CVE-2025-21658

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> btrfs: avoid NULL pointer dereference if no valid extent tree<br /> <br /> [BUG]<br /> Syzbot reported a crash with the following call trace:<br /> <br /> BTRFS info (device loop0): scrub: started on devid 1<br /> BUG: kernel NULL pointer dereference, address: 0000000000000208<br /> #PF: supervisor read access in kernel mode<br /> #PF: error_code(0x0000) - not-present page<br /> PGD 106e70067 P4D 106e70067 PUD 107143067 PMD 0<br /> Oops: Oops: 0000 [#1] PREEMPT SMP NOPTI<br /> CPU: 1 UID: 0 PID: 689 Comm: repro Kdump: loaded Tainted: G O 6.13.0-rc4-custom+ #206<br /> Tainted: [O]=OOT_MODULE<br /> Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS unknown 02/02/2022<br /> RIP: 0010:find_first_extent_item+0x26/0x1f0 [btrfs]<br /> Call Trace:<br /> <br /> scrub_find_fill_first_stripe+0x13d/0x3b0 [btrfs]<br /> scrub_simple_mirror+0x175/0x260 [btrfs]<br /> scrub_stripe+0x5d4/0x6c0 [btrfs]<br /> scrub_chunk+0xbb/0x170 [btrfs]<br /> scrub_enumerate_chunks+0x2f4/0x5f0 [btrfs]<br /> btrfs_scrub_dev+0x240/0x600 [btrfs]<br /> btrfs_ioctl+0x1dc8/0x2fa0 [btrfs]<br /> ? do_sys_openat2+0xa5/0xf0<br /> __x64_sys_ioctl+0x97/0xc0<br /> do_syscall_64+0x4f/0x120<br /> entry_SYSCALL_64_after_hwframe+0x76/0x7e<br /> <br /> <br /> [CAUSE]<br /> The reproducer is using a corrupted image where extent tree root is<br /> corrupted, thus forcing to use "rescue=all,ro" mount option to mount the<br /> image.<br /> <br /> Then it triggered a scrub, but since scrub relies on extent tree to find<br /> where the data/metadata extents are, scrub_find_fill_first_stripe()<br /> relies on an non-empty extent root.<br /> <br /> But unfortunately scrub_find_fill_first_stripe() doesn&amp;#39;t really expect<br /> an NULL pointer for extent root, it use extent_root to grab fs_info and<br /> triggered a NULL pointer dereference.<br /> <br /> [FIX]<br /> Add an extra check for a valid extent root at the beginning of<br /> scrub_find_fill_first_stripe().<br /> <br /> The new error path is introduced by 42437a6386ff ("btrfs: introduce<br /> mount option rescue=ignorebadroots"), but that&amp;#39;s pretty old, and later<br /> commit b979547513ff ("btrfs: scrub: introduce helper to find and fill<br /> sector info for a scrub_stripe") changed how we do scrub.<br /> <br /> So for kernels older than 6.6, the fix will need manual backport.