CVE-2026-97945

CVE CVE-2026-97945EUVD EUVD-2026-86799Published 2026-09-25T10:22:55.000Z

What the advisory describes

In the Linux kernel, the following vulnerability has been resolved: x86/mm: Fix user-space data loss with MADV_FREE and THP Some of users of Polars (a data analytics library) have lost production data from this bug. They seem to have just the right combination of huge pages, MADV_FREE and heavy reclaim pressure. pmd_modify() masks the old value with (_HPAGE_CHG_MASK & ~_PAGE_DIRTY), silently discarding the hardware dirty bit. The subsequent pmd_mksaveddirty() call is supposed to transfer _PAGE_DIRTY into _PAGE_SAVED_DIRTY when write-protecting, but the dirty bit was already stripped from the value, so there is nothing left to transfer. Contrast with pte_modify(), which keeps _PAGE_DIRTY_BITS in its mask, and pud_modify(), which keeps _HPAGE_CHG_MASK untouched: pmd_modify() is the odd one out. Any pmd_modify() on a writable, dirty PMD loses the dirty state. One visible consequence is data loss with MADV_FREE on PMD-mapped THP: memset(buf, 0x5A, size); // PMD-mapped THP, PMD dirty madvise(buf, size, MADV_FREE); // PMD cleaned but left writable, // folio marked lazyfree memset(buf, 0x5A, size); // hardware sets _PAGE_DIRTY again mprotect(buf, size, PROT_READ); // pmd_modify() drops the dirty bit mprotect(buf, size, PROT_READ|PROT_WRITE); // ... memory pressure ... Reclaim (e.g. under memcg pressure) then finds the lazyfree folio with no dirty bit set anywhere and frees it in __discard_anon_folio_pmd_locked(), even though the data was rewritten after MADV_FREE; subsequent reads fault in fresh zero pages. NUMA hinting alone can trigger the same loss, as do_huge_pmd_numa_page() restores the PMD through pmd_modify() as well. PMD-mapped file THPs are affected too: mprotect()/NUMA hinting dropping the dirty bit means rewritten data is never written back. Fix it by keeping _PAGE_DIRTY in the preserved mask, exactly like pte_modify() and pud_modify() do. The existing pmd_mksaveddirty()/pmd_clear_saveddirty() pair then performs the hardware-dirty <-> saved-dirty transition based on the write bit, preserving the shadow-stack encoding rules.

Source: EUVD (ENISA), in the words of the advisory.

Products the advisory names

These come from the advisory itself, not from any check we performed.

• Linux — Linux bb3aadf7d446aaf22c725b274e2c194ac5cb2111 <e4d569fe2f436a9b58c048fefabf56eab68c1b7e; patch: 7.3-rc4; bb3aadf7d446aaf22c725b274e2c194ac5cb2111 <a5ad97cbcfcec31430268441c10eedde73200177; patch: 6.12.111; patch: 6.18.53; bb3aadf7d446aaf22c725b274e2c194ac5cb2111 <55ddbc2ca6d5db64d44c3fb8f8705b4d39eb30b7; patch: 0; 6.6; patch: 7.2.7; bb3aadf7d446aaf22c725b274e2c194ac5cb2111 <f7491d7c81db0e7c304a7bd757a76d2fbeaff80e

The versions shown are the advisory's own. Patchlage compares no version numbers and derives no judgement from them — which version is installed is something a person has to look up.

Carried in the product catalogue

An estate covering these products can be recorded in Patchlage. An advisory about them appears in the next morning's situation report.

• Linux — Linux

Patchlage reports hits and suspected hits. About everything else this system says nothing — neither this page nor the situation report ever claims that an estate is safe.