CVE-2021-46958

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> btrfs: fix race between transaction aborts and fsyncs leading to use-after-free<br /> <br /> There is a race between a task aborting a transaction during a commit,<br /> a task doing an fsync and the transaction kthread, which leads to an<br /> use-after-free of the log root tree. When this happens, it results in a<br /> stack trace like the following:<br /> <br /> BTRFS info (device dm-0): forced readonly<br /> BTRFS warning (device dm-0): Skipping commit of aborted transaction.<br /> BTRFS: error (device dm-0) in cleanup_transaction:1958: errno=-5 IO failure<br /> BTRFS warning (device dm-0): lost page write due to IO error on /dev/mapper/error-test (-5)<br /> BTRFS warning (device dm-0): Skipping commit of aborted transaction.<br /> BTRFS warning (device dm-0): direct IO failed ino 261 rw 0,0 sector 0xa4e8 len 4096 err no 10<br /> BTRFS error (device dm-0): error writing primary super block to device 1<br /> BTRFS warning (device dm-0): direct IO failed ino 261 rw 0,0 sector 0x12e000 len 4096 err no 10<br /> BTRFS warning (device dm-0): direct IO failed ino 261 rw 0,0 sector 0x12e008 len 4096 err no 10<br /> BTRFS warning (device dm-0): direct IO failed ino 261 rw 0,0 sector 0x12e010 len 4096 err no 10<br /> BTRFS: error (device dm-0) in write_all_supers:4110: errno=-5 IO failure (1 errors while writing supers)<br /> BTRFS: error (device dm-0) in btrfs_sync_log:3308: errno=-5 IO failure<br /> general protection fault, probably for non-canonical address 0x6b6b6b6b6b6b6b68: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC PTI<br /> CPU: 2 PID: 2458471 Comm: fsstress Not tainted 5.12.0-rc5-btrfs-next-84 #1<br /> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.14.0-0-g155821a1990b-prebuilt.qemu.org 04/01/2014<br /> RIP: 0010:__mutex_lock+0x139/0xa40<br /> Code: c0 74 19 (...)<br /> RSP: 0018:ffff9f18830d7b00 EFLAGS: 00010202<br /> RAX: 6b6b6b6b6b6b6b68 RBX: 0000000000000001 RCX: 0000000000000002<br /> RDX: ffffffffb9c54d13 RSI: 0000000000000000 RDI: 0000000000000000<br /> RBP: ffff9f18830d7bc0 R08: 0000000000000000 R09: 0000000000000000<br /> R10: ffff9f18830d7be0 R11: 0000000000000001 R12: ffff8c6cd199c040<br /> R13: ffff8c6c95821358 R14: 00000000fffffffb R15: ffff8c6cbcf01358<br /> FS: 00007fa9140c2b80(0000) GS:ffff8c6fac600000(0000) knlGS:0000000000000000<br /> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033<br /> CR2: 00007fa913d52000 CR3: 000000013d2b4003 CR4: 0000000000370ee0<br /> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000<br /> DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400<br /> Call Trace:<br /> ? __btrfs_handle_fs_error+0xde/0x146 [btrfs]<br /> ? btrfs_sync_log+0x7c1/0xf20 [btrfs]<br /> ? btrfs_sync_log+0x7c1/0xf20 [btrfs]<br /> btrfs_sync_log+0x7c1/0xf20 [btrfs]<br /> btrfs_sync_file+0x40c/0x580 [btrfs]<br /> do_fsync+0x38/0x70<br /> __x64_sys_fsync+0x10/0x20<br /> do_syscall_64+0x33/0x80<br /> entry_SYSCALL_64_after_hwframe+0x44/0xae<br /> RIP: 0033:0x7fa9142a55c3<br /> Code: 8b 15 09 (...)<br /> RSP: 002b:00007fff26278d48 EFLAGS: 00000246 ORIG_RAX: 000000000000004a<br /> RAX: ffffffffffffffda RBX: 0000563c83cb4560 RCX: 00007fa9142a55c3<br /> RDX: 00007fff26278cb0 RSI: 00007fff26278cb0 RDI: 0000000000000005<br /> RBP: 0000000000000005 R08: 0000000000000001 R09: 00007fff26278d5c<br /> R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000340<br /> R13: 00007fff26278de0 R14: 00007fff26278d96 R15: 0000563c83ca57c0<br /> Modules linked in: btrfs dm_zero dm_snapshot dm_thin_pool (...)<br /> ---[ end trace ee2f1b19327d791d ]---<br /> <br /> The steps that lead to this crash are the following:<br /> <br /> 1) We are at transaction N;<br /> <br /> 2) We have two tasks with a transaction handle attached to transaction N.<br /> Task A and Task B. Task B is doing an fsync;<br /> <br /> 3) Task B is at btrfs_sync_log(), and has saved fs_info-&gt;log_root_tree<br /> into a local variable named &amp;#39;log_root_tree&amp;#39; at the top of<br /> btrfs_sync_log(). Task B is about to call write_all_supers(), but<br /> before that...<br /> <br /> 4) Task A calls btrfs_commit_transaction(), and after it sets the<br /> transaction state to TRANS_STATE_COMMIT_START, an error happens before<br /> it w<br /> ---truncated---