{"api_version":"1","generated_at":"2026-09-25T08:35:35+00:00","cve":"CVE-2026-93255","urls":{"html":"https://cve.report/CVE-2026-93255","api":"https://cve.report/api/cve/CVE-2026-93255.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-93255","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-93255"},"summary":{"title":"btrfs: make sure EXTENT_BUFFER_READING is cleared under refs_lock","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nbtrfs: make sure EXTENT_BUFFER_READING is cleared under refs_lock\n\n[FALSE ALERTS]\nThere is a bug report that the warning inside\ninvalidate_and_check_btree_folios() got triggered during btrfs/298:\n\n BTRFS info (device sdd): first mount of filesystem f9bf732a-a19b-44b9-99a7-614ddff168e2\n BTRFS info (device sdd): using crc32c checksum algorithm\n BTRFS error (device sdd): failed to find fsid cb2fdb42-b638-4f2f-badd-4127467ba674 when attempting to open seed devices\n BTRFS error (device sdd): failed to read chunk tree: -2\n ------------[ cut here ]------------\n WARNING: disk-io.c:3342 at invalidate_and_check_btree_folios+0x260/0x3c0 [btrfs], CPU#4: mount/125993\n CPU: 4 UID: 0 PID: 125993 Comm: mount Tainted: G        W  OE       7.1.0-rc7-custom+ #1 PREEMPT(full)\n Hardware name: QEMU KVM Virtual Machine, BIOS edk2-20250812-19.fc42 08/12/2025\n Call trace:\n  invalidate_and_check_btree_folios+0x260/0x3c0 [btrfs] (P)\n  open_ctree+0x1f50/0x23b0 [btrfs]\n  btrfs_get_tree+0x89c/0xc48 [btrfs]\n  vfs_get_tree+0x30/0x110\n  vfs_cmd_create+0x58/0xe8\n  __arm64_sys_fsconfig+0x39c/0x518\n  invoke_syscall.constprop.0+0x48/0x120\n  el0_svc_common.constprop.0+0x40/0xe8\n  do_el0_svc+0x24/0x38\n  el0_svc+0x50/0x310\n  el0t_64_sync_handler+0xa0/0xe8\n  el0t_64_sync+0x198/0x1a0\n ---[ end trace 0000000000000000 ]---\n BTRFS warning (device sdd): unable to release extent buffer 365985792 owner 3 gen 17 refs 3 flags 0x5\n\n[CAUSE]\nIn that invalidate_and_check_btree_folios() we wait for the eb to finish\nits read, then check if it's only held by us and the btree inode.\n\nIf not, then do a warning as it may be still held, and could cause\nproblems.\n\nBut there is a small window where the check can lead to false alerts:\n\n    Thread A (Read endio)         |        Thread B (Unmount)\n----------------------------------+-------------------------------------\n end_bbio_meta_read()             |\n | The eb has one extra ref held  |\n | by the reader, and has         |\n | EXTENT_BUFFER_READING flag set | invalidate_and_check_btree_folios()\n |                                | |\n |- clear_extent_buffer_reading() | |\n |                                | |- wait_on_bit_io();\n |                                | |  The EXTENT_BUFFER_READING flag is\n |                                | |  cleared\n |                                | |- if (refcount_read(eb->refs) > 2)\n |                                |    The eb is held by the read, us\n |                                |    and btree inode, thus it\n |                                |    will trigger the warning\n |- free_extent_buffer()          |\n\n[FIX]\nIntroduce a helper, free_extent_buffer_clear_reading().\n\nIf the new parameter, @clear_reading, is set, we will hold the spinlock\nat the beginning of free_extent_buffer_clear_reading() to make sure the\nEXTENT_BUFFER_READING flag is cleared inside the same critical section\nof decreasing refs.\n\nNow free_extent_buffer() will just call\nfree_extent_buffer_clear_reading() with @clear_reading set to false, so\nno behavior change.\n\nBut for end_bbio_meta_read(), it will not clear_extent_buffer_reading()\ndirectly, but pass @clear_reading as true.\n\nThen inside invalidate_and_check_btree_folios(), hold the refs_lock\nbefore reading refs.\nSo that we eliminate the race window completely.","state":"PUBLISHED","assigner":"Linux","published_at":"2026-09-24 16:17:22","updated_at":"2026-09-24 16:17:22"},"problem_types":[],"metrics":[],"references":[{"url":"https://git.kernel.org/stable/c/2073f3d97b0c5d34fe64812d1237e37bd79c17ec","name":"https://git.kernel.org/stable/c/2073f3d97b0c5d34fe64812d1237e37bd79c17ec","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/690c2accacb1aca91ab8186d15dee56da8723f31","name":"https://git.kernel.org/stable/c/690c2accacb1aca91ab8186d15dee56da8723f31","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-93255","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-93255","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 83f7e52b7ed1c3e03b79123e20b6f6adf8d886bb 2073f3d97b0c5d34fe64812d1237e37bd79c17ec git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 83f7e52b7ed1c3e03b79123e20b6f6adf8d886bb 690c2accacb1aca91ab8186d15dee56da8723f31 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected b0c33c0628c5b0678fcae2e05e7b2764149248f8 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 7.1.8 7.2 semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 7.2","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.2 semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.2.6 7.2.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.3-rc1 * original_commit_for_fix","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":null,"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"cna":{"affected":[{"defaultStatus":"unaffected","product":"Linux","programFiles":["fs/btrfs/disk-io.c","fs/btrfs/extent_io.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"lessThan":"2073f3d97b0c5d34fe64812d1237e37bd79c17ec","status":"affected","version":"83f7e52b7ed1c3e03b79123e20b6f6adf8d886bb","versionType":"git"},{"lessThan":"690c2accacb1aca91ab8186d15dee56da8723f31","status":"affected","version":"83f7e52b7ed1c3e03b79123e20b6f6adf8d886bb","versionType":"git"},{"status":"affected","version":"b0c33c0628c5b0678fcae2e05e7b2764149248f8","versionType":"git"},{"lessThan":"7.2","status":"affected","version":"7.1.8","versionType":"semver"}]},{"defaultStatus":"affected","product":"Linux","programFiles":["fs/btrfs/disk-io.c","fs/btrfs/extent_io.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"status":"affected","version":"7.2"},{"lessThan":"7.2","status":"unaffected","version":"0","versionType":"semver"},{"lessThanOrEqual":"7.2.*","status":"unaffected","version":"7.2.6","versionType":"semver"},{"lessThanOrEqual":"*","status":"unaffected","version":"7.3-rc1","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"cpeMatch":[{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.2.6","versionStartIncluding":"7.2","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.3-rc1","versionStartIncluding":"7.2","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"7.1.8","vulnerable":true}],"negate":false,"operator":"OR"}]}],"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nbtrfs: make sure EXTENT_BUFFER_READING is cleared under refs_lock\n\n[FALSE ALERTS]\nThere is a bug report that the warning inside\ninvalidate_and_check_btree_folios() got triggered during btrfs/298:\n\n BTRFS info (device sdd): first mount of filesystem f9bf732a-a19b-44b9-99a7-614ddff168e2\n BTRFS info (device sdd): using crc32c checksum algorithm\n BTRFS error (device sdd): failed to find fsid cb2fdb42-b638-4f2f-badd-4127467ba674 when attempting to open seed devices\n BTRFS error (device sdd): failed to read chunk tree: -2\n ------------[ cut here ]------------\n WARNING: disk-io.c:3342 at invalidate_and_check_btree_folios+0x260/0x3c0 [btrfs], CPU#4: mount/125993\n CPU: 4 UID: 0 PID: 125993 Comm: mount Tainted: G        W  OE       7.1.0-rc7-custom+ #1 PREEMPT(full)\n Hardware name: QEMU KVM Virtual Machine, BIOS edk2-20250812-19.fc42 08/12/2025\n Call trace:\n  invalidate_and_check_btree_folios+0x260/0x3c0 [btrfs] (P)\n  open_ctree+0x1f50/0x23b0 [btrfs]\n  btrfs_get_tree+0x89c/0xc48 [btrfs]\n  vfs_get_tree+0x30/0x110\n  vfs_cmd_create+0x58/0xe8\n  __arm64_sys_fsconfig+0x39c/0x518\n  invoke_syscall.constprop.0+0x48/0x120\n  el0_svc_common.constprop.0+0x40/0xe8\n  do_el0_svc+0x24/0x38\n  el0_svc+0x50/0x310\n  el0t_64_sync_handler+0xa0/0xe8\n  el0t_64_sync+0x198/0x1a0\n ---[ end trace 0000000000000000 ]---\n BTRFS warning (device sdd): unable to release extent buffer 365985792 owner 3 gen 17 refs 3 flags 0x5\n\n[CAUSE]\nIn that invalidate_and_check_btree_folios() we wait for the eb to finish\nits read, then check if it's only held by us and the btree inode.\n\nIf not, then do a warning as it may be still held, and could cause\nproblems.\n\nBut there is a small window where the check can lead to false alerts:\n\n    Thread A (Read endio)         |        Thread B (Unmount)\n----------------------------------+-------------------------------------\n end_bbio_meta_read()             |\n | The eb has one extra ref held  |\n | by the reader, and has         |\n | EXTENT_BUFFER_READING flag set | invalidate_and_check_btree_folios()\n |                                | |\n |- clear_extent_buffer_reading() | |\n |                                | |- wait_on_bit_io();\n |                                | |  The EXTENT_BUFFER_READING flag is\n |                                | |  cleared\n |                                | |- if (refcount_read(eb->refs) > 2)\n |                                |    The eb is held by the read, us\n |                                |    and btree inode, thus it\n |                                |    will trigger the warning\n |- free_extent_buffer()          |\n\n[FIX]\nIntroduce a helper, free_extent_buffer_clear_reading().\n\nIf the new parameter, @clear_reading, is set, we will hold the spinlock\nat the beginning of free_extent_buffer_clear_reading() to make sure the\nEXTENT_BUFFER_READING flag is cleared inside the same critical section\nof decreasing refs.\n\nNow free_extent_buffer() will just call\nfree_extent_buffer_clear_reading() with @clear_reading set to false, so\nno behavior change.\n\nBut for end_bbio_meta_read(), it will not clear_extent_buffer_reading()\ndirectly, but pass @clear_reading as true.\n\nThen inside invalidate_and_check_btree_folios(), hold the refs_lock\nbefore reading refs.\nSo that we eliminate the race window completely."}],"providerMetadata":{"dateUpdated":"2026-09-24T15:51:42.056Z","orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux"},"references":[{"url":"https://git.kernel.org/stable/c/2073f3d97b0c5d34fe64812d1237e37bd79c17ec"},{"url":"https://git.kernel.org/stable/c/690c2accacb1aca91ab8186d15dee56da8723f31"}],"title":"btrfs: make sure EXTENT_BUFFER_READING is cleared under refs_lock","x_generator":{"engine":"bippy-1.2.0"}}},"cveMetadata":{"assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","assignerShortName":"Linux","cveId":"CVE-2026-93255","datePublished":"2026-09-24T15:51:42.056Z","dateReserved":"2026-09-17T16:02:15.096Z","dateUpdated":"2026-09-24T15:51:42.056Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-09-24 16:17:22","lastModifiedDate":"2026-09-24 16:17:22","problem_types":[],"metrics":[],"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"93255","Ordinal":"1","Title":"btrfs: make sure EXTENT_BUFFER_READING is cleared under refs_loc","CVE":"CVE-2026-93255","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"93255","Ordinal":"1","NoteData":"In the Linux kernel, the following vulnerability has been resolved:\n\nbtrfs: make sure EXTENT_BUFFER_READING is cleared under refs_lock\n\n[FALSE ALERTS]\nThere is a bug report that the warning inside\ninvalidate_and_check_btree_folios() got triggered during btrfs/298:\n\n BTRFS info (device sdd): first mount of filesystem f9bf732a-a19b-44b9-99a7-614ddff168e2\n BTRFS info (device sdd): using crc32c checksum algorithm\n BTRFS error (device sdd): failed to find fsid cb2fdb42-b638-4f2f-badd-4127467ba674 when attempting to open seed devices\n BTRFS error (device sdd): failed to read chunk tree: -2\n ------------[ cut here ]------------\n WARNING: disk-io.c:3342 at invalidate_and_check_btree_folios+0x260/0x3c0 [btrfs], CPU#4: mount/125993\n CPU: 4 UID: 0 PID: 125993 Comm: mount Tainted: G        W  OE       7.1.0-rc7-custom+ #1 PREEMPT(full)\n Hardware name: QEMU KVM Virtual Machine, BIOS edk2-20250812-19.fc42 08/12/2025\n Call trace:\n  invalidate_and_check_btree_folios+0x260/0x3c0 [btrfs] (P)\n  open_ctree+0x1f50/0x23b0 [btrfs]\n  btrfs_get_tree+0x89c/0xc48 [btrfs]\n  vfs_get_tree+0x30/0x110\n  vfs_cmd_create+0x58/0xe8\n  __arm64_sys_fsconfig+0x39c/0x518\n  invoke_syscall.constprop.0+0x48/0x120\n  el0_svc_common.constprop.0+0x40/0xe8\n  do_el0_svc+0x24/0x38\n  el0_svc+0x50/0x310\n  el0t_64_sync_handler+0xa0/0xe8\n  el0t_64_sync+0x198/0x1a0\n ---[ end trace 0000000000000000 ]---\n BTRFS warning (device sdd): unable to release extent buffer 365985792 owner 3 gen 17 refs 3 flags 0x5\n\n[CAUSE]\nIn that invalidate_and_check_btree_folios() we wait for the eb to finish\nits read, then check if it's only held by us and the btree inode.\n\nIf not, then do a warning as it may be still held, and could cause\nproblems.\n\nBut there is a small window where the check can lead to false alerts:\n\n    Thread A (Read endio)         |        Thread B (Unmount)\n----------------------------------+-------------------------------------\n end_bbio_meta_read()             |\n | The eb has one extra ref held  |\n | by the reader, and has         |\n | EXTENT_BUFFER_READING flag set | invalidate_and_check_btree_folios()\n |                                | |\n |- clear_extent_buffer_reading() | |\n |                                | |- wait_on_bit_io();\n |                                | |  The EXTENT_BUFFER_READING flag is\n |                                | |  cleared\n |                                | |- if (refcount_read(eb->refs) > 2)\n |                                |    The eb is held by the read, us\n |                                |    and btree inode, thus it\n |                                |    will trigger the warning\n |- free_extent_buffer()          |\n\n[FIX]\nIntroduce a helper, free_extent_buffer_clear_reading().\n\nIf the new parameter, @clear_reading, is set, we will hold the spinlock\nat the beginning of free_extent_buffer_clear_reading() to make sure the\nEXTENT_BUFFER_READING flag is cleared inside the same critical section\nof decreasing refs.\n\nNow free_extent_buffer() will just call\nfree_extent_buffer_clear_reading() with @clear_reading set to false, so\nno behavior change.\n\nBut for end_bbio_meta_read(), it will not clear_extent_buffer_reading()\ndirectly, but pass @clear_reading as true.\n\nThen inside invalidate_and_check_btree_folios(), hold the refs_lock\nbefore reading refs.\nSo that we eliminate the race window completely.","Type":"Description","Title":"btrfs: make sure EXTENT_BUFFER_READING is cleared under refs_loc"}]}}}