CVE-2026-74674 (GCVE-0-2026-74674)

Vulnerability from cvelistv5 – Published: 2026-08-22 15:32 – Updated: 2026-08-25 05:41
VLAI
Title
mm: fix incorrect flush address in direct page table reclaim
Summary
In the Linux kernel, the following vulnerability has been resolved: mm: fix incorrect flush address in direct page table reclaim When zap_pte_range reclaims a page table, it does: pte_free_tlb(tlb, pmd_pgtable(pmdval), addr); and this is unconditionally wrong: if this code executes, addr *always* points one past the end of the range covered by the table. The addr parameter is used to flush the TLB (really the paging-structure-cache) to drop references to the to-be-freed table, and any architecture that cares about the parameter will flush the wrong address. (But they'll still free the correct page). I think it's worth contemplating why the kernel works at all. If we hit the offending line of code, we will first clear the PMD entry (line 1954, zap_empty_pte_table), then we will issue pending flushes if force_flush is set (tlb_flush_mmu_tlbonly(tlb)), then we will skip the retry on line 1979 (phew!), and then we will do the offending pte_free_tlb call. *Or* we will clear the PMD entry immediately before pte_free_tlb (line 1983, zap_pte_table_if_empty). If we have any pending flushes (i.e. we actually zapped any last-level entries) at the time we clear the PMD entry, then the flush really ought to flush all references to the table (Linus certainly seems to think it will on all architectures [0]). The condition under which we have no accumulated flushes at the time of the clear is very complex (the whole zap_pte_range function has absurdly complex control flow). If we do hit the bad case, then we will end up clearing the PMD entry after the last time the range is flushed, and any CPU is free to cache a reference to the (empty) page table. If this happens due to an ordinary read or write, it would segfault, so it would be rare. But the cache could be speculatively filled as well. Then we'll flush the wrong address and then free and possibly reuse the table. On x86, even flushing the wrong address works on non-KPTI Intel systems because INVLPG flushes *all* paging-structure-caches, not just the ones for the target address. But INVPCID does not, and flush_tlb_one_user will use INVPCID if it's available. And then we're toast. AMD systems are more susceptible: we set the EFER.TCE bit, which makes even INVLPG only flush the target address. I think this might fix an issue in ripgrep reported here: https://github.com/BurntSushi/ripgrep/issues/3494 [0] https://lore.kernel.org/all/CA+55aFzBggoXtNXQeng5d_mRoDnaMBE5Y+URs+PHR67nUpMtaw@mail.gmail.com/T/#u
Severity
7.8 (High)
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
Assigner
References
Impacted products
Vendor Product Version CPE status
Linux Linux Affected: 4c640eb4181cf8de74c8b9e7c9cf16bf8d26b73e , < 0b8ff21cbda8808c86b18f1b0ca2d0025af9a80a (git)
Affected: 4c640eb4181cf8de74c8b9e7c9cf16bf8d26b73e , < 478a1c3abebfc717db0d1281a9cdd7befafee542 (git)
guessed Create a notification for this product.
Linux Linux Affected: 7.0
Unaffected: 0 , < 7.0 (semver)
Unaffected: 7.1.9 , ≤ 7.1.* (semver)
Unaffected: 7.2 , ≤ * (original_commit_for_fix)
guessed Create a notification for this product.
Show details on NVD website
To clipboard 

{
  "containers": {
    "cna": {
      "affected": [
        {
          "defaultStatus": "unaffected",
          "product": "Linux",
          "programFiles": [
            "mm/memory.c"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "lessThan": "0b8ff21cbda8808c86b18f1b0ca2d0025af9a80a",
              "status": "affected",
              "version": "4c640eb4181cf8de74c8b9e7c9cf16bf8d26b73e",
              "versionType": "git"
            },
            {
              "lessThan": "478a1c3abebfc717db0d1281a9cdd7befafee542",
              "status": "affected",
              "version": "4c640eb4181cf8de74c8b9e7c9cf16bf8d26b73e",
              "versionType": "git"
            }
          ]
        },
        {
          "defaultStatus": "affected",
          "product": "Linux",
          "programFiles": [
            "mm/memory.c"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "status": "affected",
              "version": "7.0"
            },
            {
              "lessThan": "7.0",
              "status": "unaffected",
              "version": "0",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "7.1.*",
              "status": "unaffected",
              "version": "7.1.9",
              "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": "7.1.9",
                  "versionStartIncluding": "7.0",
                  "vulnerable": true
                },
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "7.2",
                  "versionStartIncluding": "7.0",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nmm: fix incorrect flush address in direct page table reclaim\n\nWhen zap_pte_range reclaims a page table, it does:\n\n    pte_free_tlb(tlb, pmd_pgtable(pmdval), addr);\n\nand this is unconditionally wrong: if this code executes, addr *always*\npoints one past the end of the range covered by the table.  The addr\nparameter is used to flush the TLB (really the paging-structure-cache)\nto drop references to the to-be-freed table, and any architecture that\ncares about the parameter will flush the wrong address.  (But they\u0027ll\nstill free the correct page).\n\nI think it\u0027s worth contemplating why the kernel works at all.\n\nIf we hit the offending line of code, we will first clear the PMD entry\n(line 1954, zap_empty_pte_table), then we will issue pending flushes if\nforce_flush is set (tlb_flush_mmu_tlbonly(tlb)), then we will skip the\nretry on line 1979 (phew!), and then we will do the offending\npte_free_tlb call.  *Or* we will clear the PMD entry immediately before\npte_free_tlb (line 1983, zap_pte_table_if_empty).\n\nIf we have any pending flushes (i.e. we actually zapped any last-level\nentries) at the time we clear the PMD entry, then the flush really ought\nto flush all references to the table (Linus certainly seems to think it\nwill on all architectures [0]).\n\nThe condition under which we have no accumulated flushes at the time of\nthe clear is very complex (the whole zap_pte_range function has absurdly\ncomplex control flow).  If we do hit the bad case, then we will end up\nclearing the PMD entry after the last time the range is flushed, and any\nCPU is free to cache a reference to the (empty) page table.  If this\nhappens due to an ordinary read or write, it would segfault, so it would\nbe rare.  But the cache could be speculatively filled as well.  Then\nwe\u0027ll flush the wrong address and then free and possibly reuse the\ntable.\n\nOn x86, even flushing the wrong address works on non-KPTI Intel systems\nbecause INVLPG flushes *all* paging-structure-caches, not just the ones\nfor the target address.  But INVPCID does not, and flush_tlb_one_user\nwill use INVPCID if it\u0027s available.  And then we\u0027re toast.  AMD systems\nare more susceptible: we set the EFER.TCE bit, which makes even INVLPG\nonly flush the target address.\n\nI think this might fix an issue in ripgrep reported here:\nhttps://github.com/BurntSushi/ripgrep/issues/3494\n\n[0] https://lore.kernel.org/all/CA+55aFzBggoXtNXQeng5d_mRoDnaMBE5Y+URs+PHR67nUpMtaw@mail.gmail.com/T/#u"
        }
      ],
      "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 only reachable via local memory-management syscalls\u2014primarily madvise(MADV_DONTNEED) on anonymous mappings\u2014that drive zap_pte_range() page-table reclaim; there is no network, adjacent-radio, or physical-device path to this code.\nAC:L - An unprivileged attacker can repeatedly mmap and madvise(DONTNEED) 2MB+ anonymous regions and run concurrent memory pressure (as in the ripgrep repro) to hit the faulty pte_free_tlb() flush; attacker-controlled syscalls and threading make triggering the bad case feasible rather than dependent on uncontrollable system state.\nPR:L - Exploitation requires only a local unprivileged user manipulating their own address space through standard madvise/mmap syscalls; no root, capabilities beyond normal UID, or authentication is needed, and this works within user namespaces on the caller\u0027s mm.\nUI:N - No victim interaction is required; the attacker directly issues the syscalls and memory-access patterns that cause incorrect paging-structure-cache flushes during PTE page reclaim.\nS:U - Impact is stale paging-structure-cache entries and corrupted virtual-to-physical translations within the attacker\u0027s process and kernel MM subsystem\u2014standard local memory corruption, not a VM escape, IOMMU bypass, or cross-authority sandbox breakout.\nC:H - Flushing the wrong address leaves CPUs with cached references to freed PTE pages; once reused, stale translations can map unrelated physical pages into the attacker\u0027s address space, giving arbitrary read primitives and information disclosure consistent with use-after-free severity.\nI:H - Freed page-table pages can be reallocated while stale paging-structure caches still reference them, enabling wrong virtual-to-physical mappings and demonstrated user-space memory corruption (ripgrep/musl heap corruption), exploitable for arbitrary writes and control-flow hijacking.\nA:H - Stale TLB/paging-structure-cache entries after incorrect flushes cause user-space SIGSEGV (demonstrated in ripgrep on Linux 7.0) and can destabilize or crash affected workloads; memory-corruption class bugs in the MM path carry high availability impact even when not fully weaponized."
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-08-25T05:41:26.569Z",
        "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
        "shortName": "Linux"
      },
      "references": [
        {
          "url": "https://git.kernel.org/stable/c/0b8ff21cbda8808c86b18f1b0ca2d0025af9a80a"
        },
        {
          "url": "https://git.kernel.org/stable/c/478a1c3abebfc717db0d1281a9cdd7befafee542"
        }
      ],
      "title": "mm: fix incorrect flush address in direct page table reclaim",
      "x_generator": {
        "engine": "bippy-1.2.0"
      }
    }
  },
  "cveMetadata": {
    "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
    "assignerShortName": "Linux",
    "cveId": "CVE-2026-74674",
    "datePublished": "2026-08-22T15:32:43.670Z",
    "dateReserved": "2026-08-15T05:44:03.925Z",
    "dateUpdated": "2026-08-25T05:41:26.569Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2",
  "vulnerability-lookup:meta": {
    "epss": {
      "cve": "CVE-2026-74674",
      "date": "2026-09-01",
      "epss": "0.0012",
      "percentile": "0.02082"
    },
    "nvd": "{\"cve\":{\"id\":\"CVE-2026-74674\",\"sourceIdentifier\":\"416baaa9-dc9f-4396-8d5f-8c081fb06d67\",\"published\":\"2026-08-22T16:16:41.480\",\"lastModified\":\"2026-08-25T06:18:49.990\",\"vulnStatus\":\"Received\",\"cveTags\":[],\"descriptions\":[{\"lang\":\"en\",\"value\":\"In the Linux kernel, the following vulnerability has been resolved:\\n\\nmm: fix incorrect flush address in direct page table reclaim\\n\\nWhen zap_pte_range reclaims a page table, it does:\\n\\n    pte_free_tlb(tlb, pmd_pgtable(pmdval), addr);\\n\\nand this is unconditionally wrong: if this code executes, addr *always*\\npoints one past the end of the range covered by the table.  The addr\\nparameter is used to flush the TLB (really the paging-structure-cache)\\nto drop references to the to-be-freed table, and any architecture that\\ncares about the parameter will flush the wrong address.  (But they\u0027ll\\nstill free the correct page).\\n\\nI think it\u0027s worth contemplating why the kernel works at all.\\n\\nIf we hit the offending line of code, we will first clear the PMD entry\\n(line 1954, zap_empty_pte_table), then we will issue pending flushes if\\nforce_flush is set (tlb_flush_mmu_tlbonly(tlb)), then we will skip the\\nretry on line 1979 (phew!), and then we will do the offending\\npte_free_tlb call.  *Or* we will clear the PMD entry immediately before\\npte_free_tlb (line 1983, zap_pte_table_if_empty).\\n\\nIf we have any pending flushes (i.e. we actually zapped any last-level\\nentries) at the time we clear the PMD entry, then the flush really ought\\nto flush all references to the table (Linus certainly seems to think it\\nwill on all architectures [0]).\\n\\nThe condition under which we have no accumulated flushes at the time of\\nthe clear is very complex (the whole zap_pte_range function has absurdly\\ncomplex control flow).  If we do hit the bad case, then we will end up\\nclearing the PMD entry after the last time the range is flushed, and any\\nCPU is free to cache a reference to the (empty) page table.  If this\\nhappens due to an ordinary read or write, it would segfault, so it would\\nbe rare.  But the cache could be speculatively filled as well.  Then\\nwe\u0027ll flush the wrong address and then free and possibly reuse the\\ntable.\\n\\nOn x86, even flushing the wrong address works on non-KPTI Intel systems\\nbecause INVLPG flushes *all* paging-structure-caches, not just the ones\\nfor the target address.  But INVPCID does not, and flush_tlb_one_user\\nwill use INVPCID if it\u0027s available.  And then we\u0027re toast.  AMD systems\\nare more susceptible: we set the EFER.TCE bit, which makes even INVLPG\\nonly flush the target address.\\n\\nI think this might fix an issue in ripgrep reported here:\\nhttps://github.com/BurntSushi/ripgrep/issues/3494\\n\\n[0] https://lore.kernel.org/all/CA+55aFzBggoXtNXQeng5d_mRoDnaMBE5Y+URs+PHR67nUpMtaw@mail.gmail.com/T/#u\"}],\"affected\":[{\"source\":\"416baaa9-dc9f-4396-8d5f-8c081fb06d67\",\"affectedData\":[{\"vendor\":\"Linux\",\"product\":\"Linux\",\"defaultStatus\":\"unaffected\",\"programFiles\":[\"mm/memory.c\"],\"repo\":\"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git\",\"versions\":[{\"version\":\"4c640eb4181cf8de74c8b9e7c9cf16bf8d26b73e\",\"lessThan\":\"0b8ff21cbda8808c86b18f1b0ca2d0025af9a80a\",\"versionType\":\"git\",\"status\":\"affected\"},{\"version\":\"4c640eb4181cf8de74c8b9e7c9cf16bf8d26b73e\",\"lessThan\":\"478a1c3abebfc717db0d1281a9cdd7befafee542\",\"versionType\":\"git\",\"status\":\"affected\"}]},{\"vendor\":\"Linux\",\"product\":\"Linux\",\"defaultStatus\":\"affected\",\"programFiles\":[\"mm/memory.c\"],\"repo\":\"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git\",\"versions\":[{\"version\":\"7.0\",\"status\":\"affected\"},{\"version\":\"0\",\"lessThan\":\"7.0\",\"versionType\":\"semver\",\"status\":\"unaffected\"},{\"version\":\"7.1.9\",\"lessThanOrEqual\":\"7.1.*\",\"versionType\":\"semver\",\"status\":\"unaffected\"},{\"version\":\"7.2\",\"lessThanOrEqual\":\"*\",\"versionType\":\"original_commit_for_fix\",\"status\":\"unaffected\"}]}]}],\"metrics\":{\"cvssMetricV31\":[{\"source\":\"416baaa9-dc9f-4396-8d5f-8c081fb06d67\",\"type\":\"Secondary\",\"cvssData\":{\"version\":\"3.1\",\"vectorString\":\"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H\",\"baseScore\":7.8,\"baseSeverity\":\"HIGH\",\"attackVector\":\"LOCAL\",\"attackComplexity\":\"LOW\",\"privilegesRequired\":\"LOW\",\"userInteraction\":\"NONE\",\"scope\":\"UNCHANGED\",\"confidentialityImpact\":\"HIGH\",\"integrityImpact\":\"HIGH\",\"availabilityImpact\":\"HIGH\"},\"exploitabilityScore\":1.8,\"impactScore\":5.9}]},\"references\":[{\"url\":\"https://git.kernel.org/stable/c/0b8ff21cbda8808c86b18f1b0ca2d0025af9a80a\",\"source\":\"416baaa9-dc9f-4396-8d5f-8c081fb06d67\"},{\"url\":\"https://git.kernel.org/stable/c/478a1c3abebfc717db0d1281a9cdd7befafee542\",\"source\":\"416baaa9-dc9f-4396-8d5f-8c081fb06d67\"}]}}",
    "redhat_vex": {
      "aggregate_severity": "Moderate",
      "current_release_date": "2026-08-25T08:54:04+00:00",
      "cve": "CVE-2026-74674",
      "id": "CVE-2026-74674",
      "initial_release_date": "2026-08-22T00:00:00+00:00",
      "product_status:known_affected": "14",
      "product_status:known_not_affected": "260",
      "source": "Red Hat CSAF VEX",
      "status": "final",
      "title": "kernel: mm: fix incorrect flush address in direct page table reclaim",
      "url": "https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-74674.json",
      "version": "3"
    }
  }
}



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.