1927ce2038629539a37f8b"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/efb1dadaeb897b22b9e118d398d0f41e70e9ccca"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    }
  ]
}

CVE-2026-74357 (GCVE-0-2026-74357)

Vulnerability from cvelistv5 – Published: 2026-08-15 05:58 – Updated: 2026-08-17 05:46

Title

drm/amdgpu: fix KASAN slab-out-of-bounds in amdgpu_coredump ring dump

Summary

In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: fix KASAN slab-out-of-bounds in amdgpu_coredump ring dump The ring content dump in amdgpu_coredump() uses two separate loops over adev->rings[]: the first counts rings with unsignalled fences to size the allocation, and the second copies ring data into the allocated buffers. Both loops use the same condition to skip rings: atomic_read(&ring->fence_drv.last_seq) == ring->fence_drv.sync_seq Because last_seq is an atomic that is updated concurrently by the fence signalling path, additional rings may appear unsignalled in the second loop that were signalled during the first. When this happens, idx exceeds the allocated ring_count and the store to coredump->rings[idx] writes past the end of the kcalloc-ed buffer. This was found during IGT stressful test amd_queue_reset which triggers random GPU resets. The OVERSIZE subtest (CMD_STREAM_EXEC_INVALID_PACKET_LENGTH_OVERSIZE on GFX ring) provokes a ring timeout and subsequent coredump, which hits the race between the counting and copying loops. The failure is non-deterministic and depends on fence signalling timing during the reset. KASAN log: BUG: KASAN: slab-out-of-bounds in amdgpu_coredump+0x1274/0x12f0 [amdgpu] Write of size 4 at addr ffff888106154258 by task kworker/u128:5/23625 CPU: 16 UID: 0 PID: 23625 Comm: kworker/u128:5 Not tainted 6.19.0+ #35 Workqueue: amdgpu-reset-dev drm_sched_job_timedout [gpu_sched] Call Trace: <TASK> dump_stack_lvl+0xa5/0x110 print_report+0xd1/0x660 kasan_report+0xf3/0x130 __asan_report_store4_noabort+0x17/0x30 amdgpu_coredump+0x1274/0x12f0 [amdgpu] amdgpu_job_timedout+0xef0/0x16c0 [amdgpu] drm_sched_job_timedout+0x194/0x5c0 [gpu_sched] process_one_work+0x84b/0x1990 worker_thread+0x6b8/0x11b0 </TASK> Allocated by task 23625: kasan_save_stack+0x39/0x70 __kasan_kmalloc+0xc3/0xd0 __kmalloc_noprof+0x2ec/0x910 amdgpu_coredump+0x5c5/0x12f0 [amdgpu] amdgpu_job_timedout+0xef0/0x16c0 [amdgpu] The buggy address belongs to the object at ffff888106154200 which belongs to the cache kmalloc-rnd-09-96 of size 96 The buggy address is located 16 bytes to the right of allocated 72-byte region [ffff888106154200, ffff888106154248) 72 bytes = 3 * sizeof(struct amdgpu_coredump_ring), so ring_count was 3 but idx reached 3+, writing ring_index (at struct offset 16) 16 bytes past the allocation. Fix by adding an idx < ring_count guard to the copy loop so it cannot exceed the allocated count even when the fence state changes between the two passes.

Severity

7.8 (High)


                        
                          CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

Assigner

Linux mitre root

References

2 references

URL	Tags
https://git.kernel.org/stable/c/efb1dadaeb897b22b…
https://git.kernel.org/stable/c/08ac3a7879d300302…

Impacted products

2 products

Vendor	Product	Version
Linux	Linux	Affected: eea85914d15bfe3bdf9f8f80a479f0dee0aa7d73 , < efb1dadaeb897b22b9e118d398d0f41e70e9ccca (git) Affected: eea85914d15bfe3bdf9f8f80a479f0dee0aa7d73 , < 08ac3a7879d300302a1927ce2038629539a37f8b (git)
Linux	Linux	Affected: 7.1 Unaffected: 0 , < 7.1 (semver) Unaffected: 7.1.5 , ≤ 7.1.* (semver) Unaffected: 7.2 , ≤ * (original_commit_for_fix)

Show details on NVD website

JSON

To clipboard

{
  "containers": {
    "cna": {
      "affected": [
        {
          "defaultStatus": "unaffected",
          "product": "Linux",
          "programFiles": [
            "drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "lessThan": "efb1dadaeb897b22b9e118d398d0f41e70e9ccca",
              "status": "affected",
              "version": "eea85914d15bfe3bdf9f8f80a479f0dee0aa7d73",
              "versionType": "git"
            },
            {
              "lessThan": "08ac3a7879d300302a1927ce2038629539a37f8b",
              "status": "affected",
              "version": "eea85914d15bfe3bdf9f8f80a479f0dee0aa7d73",
              "versionType": "git"
            }
          ]
        },
        {
          "defaultStatus": "affected",
          "product": "Linux",
          "programFiles": [
            "drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "status": "affected",
              "version": "7.1"
            },
            {
              "lessThan": "7.1",
              "status": "unaffected",
              "version": "0",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "7.1.*",
              "status": "unaffected",
              "version": "7.1.5",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "*",
              "status": "unaffected",
              "version": "7.2",
              "versionType": "original_commit_for_fix"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "7.1.5",
                  "versionStartIncluding": "7.1",
                  "vulnerable": true
                },
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "7.2",
                  "versionStartIncluding": "7.1",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/amdgpu: fix KASAN slab-out-of-bounds in amdgpu_coredump ring dump\n\nThe ring content dump in amdgpu_coredump() uses two separate loops over\nadev-\u003erings[]: the first counts rings with unsignalled fences to size\nthe allocation, and the second copies ring data into the allocated\nbuffers.\n\nBoth loops use the same condition to skip rings:\n\n    atomic_read(\u0026ring-\u003efence_drv.last_seq) == ring-\u003efence_drv.sync_seq\n\nBecause last_seq is an atomic that is updated concurrently by the fence\nsignalling path, additional rings may appear unsignalled in the second\nloop that were signalled during the first. When this happens, idx\nexceeds the allocated ring_count and the store to coredump-\u003erings[idx]\nwrites past the end of the kcalloc-ed buffer.\n\nThis was found during IGT stressful test amd_queue_reset which\ntriggers random GPU resets. The OVERSIZE subtest\n(CMD_STREAM_EXEC_INVALID_PACKET_LENGTH_OVERSIZE on GFX ring) provokes\na ring timeout and subsequent coredump, which hits the race between\nthe counting and copying loops. The failure is non-deterministic and\ndepends on fence signalling timing during the reset.\n\nKASAN log:\n\n  BUG: KASAN: slab-out-of-bounds in amdgpu_coredump+0x1274/0x12f0 [amdgpu]\n  Write of size 4 at addr ffff888106154258 by task kworker/u128:5/23625\n  CPU: 16 UID: 0 PID: 23625 Comm: kworker/u128:5 Not tainted 6.19.0+ #35\n  Workqueue: amdgpu-reset-dev drm_sched_job_timedout [gpu_sched]\n  Call Trace:\n   \u003cTASK\u003e\n   dump_stack_lvl+0xa5/0x110\n   print_report+0xd1/0x660\n   kasan_report+0xf3/0x130\n   __asan_report_store4_noabort+0x17/0x30\n   amdgpu_coredump+0x1274/0x12f0 [amdgpu]\n   amdgpu_job_timedout+0xef0/0x16c0 [amdgpu]\n   drm_sched_job_timedout+0x194/0x5c0 [gpu_sched]\n   process_one_work+0x84b/0x1990\n   worker_thread+0x6b8/0x11b0\n   \u003c/TASK\u003e\n\n  Allocated by task 23625:\n   kasan_save_stack+0x39/0x70\n   __kasan_kmalloc+0xc3/0xd0\n   __kmalloc_noprof+0x2ec/0x910\n   amdgpu_coredump+0x5c5/0x12f0 [amdgpu]\n   amdgpu_job_timedout+0xef0/0x16c0 [amdgpu]\n\n  The buggy address belongs to the object at ffff888106154200\n   which belongs to the cache kmalloc-rnd-09-96 of size 96\n  The buggy address is located 16 bytes to the right of\n   allocated 72-byte region [ffff888106154200, ffff888106154248)\n\n72 bytes = 3 * sizeof(struct amdgpu_coredump_ring), so ring_count was 3\nbut idx reached 3+, writing ring_index (at struct offset 16) 16 bytes\npast the allocation.\n\nFix by adding an idx \u003c ring_count guard to the copy loop so it cannot\nexceed the allocated count even when the fence state changes between\nthe two passes."
        }
      ],
      "metrics": [
        {
          "cvssV3_1": {
            "baseScore": 7.8,
            "baseSeverity": "HIGH",
            "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
            "version": "3.1"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "AV:L - Exploitation requires submitting AMDGPU_CS ioctls on /dev/dri/renderD* to hang a GPU ring, which triggers amdgpu_coredump() from the local drm_sched_job_timedout workqueue during reset recovery.\nAC:L - An attacker can reliably provoke ring timeouts with crafted GFX packets (e.g., invalid packet length oversize per IGT) and repeatedly retry to win the fence-signalling race between the counting and copy loops.\nPR:L - AMDGPU_CS and related ioctls are allowed for DRM render clients on /dev/dri/renderD*, reachable by unprivileged local users in the render/video group on workstations, gaming PCs, and cloud GPU tenants.\nUI:N - No victim interaction is required; the attacker alone submits malicious GPU command streams that cause the timeout and subsequent vulnerable coredump path.\nS:U - The slab-out-of-bounds write corrupts kernel heap memory within the same kernel security authority and does not cross a VM, container, or IOMMU boundary by itself.\nC:H - The bug performs out-of-bounds writes past the kcalloc\u0027d rings[] array and can also memcpy past the undersized rings_dw buffer, corrupting adjacent kmalloc objects and enabling kernel memory disclosure.\nI:H - Out-of-bounds heap writes of ring metadata and full ring buffer contents provide memory corruption primitives that can be leveraged for arbitrary kernel data modification and code execution.\nA:H - Slab-out-of-bounds corruption during GPU reset coredump in a kernel worker can cause kernel Oops/panic or failed recovery, denying availability on AMD GPU systems."
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-08-17T05:46:18.298Z",
        "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
        "shortName": "Linux"
      },
      "references": [
        {
          "url": "https://git.kernel.org/stable/c/efb1dadaeb897b22b9e118d398d0f41e70e9ccca"
        },
        {
          "url": "https://git.kernel.org/stable/c/08ac3a7879d300302a1927ce2038629539a37f8b"
        }
      ],
      "title": "drm/amdgpu: fix KASAN slab-out-of-bounds in amdgpu_coredump ring dump",
      "x_generator": {
        "engine": "bippy-1.2.0"
      }
    }
  },
  "cveMetadata": {
    "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
    "assignerShortName": "Linux",
    "cveId": "CVE-2026-74357",
    "datePublished": "2026-08-15T05:58:42.116Z",
    "dateReserved": "2026-08-15T05:44:03.887Z",
    "dateUpdated": "2026-08-17T05:46:18.298Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}

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.

The MITRE ATT&CK techniques below are AI-generated suggestions, inferred from the description of the vulnerability by the CIRCL/vulnerability-attack-technique-classification-roberta-base model, served locally by ML-Gateway. They have not been verified by an analyst and are provided for guidance only.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.

Action not permitted

GHSA-RF2R-54VM-QV57

CVE-2026-74357 (GCVE-0-2026-74357)

Tags

Sightings

Nomenclature