Re: WARNING in request_end

[Kirill Tkhai]

On 24.09.2018 17:44, Miklos Szeredi wrote:
> On Mon, Sep 24, 2018 at 2:29 PM, syzbot
> <syzbot+ef054c4d3f64cd7f7cec@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
>> Hello,
>>
>> syzbot found the following crash on:
>>
>> HEAD commit:    6bf4ca7fbc85 Linux 4.19-rc5
>> git tree:       upstream
>> console output: https://syzkaller.appspot.com/x/log.txt?x=159149c6400000
>> kernel config:  https://syzkaller.appspot.com/x/.config?x=22a62640793a83c9
>> dashboard link: https://syzkaller.appspot.com/bug?extid=ef054c4d3f64cd7f7cec
>> compiler:       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+ef054c4d3f64cd7f7cec@xxxxxxxxxxxxxxxxxxxxxxxxx
>>
>> WARNING: CPU: 0 PID: 9445 at fs/fuse/dev.c:390 request_end+0x82e/0xaa0
> 
> And  there we have the bug likely caused by the set_bit(FR_SENT, ...)
> not being inside the fpq->lock-ed region.
> 
> So that needs to be fixed anyway, apparently.

I can't confirm, since I haven't found yet the direct way, that set_bit() results
in this stack...

We have one more (unrelated) possible use-after-free here:

cpu0                                       cpu1
fuse_dev_do_write()                     fuse_dev_do_write()
  req = request_find(fpq, oh.unique)    ...
  spin_unlock(&fpq->lock)               ...
  ...                                   req = request_find(fpq, oh.unique)
  ...                                   spin_unlock(&fpq->lock)
  queue_interrupt(&fc->iq, req);        ...
  ...                                   ...
  ...                                   ...
request freed                           ...
...                                     queue_interrupt(&fc->iq, req); <- use after free

Something like below is needed:

@@ -1875,16 +1877,20 @@ static ssize_t fuse_dev_do_write(struct fuse_dev *fud,
 
 	/* Is it an interrupt reply? */
 	if (req->intr_unique == oh.unique) {
+		__fuse_get_request(req);
 		spin_unlock(&fpq->lock);
 
 		err = -EINVAL;
-		if (nbytes != sizeof(struct fuse_out_header))
+		if (nbytes != sizeof(struct fuse_out_header)) {
+			fuse_put_request(fc, req);
 			goto err_finish;
+		}
 
 		if (oh.error == -ENOSYS)
 			fc->no_interrupt = 1;
 		else if (oh.error == -EAGAIN)
 			queue_interrupt(&fc->iq, req);
+		fuse_put_request(fc, req);
 
 		fuse_copy_finish(cs);
 		return nbytes;