Vulnerability-Lookup

GHSA-3995-93H6-G37X

Vulnerability from github – Published: 2026-08-10 15:33 – Updated: 2026-08-14 00:31

Details

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

mm/page_vma_mapped: fix device-private PMD handling

Commit 65edfda6f3f2 ("mm/rmap: extend rmap and migration support device-private entries") introduced the concept of device-private PMD entries, but did not correctly update the rmap walk code to account for them.

As a result, when page_vma_mapped_walk() encounters device-private PMD entries, it takes no action other than to acquire the PMD lock and exit.

However this is highly problematic for two reasons - firstly, device private entries possess a PFN so check_pmd() needs to be called to ensure an overlapping PFN range.

Secondly, and more importantly, if PVMW_MIGRATION is set the caller assumes the returned entry is a migration entry, resulting in memory corruption when the caller tries to interpret the device private entry as such.

In addition, commit 146287290023 ("mm/huge_memory: implement device-private THP splitting") allowed device private PMDs to be split like THP mappings, but again did not update this code path.

As a result, we might race a PMD split prior to acquiring the PMD lock.

This patch addresses all of these issues by invoking check_pmd(), ensuring PMVW_MIGRATION is not set and checks whether a split raced us we do for PMD THP and migration entries.

Instead of checking for a subset of the cases after taking the pmd_lock(), put device-private along with pmd_trans_huge() and pmd_is_migration_entry(). Also remove thp_migration_supported() as it is already guarded by pmd_is_migration_entry().

[akpm@linux-foundation.org: fix Raspberry Pi 1 build, per David]

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-2026-68163"
  ],
  "database_specific": {
    "cwe_ids": [],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2026-08-10T13:20:02Z",
    "severity": "HIGH"
  },
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nmm/page_vma_mapped: fix device-private PMD handling\n\nCommit 65edfda6f3f2 (\"mm/rmap: extend rmap and migration support\ndevice-private entries\") introduced the concept of device-private PMD\nentries, but did not correctly update the rmap walk code to account for\nthem.\n\nAs a result, when page_vma_mapped_walk() encounters device-private PMD\nentries, it takes no action other than to acquire the PMD lock and exit.\n\nHowever this is highly problematic for two reasons - firstly, device\nprivate entries possess a PFN so check_pmd() needs to be called to ensure\nan overlapping PFN range.\n\nSecondly, and more importantly, if PVMW_MIGRATION is set the caller\nassumes the returned entry is a migration entry, resulting in memory\ncorruption when the caller tries to interpret the device private entry as\nsuch.\n\nIn addition, commit 146287290023 (\"mm/huge_memory: implement\ndevice-private THP splitting\") allowed device private PMDs to be split\nlike THP mappings, but again did not update this code path.\n\nAs a result, we might race a PMD split prior to acquiring the PMD lock.\n\nThis patch addresses all of these issues by invoking check_pmd(), ensuring\nPMVW_MIGRATION is not set and checks whether a split raced us we do for\nPMD THP and migration entries.\n\nInstead of checking for a subset of the cases after taking the pmd_lock(),\nput device-private along with pmd_trans_huge() and\npmd_is_migration_entry().  Also remove thp_migration_supported() as it is\nalready guarded by pmd_is_migration_entry().\n\n[akpm@linux-foundation.org: fix Raspberry Pi 1 build, per David]",
  "id": "GHSA-3995-93h6-g37x",
  "modified": "2026-08-14T00:31:52Z",
  "published": "2026-08-10T15:33:38Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-68163"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/ab6209f4b48a98ef14d6766acdb62aa9bb32e670"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/f84ca9b1888d8fce7dfefe0e750fa971f8797486"
    }
  ],
  "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-2026-68163 (GCVE-0-2026-68163)

Vulnerability from cvelistv5 – Published: 2026-08-10 11:59 – Updated: 2026-08-17 05:00

Title

mm/page_vma_mapped: fix device-private PMD handling

Summary

In the Linux kernel, the following vulnerability has been resolved: mm/page_vma_mapped: fix device-private PMD handling Commit 65edfda6f3f2 ("mm/rmap: extend rmap and migration support device-private entries") introduced the concept of device-private PMD entries, but did not correctly update the rmap walk code to account for them. As a result, when page_vma_mapped_walk() encounters device-private PMD entries, it takes no action other than to acquire the PMD lock and exit. However this is highly problematic for two reasons - firstly, device private entries possess a PFN so check_pmd() needs to be called to ensure an overlapping PFN range. Secondly, and more importantly, if PVMW_MIGRATION is set the caller assumes the returned entry is a migration entry, resulting in memory corruption when the caller tries to interpret the device private entry as such. In addition, commit 146287290023 ("mm/huge_memory: implement device-private THP splitting") allowed device private PMDs to be split like THP mappings, but again did not update this code path. As a result, we might race a PMD split prior to acquiring the PMD lock. This patch addresses all of these issues by invoking check_pmd(), ensuring PMVW_MIGRATION is not set and checks whether a split raced us we do for PMD THP and migration entries. Instead of checking for a subset of the cases after taking the pmd_lock(), put device-private along with pmd_trans_huge() and pmd_is_migration_entry(). Also remove thp_migration_supported() as it is already guarded by pmd_is_migration_entry(). [akpm@linux-foundation.org: fix Raspberry Pi 1 build, per David]

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/ab6209f4b48a98ef1…
https://git.kernel.org/stable/c/f84ca9b1888d8fce7…

Impacted products

2 products

Vendor	Product	Version
Linux	Linux	Affected: 65edfda6f3f2e58f757485a056e4f1775a1404a8 , < ab6209f4b48a98ef14d6766acdb62aa9bb32e670 (git) Affected: 65edfda6f3f2e58f757485a056e4f1775a1404a8 , < f84ca9b1888d8fce7dfefe0e750fa971f8797486 (git)
Linux	Linux	Affected: 6.19 Unaffected: 0 , < 6.19 (semver) Unaffected: 7.1.6 , ≤ 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": [
            "mm/page_vma_mapped.c"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "lessThan": "ab6209f4b48a98ef14d6766acdb62aa9bb32e670",
              "status": "affected",
              "version": "65edfda6f3f2e58f757485a056e4f1775a1404a8",
              "versionType": "git"
            },
            {
              "lessThan": "f84ca9b1888d8fce7dfefe0e750fa971f8797486",
              "status": "affected",
              "version": "65edfda6f3f2e58f757485a056e4f1775a1404a8",
              "versionType": "git"
            }
          ]
        },
        {
          "defaultStatus": "affected",
          "product": "Linux",
          "programFiles": [
            "mm/page_vma_mapped.c"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "status": "affected",
              "version": "6.19"
            },
            {
              "lessThan": "6.19",
              "status": "unaffected",
              "version": "0",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "7.1.*",
              "status": "unaffected",
              "version": "7.1.6",
              "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": "7.1.6",
                  "versionStartIncluding": "6.19",
                  "vulnerable": true
                },
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "7.2",
                  "versionStartIncluding": "6.19",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nmm/page_vma_mapped: fix device-private PMD handling\n\nCommit 65edfda6f3f2 (\"mm/rmap: extend rmap and migration support\ndevice-private entries\") introduced the concept of device-private PMD\nentries, but did not correctly update the rmap walk code to account for\nthem.\n\nAs a result, when page_vma_mapped_walk() encounters device-private PMD\nentries, it takes no action other than to acquire the PMD lock and exit.\n\nHowever this is highly problematic for two reasons - firstly, device\nprivate entries possess a PFN so check_pmd() needs to be called to ensure\nan overlapping PFN range.\n\nSecondly, and more importantly, if PVMW_MIGRATION is set the caller\nassumes the returned entry is a migration entry, resulting in memory\ncorruption when the caller tries to interpret the device private entry as\nsuch.\n\nIn addition, commit 146287290023 (\"mm/huge_memory: implement\ndevice-private THP splitting\") allowed device private PMDs to be split\nlike THP mappings, but again did not update this code path.\n\nAs a result, we might race a PMD split prior to acquiring the PMD lock.\n\nThis patch addresses all of these issues by invoking check_pmd(), ensuring\nPMVW_MIGRATION is not set and checks whether a split raced us we do for\nPMD THP and migration entries.\n\nInstead of checking for a subset of the cases after taking the pmd_lock(),\nput device-private along with pmd_trans_huge() and\npmd_is_migration_entry().  Also remove thp_migration_supported() as it is\nalready guarded by pmd_is_migration_entry().\n\n[akpm@linux-foundation.org: fix Raspberry Pi 1 build, per David]"
        }
      ],
      "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 bug is in the MM rmap/page-table walk reached only from in-kernel memory-management paths (migrate_vma finalize, rmap unmap/migration) triggered by local syscalls and GPU driver ioctls (AMD KFD, nouveau, drm_pagemap), not from network packet handling.\nAC:L - An attacker with GPU/HMM access can reliably create device-private THP PMDs and drive migrate_vma/setup/finalize while racing PMD splits via concurrent threads; both sides of the race are attacker-controlled, not dependent on admin timing or uncontrollable layout.\nPR:L - Exploitation needs only an unprivileged local user who can use the GPU/HMM stack (typical render/video group access on ML/cloud GPU nodes) to allocate device-private THP and invoke migration; it does not require real root in the init namespace.\nUI:N - No victim interaction is required; memory migration and PMD splitting are driven entirely by the attacker\u0027s own process through normal GPU unified-memory/migration APIs and concurrent threads.\nS:U - Impact is kernel page-table corruption and local privilege escalation within the host kernel security domain; it is not a VM guest-to-host escape, IOMMU bypass, or cross-authority sandbox breakout per kernel CVSS scope guidance.\nC:H - Misinterpreting a device-private PMD as a migration entry corrupts page tables and rmap state; such MM corruption is readily weaponizable for arbitrary kernel memory reads, not merely a bounded leak.\nI:H - The faulty walk causes remove_migration_pmd() and related callers to install incorrect PMD/PTE contents from mis-typed entries, enabling arbitrary kernel memory writes and control-flow hijack via page-table manipulation.\nA:H - Incorrect PMD handling during migration or concurrent THP splitting can cause kernel oops/panic, use-after-free conditions, or hung migration paths, producing full system denial of service even before successful exploitation."
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-08-17T05:00:02.619Z",
        "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
        "shortName": "Linux"
      },
      "references": [
        {
          "url": "https://git.kernel.org/stable/c/ab6209f4b48a98ef14d6766acdb62aa9bb32e670"
        },
        {
          "url": "https://git.kernel.org/stable/c/f84ca9b1888d8fce7dfefe0e750fa971f8797486"
        }
      ],
      "title": "mm/page_vma_mapped: fix device-private PMD handling",
      "x_generator": {
        "engine": "bippy-1.2.0"
      }
    }
  },
  "cveMetadata": {
    "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
    "assignerShortName": "Linux",
    "cveId": "CVE-2026-68163",
    "datePublished": "2026-08-10T11:59:30.633Z",
    "dateReserved": "2026-07-30T09:28:09.372Z",
    "dateUpdated": "2026-08-17T05:00:02.619Z",
    "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.
Browse all ATT&CK techniques and the vulnerabilities related to each.

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.