Re: KASAN: invalid-free in io_req_caches_free

[Jens Axboe]

On 2/22/21 7:25 PM, Hillf Danton wrote:
> Mon, 22 Feb 2021 01:44:21 -0800
>> Hello,
>>
>> syzbot found the following issue on:
>>
>> HEAD commit:    31caf8b2 Merge branch 'linus' of git://git.kernel.org/pub/..
>> git tree:       upstream
>> console output: https://syzkaller.appspot.com/x/log.txt?x=15a8afa6d00000
>> kernel config:  https://syzkaller.appspot.com/x/.config?x=5a8f3a57fabb4015
>> dashboard link: https://syzkaller.appspot.com/bug?extid=30b4936dcdb3aafa4fb4
>>
>> Unfortunately, I don't have any reproducer for this issue yet.
>>
>> IMPORTANT: if you fix the issue, please add the following tag to the commit:
>> Reported-by: syzbot+30b4936dcdb3aafa4fb4@xxxxxxxxxxxxxxxxxxxxxxxxx
>>
>> ==================================================================
>> BUG: KASAN: double-free or invalid-free in io_req_caches_free.constprop.0+0x3ce/0x530 fs/io_uring.c:8709
>>
>> CPU: 1 PID: 243 Comm: kworker/u4:6 Not tainted 5.11.0-syzkaller #0
>> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
>> Workqueue: events_unbound io_ring_exit_work
>> Call Trace:
>>  __dump_stack lib/dump_stack.c:79 [inline]
>>  dump_stack+0xfa/0x151 lib/dump_stack.c:120
>>  print_address_description.constprop.0.cold+0x5b/0x2c6 mm/kasan/report.c:230
>>  kasan_report_invalid_free+0x51/0x80 mm/kasan/report.c:355
>>  ____kasan_slab_free+0xcc/0xe0 mm/kasan/common.c:341
>>  kasan_slab_free include/linux/kasan.h:192 [inline]
>>  __cache_free mm/slab.c:3424 [inline]
>>  kmem_cache_free_bulk+0x4b/0x1b0 mm/slab.c:3744
>>  io_req_caches_free.constprop.0+0x3ce/0x530 fs/io_uring.c:8709
>>  io_ring_ctx_free fs/io_uring.c:8764 [inline]
>>  io_ring_exit_work+0x518/0x6b0 fs/io_uring.c:8846
>>  process_one_work+0x98d/0x1600 kernel/workqueue.c:2275
>>  worker_thread+0x64c/0x1120 kernel/workqueue.c:2421
>>  kthread+0x3b1/0x4a0 kernel/kthread.c:292
>>  ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:294
>>
>> Allocated by task 11900:
>>  kasan_save_stack+0x1b/0x40 mm/kasan/common.c:38
>>  kasan_set_track mm/kasan/common.c:46 [inline]
>>  set_alloc_info mm/kasan/common.c:401 [inline]
>>  ____kasan_kmalloc.constprop.0+0x7f/0xa0 mm/kasan/common.c:429
>>  kasan_slab_alloc include/linux/kasan.h:209 [inline]
>>  slab_post_alloc_hook mm/slab.h:512 [inline]
>>  kmem_cache_alloc_bulk+0x2c2/0x460 mm/slab.c:3534
>>  io_alloc_req fs/io_uring.c:2014 [inline]
>>  io_submit_sqes+0x18e8/0x2b60 fs/io_uring.c:6915
>>  __do_sys_io_uring_enter+0x1154/0x1f50 fs/io_uring.c:9454
>>  do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46
>>  entry_SYSCALL_64_after_hwframe+0x44/0xae
>>
>> Freed by task 11900:
>>  kasan_save_stack+0x1b/0x40 mm/kasan/common.c:38
>>  kasan_set_track+0x1c/0x30 mm/kasan/common.c:46
>>  kasan_set_free_info+0x20/0x30 mm/kasan/generic.c:356
>>  ____kasan_slab_free+0xb0/0xe0 mm/kasan/common.c:362
>>  kasan_slab_free include/linux/kasan.h:192 [inline]
>>  __cache_free mm/slab.c:3424 [inline]
>>  kmem_cache_free_bulk+0x4b/0x1b0 mm/slab.c:3744
>>  io_req_caches_free.constprop.0+0x3ce/0x530 fs/io_uring.c:8709
>>  io_uring_flush+0x483/0x6e0 fs/io_uring.c:9237
>>  filp_close+0xb4/0x170 fs/open.c:1286
>>  close_files fs/file.c:403 [inline]
>>  put_files_struct fs/file.c:418 [inline]
>>  put_files_struct+0x1d0/0x350 fs/file.c:415
>>  exit_files+0x7e/0xa0 fs/file.c:435
>>  do_exit+0xc27/0x2ae0 kernel/exit.c:820
>>  do_group_exit+0x125/0x310 kernel/exit.c:922
>>  get_signal+0x42c/0x2100 kernel/signal.c:2773
>>  arch_do_signal_or_restart+0x2a8/0x1eb0 arch/x86/kernel/signal.c:811
>>  handle_signal_work kernel/entry/common.c:147 [inline]
>>  exit_to_user_mode_loop kernel/entry/common.c:171 [inline]
>>  exit_to_user_mode_prepare+0x148/0x250 kernel/entry/common.c:208
>>  __syscall_exit_to_user_mode_work kernel/entry/common.c:290 [inline]
>>  syscall_exit_to_user_mode+0x19/0x50 kernel/entry/common.c:301
>>  entry_SYSCALL_64_after_hwframe+0x44/0xae
>>
>> The buggy address belongs to the object at ffff888012eaae00
>>  which belongs to the cache io_kiocb of size 208
>> The buggy address is located 0 bytes inside of
>>  208-byte region [ffff888012eaae00, ffff888012eaaed0)
>> The buggy address belongs to the page:
>> page:0000000091458aed refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x12eaa
>> flags: 0xfff00000000200(slab)
>> raw: 00fff00000000200 ffff8880188dcd50 ffff8880188dcd50 ffff888141b4ff00
>> raw: 0000000000000000 ffff888012eaa040 000000010000000c 0000000000000000
>> page dumped because: kasan: bad access detected
>>
>> Memory state around the buggy address:
>>  ffff888012eaad00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>>  ffff888012eaad80: fb fb fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>>> ffff888012eaae00: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>>                    ^
>>  ffff888012eaae80: fb fb fb fb fb fb fb fb fb fb fc fc fc fc fc fc
>>  ffff888012eaaf00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>> ==================================================================
> 
> Fix double free by reverting 41be53e94fb0 ("io_uring: kill cached
> requests from exiting task closing the ring") - it ruins kworker's life.
> 
> --- x/fs/io_uring.c
> +++ y/fs/io_uring.c
> @@ -9234,7 +9234,6 @@ static int io_uring_flush(struct file *f
>  
>  	if (fatal_signal_pending(current) || (current->flags & PF_EXITING)) {
>  		io_uring_cancel_task_requests(ctx, NULL);
> -		io_req_caches_free(ctx, current);
>  	}
>  
>  	if (!tctx)

Already merged the right fix this morning:

https://git.kernel.dk/cgit/linux-block/commit/?h=for-5.12/io_uring&id=8e5c66c485a8af3f39a8b0358e9e09f002016d92

-- 
Jens Axboe