CVE-2024-49876 (GCVE-0-2024-49876)
Vulnerability from cvelistv5
Published
2024-10-21 18:01
Modified
2026-08-05 11:40
Severity ?
7.8 (High) - CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
Summary
In the Linux kernel, the following vulnerability has been resolved: drm/xe: fix UAF around queue destruction We currently do stuff like queuing the final destruction step on a random system wq, which will outlive the driver instance. With bad timing we can teardown the driver with one or more work workqueue still being alive leading to various UAF splats. Add a fini step to ensure user queues are properly torn down. At this point GuC should already be nuked so queue itself should no longer be referenced from hw pov. v2 (Matt B) - Looks much safer to use a waitqueue and then just wait for the xa_array to become empty before triggering the drain. (cherry picked from commit 861108666cc0e999cffeab6aff17b662e68774e3)
References
Impacted products
Vendor Product Version
Linux Linux Version: dd08ebf6c3525a7ea2186e636df064ea47281987
Version: dd08ebf6c3525a7ea2186e636df064ea47281987
Version: dd08ebf6c3525a7ea2186e636df064ea47281987
 Create a notification for this product.
Show details on NVD website

To clipboard 

{
  "containers": {
    "adp": [
      {
        "metrics": [
          {
            "other": {
              "content": {
                "id": "CVE-2024-49876",
                "options": [
                  {
                    "Exploitation": "none"
                  },
                  {
                    "Automatable": "no"
                  },
                  {
                    "Technical Impact": "partial"
                  }
                ],
                "role": "CISA Coordinator",
                "timestamp": "2024-10-22T13:46:17.394123Z",
                "version": "2.0.3"
              },
              "type": "ssvc"
            }
          }
        ],
        "providerMetadata": {
          "dateUpdated": "2024-10-22T13:48:51.208Z",
          "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
          "shortName": "CISA-ADP"
        },
        "title": "CISA ADP Vulnrichment"
      }
    ],
    "cna": {
      "affected": [
        {
          "defaultStatus": "unaffected",
          "product": "Linux",
          "programFiles": [
            "drivers/gpu/drm/xe/xe_device.c",
            "drivers/gpu/drm/xe/xe_device_types.h",
            "drivers/gpu/drm/xe/xe_guc_submit.c",
            "drivers/gpu/drm/xe/xe_guc_types.h"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "lessThan": "272b0e78874586d6ccae04079d75b27b47705544",
              "status": "affected",
              "version": "dd08ebf6c3525a7ea2186e636df064ea47281987",
              "versionType": "git"
            },
            {
              "lessThan": "421c74670b0f9d5c007f1276d3647aa58f407fde",
              "status": "affected",
              "version": "dd08ebf6c3525a7ea2186e636df064ea47281987",
              "versionType": "git"
            },
            {
              "lessThan": "2d2be279f1ca9e7288282d4214f16eea8a727cdb",
              "status": "affected",
              "version": "dd08ebf6c3525a7ea2186e636df064ea47281987",
              "versionType": "git"
            }
          ]
        },
        {
          "defaultStatus": "affected",
          "product": "Linux",
          "programFiles": [
            "drivers/gpu/drm/xe/xe_device.c",
            "drivers/gpu/drm/xe/xe_device_types.h",
            "drivers/gpu/drm/xe/xe_guc_submit.c",
            "drivers/gpu/drm/xe/xe_guc_types.h"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "status": "affected",
              "version": "6.8"
            },
            {
              "lessThan": "6.8",
              "status": "unaffected",
              "version": "0",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "6.10.*",
              "status": "unaffected",
              "version": "6.10.14",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "6.11.*",
              "status": "unaffected",
              "version": "6.11.3",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "*",
              "status": "unaffected",
              "version": "6.12",
              "versionType": "original_commit_for_fix"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "6.10.14",
                  "versionStartIncluding": "6.8",
                  "vulnerable": true
                },
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "6.11.3",
                  "versionStartIncluding": "6.8",
                  "vulnerable": true
                },
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "6.12",
                  "versionStartIncluding": "6.8",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/xe: fix UAF around queue destruction\n\nWe currently do stuff like queuing the final destruction step on a\nrandom system wq, which will outlive the driver instance. With bad\ntiming we can teardown the driver with one or more work workqueue still\nbeing alive leading to various UAF splats. Add a fini step to ensure\nuser queues are properly torn down. At this point GuC should already be\nnuked so queue itself should no longer be referenced from hw pov.\n\nv2 (Matt B)\n - Looks much safer to use a waitqueue and then just wait for the\n   xa_array to become empty before triggering the drain.\n\n(cherry picked from commit 861108666cc0e999cffeab6aff17b662e68774e3)"
        }
      ],
      "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 - The vulnerable exec-queue create/destroy paths are reached only through DRM ioctls (`XE_EXEC_QUEUE_CREATE`/`XE_EXEC_QUEUE_DESTROY`) and fd close on the local device node `/dev/dri/renderD128`. No network or adjacent-network exposure exists.\nAC:L - The attacker fully controls the side of the race that creates the danger \u2014 a create/destroy loop keeps an unsynchronized `system_wq` fini work item outstanding essentially continuously, so no timing window must be won. Device unbind/remove (VF teardown on multi-tenant GPU hosts, eGPU hot-remove, driver reload) is a routine event after which the UAF is near-deterministic rather than a narrow race.\nPR:L - Both ioctls are marked `DRM_RENDER_ALLOW` with no capability check, so any local user with render-node access (default for desktop/logind sessions, and for any app on Android/embedded GPU stacks) can create and destroy exec queues. No root or CAP_SYS_ADMIN is needed.\nUI:N - The attacker\u0027s own process (queue destroy loop plus fd close) drives the vulnerable path, and the co-occurring device removal happens autonomously in realistic deployments \u2014 SR-IOV VF teardown by the hypervisor management stack, PCI hot-remove, or automated driver rebind \u2014 with no victim action required.\nS:U - The use-after-free corrupts kernel heap objects belonging to the same kernel security authority that the attacker is already interacting with. There is no VM, IOMMU, or sandbox boundary crossed.\nC:H - The stale worker reads a freed `xe_device`/`xe_guc` slab object that the attacker can reallocate and control, and follows pointers out of it (`guc_to_xe()`, `ge-\u003eq`, LRC/BO chains), enabling disclosure of arbitrary reallocated kernel memory.\nI:H - The worker performs numerous writes into the freed allocation \u2014 `mutex_lock()`, `xa_erase()`, GuC ID bitmap release, DRM scheduler entity/sched list teardown, `kfree()` of an attacker-influenced pointer \u2014 which is a classic heap-spray-to-arbitrary-write primitive; on module unload the work function pointer itself dangles into freed module text, allowing control-flow hijack.\nA:H - Even unweaponized, the bug produces the \"various UAF splats\" described by the author \u2014 KASAN reports, oopses, and panics from dereferencing the freed device, GuC, and scheduler state \u2014 resulting in a full kernel crash."
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-08-05T11:40:19.265Z",
        "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
        "shortName": "Linux"
      },
      "references": [
        {
          "url": "https://git.kernel.org/stable/c/272b0e78874586d6ccae04079d75b27b47705544"
        },
        {
          "url": "https://git.kernel.org/stable/c/421c74670b0f9d5c007f1276d3647aa58f407fde"
        },
        {
          "url": "https://git.kernel.org/stable/c/2d2be279f1ca9e7288282d4214f16eea8a727cdb"
        }
      ],
      "title": "drm/xe: fix UAF around queue destruction",
      "x_generator": {
        "engine": "bippy-1.2.0"
      }
    }
  },
  "cveMetadata": {
    "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
    "assignerShortName": "Linux",
    "cveId": "CVE-2024-49876",
    "datePublished": "2024-10-21T18:01:16.098Z",
    "dateReserved": "2024-10-21T12:17:06.020Z",
    "dateUpdated": "2026-08-05T11:40:19.265Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2",
  "vulnerability-lookup:meta": {
    "vulnrichment": {
      "containers": "{\"adp\": [{\"title\": \"CISA ADP Vulnrichment\", \"metrics\": [{\"other\": {\"type\": \"ssvc\", \"content\": {\"id\": \"CVE-2024-49876\", \"role\": \"CISA Coordinator\", \"options\": [{\"Exploitation\": \"none\"}, {\"Automatable\": \"no\"}, {\"Technical Impact\": \"partial\"}], \"version\": \"2.0.3\", \"timestamp\": \"2024-10-22T13:46:17.394123Z\"}}}], \"providerMetadata\": {\"orgId\": \"134c704f-9b21-4f2e-91b3-4a467353bcc0\", \"shortName\": \"CISA-ADP\", \"dateUpdated\": \"2024-10-22T13:46:20.517Z\"}}], \"cna\": {\"title\": \"drm/xe: fix UAF around queue destruction\", \"metrics\": [{\"cvssV3_1\": {\"version\": \"3.1\", \"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\"}, \"scenarios\": [{\"lang\": \"en\", \"value\": \"AV:L - The vulnerable exec-queue create/destroy paths are reached only through DRM ioctls (`XE_EXEC_QUEUE_CREATE`/`XE_EXEC_QUEUE_DESTROY`) and fd close on the local device node `/dev/dri/renderD128`. No network or adjacent-network exposure exists.\\nAC:L - The attacker fully controls the side of the race that creates the danger \\u2014 a create/destroy loop keeps an unsynchronized `system_wq` fini work item outstanding essentially continuously, so no timing window must be won. Device unbind/remove (VF teardown on multi-tenant GPU hosts, eGPU hot-remove, driver reload) is a routine event after which the UAF is near-deterministic rather than a narrow race.\\nPR:L - Both ioctls are marked `DRM_RENDER_ALLOW` with no capability check, so any local user with render-node access (default for desktop/logind sessions, and for any app on Android/embedded GPU stacks) can create and destroy exec queues. No root or CAP_SYS_ADMIN is needed.\\nUI:N - The attacker\u0027s own process (queue destroy loop plus fd close) drives the vulnerable path, and the co-occurring device removal happens autonomously in realistic deployments \\u2014 SR-IOV VF teardown by the hypervisor management stack, PCI hot-remove, or automated driver rebind \\u2014 with no victim action required.\\nS:U - The use-after-free corrupts kernel heap objects belonging to the same kernel security authority that the attacker is already interacting with. There is no VM, IOMMU, or sandbox boundary crossed.\\nC:H - The stale worker reads a freed `xe_device`/`xe_guc` slab object that the attacker can reallocate and control, and follows pointers out of it (`guc_to_xe()`, `ge-\u003eq`, LRC/BO chains), enabling disclosure of arbitrary reallocated kernel memory.\\nI:H - The worker performs numerous writes into the freed allocation \\u2014 `mutex_lock()`, `xa_erase()`, GuC ID bitmap release, DRM scheduler entity/sched list teardown, `kfree()` of an attacker-influenced pointer \\u2014 which is a classic heap-spray-to-arbitrary-write primitive; on module unload the work function pointer itself dangles into freed module text, allowing control-flow hijack.\\nA:H - Even unweaponized, the bug produces the \\\"various UAF splats\\\" described by the author \\u2014 KASAN reports, oopses, and panics from dereferencing the freed device, GuC, and scheduler state \\u2014 resulting in a full kernel crash.\"}]}], \"affected\": [{\"repo\": \"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git\", \"vendor\": \"Linux\", \"product\": \"Linux\", \"versions\": [{\"status\": \"affected\", \"version\": \"dd08ebf6c3525a7ea2186e636df064ea47281987\", \"lessThan\": \"272b0e78874586d6ccae04079d75b27b47705544\", \"versionType\": \"git\"}, {\"status\": \"affected\", \"version\": \"dd08ebf6c3525a7ea2186e636df064ea47281987\", \"lessThan\": \"421c74670b0f9d5c007f1276d3647aa58f407fde\", \"versionType\": \"git\"}, {\"status\": \"affected\", \"version\": \"dd08ebf6c3525a7ea2186e636df064ea47281987\", \"lessThan\": \"2d2be279f1ca9e7288282d4214f16eea8a727cdb\", \"versionType\": \"git\"}], \"programFiles\": [\"drivers/gpu/drm/xe/xe_device.c\", \"drivers/gpu/drm/xe/xe_device_types.h\", \"drivers/gpu/drm/xe/xe_guc_submit.c\", \"drivers/gpu/drm/xe/xe_guc_types.h\"], \"defaultStatus\": \"unaffected\"}, {\"repo\": \"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git\", \"vendor\": \"Linux\", \"product\": \"Linux\", \"versions\": [{\"status\": \"affected\", \"version\": \"6.8\"}, {\"status\": \"unaffected\", \"version\": \"0\", \"lessThan\": \"6.8\", \"versionType\": \"semver\"}, {\"status\": \"unaffected\", \"version\": \"6.10.14\", \"versionType\": \"semver\", \"lessThanOrEqual\": \"6.10.*\"}, {\"status\": \"unaffected\", \"version\": \"6.11.3\", \"versionType\": \"semver\", \"lessThanOrEqual\": \"6.11.*\"}, {\"status\": \"unaffected\", \"version\": \"6.12\", \"versionType\": \"original_commit_for_fix\", \"lessThanOrEqual\": \"*\"}], \"programFiles\": [\"drivers/gpu/drm/xe/xe_device.c\", \"drivers/gpu/drm/xe/xe_device_types.h\", \"drivers/gpu/drm/xe/xe_guc_submit.c\", \"drivers/gpu/drm/xe/xe_guc_types.h\"], \"defaultStatus\": \"affected\"}], \"references\": [{\"url\": \"https://git.kernel.org/stable/c/272b0e78874586d6ccae04079d75b27b47705544\"}, {\"url\": \"https://git.kernel.org/stable/c/421c74670b0f9d5c007f1276d3647aa58f407fde\"}, {\"url\": \"https://git.kernel.org/stable/c/2d2be279f1ca9e7288282d4214f16eea8a727cdb\"}], \"x_generator\": {\"engine\": \"bippy-1.2.0\"}, \"descriptions\": [{\"lang\": \"en\", \"value\": \"In the Linux kernel, the following vulnerability has been resolved:\\n\\ndrm/xe: fix UAF around queue destruction\\n\\nWe currently do stuff like queuing the final destruction step on a\\nrandom system wq, which will outlive the driver instance. With bad\\ntiming we can teardown the driver with one or more work workqueue still\\nbeing alive leading to various UAF splats. Add a fini step to ensure\\nuser queues are properly torn down. At this point GuC should already be\\nnuked so queue itself should no longer be referenced from hw pov.\\n\\nv2 (Matt B)\\n - Looks much safer to use a waitqueue and then just wait for the\\n   xa_array to become empty before triggering the drain.\\n\\n(cherry picked from commit 861108666cc0e999cffeab6aff17b662e68774e3)\"}], \"cpeApplicability\": [{\"nodes\": [{\"negate\": false, \"cpeMatch\": [{\"criteria\": \"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*\", \"vulnerable\": true, \"versionEndExcluding\": \"6.10.14\", \"versionStartIncluding\": \"6.8\"}, {\"criteria\": \"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*\", \"vulnerable\": true, \"versionEndExcluding\": \"6.11.3\", \"versionStartIncluding\": \"6.8\"}, {\"criteria\": \"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*\", \"vulnerable\": true, \"versionEndExcluding\": \"6.12\", \"versionStartIncluding\": \"6.8\"}], \"operator\": \"OR\"}]}], \"providerMetadata\": {\"orgId\": \"416baaa9-dc9f-4396-8d5f-8c081fb06d67\", \"shortName\": \"Linux\", \"dateUpdated\": \"2026-08-05T11:40:19.265Z\"}}}",
      "cveMetadata": "{\"cveId\": \"CVE-2024-49876\", \"state\": \"PUBLISHED\", \"dateUpdated\": \"2026-08-05T11:40:19.265Z\", \"dateReserved\": \"2024-10-21T12:17:06.020Z\", \"assignerOrgId\": \"416baaa9-dc9f-4396-8d5f-8c081fb06d67\", \"datePublished\": \"2024-10-21T18:01:16.098Z\", \"assignerShortName\": \"Linux\"}",
      "dataType": "CVE_RECORD",
      "dataVersion": "5.2"
    }
  }
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.




Sightings

Author Source Type Date

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or seen somewhere by the user.
  • Confirmed: The vulnerability is confirmed from an analyst perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: This vulnerability was exploited and seen by the user reporting the sighting.
  • Patched: This vulnerability was successfully patched by the user reporting the sighting.
  • Not exploited: This vulnerability was not exploited or seen by the user reporting the sighting.
  • Not confirmed: The user expresses doubt about the veracity of the vulnerability.
  • Not patched: This vulnerability was not successfully patched by the user reporting the sighting.