{ "dataType": "CVE_RECORD", "dataVersion": "5.2", "cveMetadata": { "cveId": "CVE-2025-22115", "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "state": "PUBLISHED", "assignerShortName": "Linux", "dateReserved": "2024-12-29T08:45:45.823Z", "datePublished": "2025-04-16T14:13:01.293Z", "dateUpdated": "2026-08-05T11:56:53.298Z" }, "containers": { "cna": { "providerMetadata": { "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "shortName": "Linux", "dateUpdated": "2026-08-05T11:56:53.298Z" }, "descriptions": [ { "lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nbtrfs: fix block group refcount race in btrfs_create_pending_block_groups()\n\nBlock group creation is done in two phases, which results in a slightly\nunintuitive property: a block group can be allocated/deallocated from\nafter btrfs_make_block_group() adds it to the space_info with\nbtrfs_add_bg_to_space_info(), but before creation is completely completed\nin btrfs_create_pending_block_groups(). As a result, it is possible for a\nblock group to go unused and have 'btrfs_mark_bg_unused' called on it\nconcurrently with 'btrfs_create_pending_block_groups'. This causes a\nnumber of issues, which were fixed with the block group flag\n'BLOCK_GROUP_FLAG_NEW'.\n\nHowever, this fix is not quite complete. Since it does not use the\nunused_bg_lock, it is possible for the following race to occur:\n\nbtrfs_create_pending_block_groups btrfs_mark_bg_unused\n if list_empty // false\n list_del_init\n clear_bit\n else if (test_bit) // true\n list_move_tail\n\nAnd we get into the exact same broken ref count and invalid new_bgs\nstate for transaction cleanup that BLOCK_GROUP_FLAG_NEW was designed to\nprevent.\n\nThe broken refcount aspect will result in a warning like:\n\n [1272.943527] refcount_t: underflow; use-after-free.\n [1272.943967] WARNING: CPU: 1 PID: 61 at lib/refcount.c:28 refcount_warn_saturate+0xba/0x110\n [1272.944731] Modules linked in: btrfs virtio_net xor zstd_compress raid6_pq null_blk [last unloaded: btrfs]\n [1272.945550] CPU: 1 UID: 0 PID: 61 Comm: kworker/u32:1 Kdump: loaded Tainted: G W 6.14.0-rc5+ #108\n [1272.946368] Tainted: [W]=WARN\n [1272.946585] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Arch Linux 1.16.3-1-1 04/01/2014\n [1272.947273] Workqueue: btrfs_discard btrfs_discard_workfn [btrfs]\n [1272.947788] RIP: 0010:refcount_warn_saturate+0xba/0x110\n [1272.949532] RSP: 0018:ffffbf1200247df0 EFLAGS: 00010282\n [1272.949901] RAX: 0000000000000000 RBX: ffffa14b00e3f800 RCX: 0000000000000000\n [1272.950437] RDX: 0000000000000000 RSI: ffffbf1200247c78 RDI: 00000000ffffdfff\n [1272.950986] RBP: ffffa14b00dc2860 R08: 00000000ffffdfff R09: ffffffff90526268\n [1272.951512] R10: ffffffff904762c0 R11: 0000000063666572 R12: ffffa14b00dc28c0\n [1272.952024] R13: 0000000000000000 R14: ffffa14b00dc2868 R15: 000001285dcd12c0\n [1272.952850] FS: 0000000000000000(0000) GS:ffffa14d33c40000(0000) knlGS:0000000000000000\n [1272.953458] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033\n [1272.953931] CR2: 00007f838cbda000 CR3: 000000010104e000 CR4: 00000000000006f0\n [1272.954474] Call Trace:\n [1272.954655] \n [1272.954812] ? refcount_warn_saturate+0xba/0x110\n [1272.955173] ? __warn.cold+0x93/0xd7\n [1272.955487] ? refcount_warn_saturate+0xba/0x110\n [1272.955816] ? report_bug+0xe7/0x120\n [1272.956103] ? handle_bug+0x53/0x90\n [1272.956424] ? exc_invalid_op+0x13/0x60\n [1272.956700] ? asm_exc_invalid_op+0x16/0x20\n [1272.957011] ? refcount_warn_saturate+0xba/0x110\n [1272.957399] btrfs_discard_cancel_work.cold+0x26/0x2b [btrfs]\n [1272.957853] btrfs_put_block_group.cold+0x5d/0x8e [btrfs]\n [1272.958289] btrfs_discard_workfn+0x194/0x380 [btrfs]\n [1272.958729] process_one_work+0x130/0x290\n [1272.959026] worker_thread+0x2ea/0x420\n [1272.959335] ? __pfx_worker_thread+0x10/0x10\n [1272.959644] kthread+0xd7/0x1c0\n [1272.959872] ? __pfx_kthread+0x10/0x10\n [1272.960172] ret_from_fork+0x30/0x50\n [1272.960474] ? __pfx_kthread+0x10/0x10\n [1272.960745] ret_from_fork_asm+0x1a/0x30\n [1272.961035] \n [1272.961238] ---[ end trace 0000000000000000 ]---\n\nThough we have seen them in the async discard workfn as well. It is\nmost likely to happen after a relocation finishes which cancels discard,\ntears down the block group, etc.\n\nFix this fully by taking the lock arou\n---truncated---" } ], "metrics": [ { "cvssV3_1": { "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" }, "scenarios": [ { "lang": "en", "value": "AV:L - The vulnerability is reached through ordinary filesystem write/delete syscalls against a mounted btrfs filesystem, requiring a local account on the system. There is no network-facing entry point into btrfs chunk/block-group allocation.\nAC:L - The attacker drives both sides of the race — one thread forces new chunk/block-group creation via writes, another frees the last extents in that group so btrfs_mark_bg_unused() fires — and can retry the interleaving indefinitely with high thread concurrency. No special kernel config or mount option is required, since the default (non-discard=async) btrfs_update_block_group() path calls btrfs_mark_bg_unused() directly.\nPR:L - Only write access to any btrfs filesystem is needed, which an unprivileged local user has on the common case of a btrfs root/home/tmp; no capability, CAP_SYS_ADMIN, or ioctl gate exists on this path. No privileged operation such as balance is required — the generic extent-free path suffices.\nUI:N - The filesystem is already mounted in the normal case and the attacker triggers everything itself with plain write/unlink activity. No victim action of any kind is needed.\nS:U - The corrupted object and all its consequences are inside the kernel's own security authority; there is no crossing of a VM, IOMMU, or sandbox boundary. This is a standard kernel-local memory-safety issue.\nC:H - The premature refcount drop frees struct btrfs_block_group while it remains reachable from the block-group rbtree, space_info lists, and discard/dirty lists, so kernel workers keep reading attacker-reclaimable slab memory. Since the object comes from a generic kmalloc cache, the attacker can spray it and turn the stale dereferences into an arbitrary kernel-memory read primitive.\nI:H - The freed-but-still-linked block group is later written through and unlinked (list_del/list_move on bg_list, discard_list, dirty_list), giving a classic unlink write-what-where once the slot is reclaimed with attacker-controlled data. Its embedded pointers (free_space_ctl, physical_map, inode, caching_ctl) are dereferenced afterwards, making control-flow hijack and privilege escalation feasible.\nA:H - Even unexploited, the bug produces refcount underflow WARNs, corrupted new_bgs/unused_bgs list state, and use-after-free of a structure central to allocation, which reliably leads to oops/panic and an unusable filesystem. The reported crash occurs in a btrfs kernel worker thread, taking down the whole system rather than just the attacking task." } ] } ], "affected": [ { "product": "Linux", "vendor": "Linux", "defaultStatus": "unaffected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "fs/btrfs/block-group.c" ], "versions": [ { "version": "0657b20c5a76c938612f8409735a8830d257866e", "lessThan": "ee56da95f8962b86fec4ef93f866e64c8d025a58", "status": "affected", "versionType": "git" }, { "version": "0657b20c5a76c938612f8409735a8830d257866e", "lessThan": "9d383a6fc59271aaaf07a33b23b2eac5b9268b7a", "status": "affected", "versionType": "git" }, { "version": "0657b20c5a76c938612f8409735a8830d257866e", "lessThan": "2d8e5168d48a91e7a802d3003e72afb4304bebfa", "status": "affected", "versionType": "git" }, { "version": "6297644db23f77c02ae7961cc542d162629ae2c4", "status": "affected", "versionType": "git" }, { "version": "7569c4294ba6ff9f194635b14876198f8a687c4a", "status": "affected", "versionType": "git" }, { "version": "6.1.47", "lessThan": "6.2", "status": "affected", "versionType": "semver" }, { "version": "6.4.12", "lessThan": "6.5", "status": "affected", "versionType": "semver" } ] }, { "product": "Linux", "vendor": "Linux", "defaultStatus": "affected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "fs/btrfs/block-group.c" ], "versions": [ { "version": "6.5", "status": "affected" }, { "version": "0", "lessThan": "6.5", "status": "unaffected", "versionType": "semver" }, { "version": "6.12.40", "lessThanOrEqual": "6.12.*", "status": "unaffected", "versionType": "semver" }, { "version": "6.14.2", "lessThanOrEqual": "6.14.*", "status": "unaffected", "versionType": "semver" }, { "version": "6.15", "lessThanOrEqual": "*", "status": "unaffected", "versionType": "original_commit_for_fix" } ] } ], "cpeApplicability": [ { "nodes": [ { "operator": "OR", "negate": false, "cpeMatch": [ { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.5", "versionEndExcluding": "6.12.40" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.5", "versionEndExcluding": "6.14.2" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.5", "versionEndExcluding": "6.15" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.1.47" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.4.12" } ] } ] } ], "references": [ { "url": "https://git.kernel.org/stable/c/ee56da95f8962b86fec4ef93f866e64c8d025a58" }, { "url": "https://git.kernel.org/stable/c/9d383a6fc59271aaaf07a33b23b2eac5b9268b7a" }, { "url": "https://git.kernel.org/stable/c/2d8e5168d48a91e7a802d3003e72afb4304bebfa" } ], "title": "btrfs: fix block group refcount race in btrfs_create_pending_block_groups()", "x_generator": { "engine": "bippy-1.2.0" } } } }