Re: KASAN: use-after-free Read in fuse_dev_do_read

From: Kirill Tkhai
Date: Thu Sep 13 2018 - 04:51:41 EST


Hi,

On 13.09.2018 11:00, syzbot wrote:
> Hello,
>
> syzbot found the following crash on:
>
> HEAD commit:ÂÂÂ 54eda9df17f3 Merge tag 'pci-v4.19-fixes-1' of git://git.ke..
> git tree:ÂÂÂÂÂÂ upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=109a3d9e400000

what are requests id written here:

write$FUSE_INIT(r0, &(0x7f0000000100)={0x50, 0x0, 0x1}, 0x50)
write$FUSE_STATFS(r0, &(0x7f0000000040)={0x60, 0x0, 0x2, {{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0xfffffffffffffffa}}}, 0x60)

?

In case of parallel thread called write with id, we just queued to processing
in fuse_dev_do_read(), we may bump to this stack (despite read has not finished
yet, syzbot may write just everything, any id).

fuse_dev_do_read() fuse_dev_do_write()
list_move_tail(&req->list, &fpq->processing); ...
spin_unlock(&fpq->lock); ...
... request_end(fc, req);
... fuse_put_request(fc, req);
if (test_bit(FR_INTERRUPTED, &req->flags))
queue_interrupt(fiq, req);

We should keep req refcount in fuse_dev_do_read(), till we haven't finished to use it.
Something like this (compile tested only):

diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index 11ea2c4a38ab..675caed3e655 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -1311,12 +1311,14 @@ static ssize_t fuse_dev_do_read(struct fuse_dev *fud, struct file *file,
goto out_end;
}
list_move_tail(&req->list, &fpq->processing);
+ __fuse_get_request(req);
spin_unlock(&fpq->lock);
set_bit(FR_SENT, &req->flags);
/* matches barrier in request_wait_answer() */
smp_mb__after_atomic();
if (test_bit(FR_INTERRUPTED, &req->flags))
queue_interrupt(fiq, req);
+ fuse_put_request(fc, req);

return reqsize;



> kernel config:Â https://syzkaller.appspot.com/x/.config?x=b8f349d23d3c4835
> dashboard link: https://syzkaller.appspot.com/bug?extid=4e975615ca01f2277bdd
> compiler:ÂÂÂÂÂÂ gcc (GCC) 8.0.1 20180413 (experimental)
>
> Unfortunately, I don't have any reproducer for this crash yet.
>
> IMPORTANT: if you fix the bug, please add the following tag to the commit:
> Reported-by: syzbot+4e975615ca01f2277bdd@xxxxxxxxxxxxxxxxxxxxxxxxx
>
> ==================================================================
> BUG: KASAN: use-after-free in constant_test_bit arch/x86/include/asm/bitops.h:328 [inline]
> BUG: KASAN: use-after-free in fuse_dev_do_read.isra.27+0x1659/0x1920 fs/fuse/dev.c:1318
> Read of size 8 at addr ffff8801cbd4ea30 by task syz-executor0/28821
>
> CPU: 1 PID: 28821 Comm: syz-executor0 Not tainted 4.19.0-rc3+ #11
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
> Call Trace:
> Â__dump_stack lib/dump_stack.c:77 [inline]
> Âdump_stack+0x1c4/0x2b4 lib/dump_stack.c:113
> Âprint_address_description.cold.8+0x9/0x1ff mm/kasan/report.c:256
> Âkasan_report_error mm/kasan/report.c:354 [inline]
> Âkasan_report.cold.9+0x242/0x309 mm/kasan/report.c:412
> Â__asan_report_load8_noabort+0x14/0x20 mm/kasan/report.c:433
> Âconstant_test_bit arch/x86/include/asm/bitops.h:328 [inline]
> Âfuse_dev_do_read.isra.27+0x1659/0x1920 fs/fuse/dev.c:1318
> Âfuse_dev_read+0x1a9/0x250 fs/fuse/dev.c:1360
> Âcall_read_iter include/linux/fs.h:1801 [inline]
> Ânew_sync_read fs/read_write.c:406 [inline]
> Â__vfs_read+0x6ac/0x9b0 fs/read_write.c:418
> Âvfs_read+0x17f/0x3c0 fs/read_write.c:452
> Âksys_read+0x101/0x260 fs/read_write.c:578
> Â__do_sys_read fs/read_write.c:588 [inline]
> Â__se_sys_read fs/read_write.c:586 [inline]
> Â__x64_sys_read+0x73/0xb0 fs/read_write.c:586
> Âdo_syscall_64+0x1b9/0x820 arch/x86/entry/common.c:290
> Âentry_SYSCALL_64_after_hwframe+0x49/0xbe
> RIP: 0033:0x4572d9
> Code: fd b4 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 0f 83 cb b4 fb ff c3 66 2e 0f 1f 84 00 00 00 00
> RSP: 002b:00007faaeefe4c78 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
> RAX: ffffffffffffffda RBX: 00007faaeefe56d4 RCX: 00000000004572d9
> RDX: 0000000000001000 RSI: 00000000200040c0 RDI: 0000000000000003
> RBP: 00000000009300a0 R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000000 R11: 0000000000000246 R12: 00000000ffffffff
> R13: 00000000004d4508 R14: 00000000004c8d44 R15: 0000000000000000
>
> Allocated by task 28825:
> Âsave_stack+0x43/0xd0 mm/kasan/kasan.c:448
> Âset_track mm/kasan/kasan.c:460 [inline]
> Âkasan_kmalloc+0xc7/0xe0 mm/kasan/kasan.c:553
> Âkasan_slab_alloc+0x12/0x20 mm/kasan/kasan.c:490
> Âkmem_cache_alloc+0x12e/0x730 mm/slab.c:3554
> Â__fuse_request_alloc+0x27/0xf0 fs/fuse/dev.c:58
> Âfuse_request_alloc fs/fuse/dev.c:89 [inline]
> Â__fuse_get_req+0x1f7/0x9e0 fs/fuse/dev.c:164
> Âfuse_get_req fs/fuse/dev.c:194 [inline]
> Âfuse_simple_request+0x28/0x730 fs/fuse/dev.c:549
> Âfuse_statfs+0x368/0x8a0 fs/fuse/inode.c:442
> Âstatfs_by_dentry+0x136/0x210 fs/statfs.c:64
> Âvfs_statfs+0x47/0x2e0 fs/statfs.c:74
> Âuser_statfs+0xbd/0x150 fs/statfs.c:89
> Â__do_sys_statfs+0x7f/0xf0 fs/statfs.c:179
> Â__se_sys_statfs fs/statfs.c:176 [inline]
> Â__x64_sys_statfs+0x54/0x80 fs/statfs.c:176
> Âdo_syscall_64+0x1b9/0x820 arch/x86/entry/common.c:290
> Âentry_SYSCALL_64_after_hwframe+0x49/0xbe
>
> Freed by task 28825:
> Âsave_stack+0x43/0xd0 mm/kasan/kasan.c:448
> Âset_track mm/kasan/kasan.c:460 [inline]
> Â__kasan_slab_free+0x102/0x150 mm/kasan/kasan.c:521
> Âkasan_slab_free+0xe/0x10 mm/kasan/kasan.c:528
> Â__cache_free mm/slab.c:3498 [inline]
> Âkmem_cache_free+0x83/0x290 mm/slab.c:3756
> Âfuse_request_free+0x8b/0xa0 fs/fuse/dev.c:104
> Âfuse_put_request+0x2a6/0x350 fs/fuse/dev.c:304
> Âfuse_simple_request+0x453/0x730 fs/fuse/dev.c:571
> Âfuse_statfs+0x368/0x8a0 fs/fuse/inode.c:442
> Âstatfs_by_dentry+0x136/0x210 fs/statfs.c:64
> Âvfs_statfs+0x47/0x2e0 fs/statfs.c:74
> Âuser_statfs+0xbd/0x150 fs/statfs.c:89
> Â__do_sys_statfs+0x7f/0xf0 fs/statfs.c:179
> Â__se_sys_statfs fs/statfs.c:176 [inline]
> Â__x64_sys_statfs+0x54/0x80 fs/statfs.c:176
> Âdo_syscall_64+0x1b9/0x820 arch/x86/entry/common.c:290
> Âentry_SYSCALL_64_after_hwframe+0x49/0xbe
>
> The buggy address belongs to the object at ffff8801cbd4ea00
> Âwhich belongs to the cache fuse_request of size 448
> The buggy address is located 48 bytes inside of
> Â448-byte region [ffff8801cbd4ea00, ffff8801cbd4ebc0)
> The buggy address belongs to the page:
> page:ffffea00072f5380 count:1 mapcount:0 mapping:ffff8801d4ad4840 index:0x0
> flags: 0x2fffc0000000100(slab)
> raw: 02fffc0000000100 ffffea000717eb48 ffff8801d4ad5c48 ffff8801d4ad4840
> raw: 0000000000000000 ffff8801cbd4e000 0000000100000008 0000000000000000
> page dumped because: kasan: bad access detected
>
> Memory state around the buggy address:
> Âffff8801cbd4e900: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> Âffff8801cbd4e980: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>> ffff8801cbd4ea00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ ^
> Âffff8801cbd4ea80: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> Âffff8801cbd4eb00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> ==================================================================
>
>
> ---
> This bug is generated by a bot. It may contain errors.
> See https://goo.gl/tpsmEJ for more information about syzbot.
> syzbot engineers can be reached at syzkaller@xxxxxxxxxxxxxxxxx
>
> syzbot will keep track of this bug report. See:
> https://goo.gl/tpsmEJ#bug-status-tracking for how to communicate with syzbot.