Re: [PATCH v2] Input: uinput: Avoid Object-Already-Free with a global lock

[Mukesh Ojha]

On 4/24/2019 6:37 PM, Al Viro wrote:
> On Wed, Apr 24, 2019 at 05:40:40PM +0530, Mukesh Ojha wrote:
> > Al,
> >
> > i tried to put traceprintk inside ioctl after fdget and fdput on a simple
> > call of open => ioctl => close
> in a loop, and multithreaded, presumably?
>
> > on /dev/uinput.
> >
> >  ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [002] ....ÂÂÂ 45.312044: SYSC_ioctl: 2 Â Â <= f_count
> > >  ÂÂÂ <After fdget()
> >  ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [002] ....ÂÂÂ 45.312055: SYSC_ioctl: 2
> > <After fdput()
> >  ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313766: uinput_open: uinput: 1
> >  ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313783: SYSC_ioctl: 1
> >  ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313788: uinput_ioctl_handler:
> > uinput: uinput_ioctl_handler, 1
> >  ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313835: SYSC_ioctl: 1
> >  ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313843: uinput_release: uinput:Â 0
> >
> >
> > So while a ioctl is running the f_count is 1, so a fput could be run and do
> > atomic_long_dec_and_test
> > this could call release right ?
> Look at ksys_ioctl():
> int ksys_ioctl(unsigned int fd, unsigned int cmd, unsigned long arg)
> {
>          int error;
>          struct fd f = fdget(fd);
> an error or refcount bumped
>          if (!f.file)
>                  return -EBADF;
> not an error, then.  We know that ->release() won't be called
> until we drop the reference we've just acquired.
>          error = security_file_ioctl(f.file, cmd, arg);
>          if (!error)
>                  error = do_vfs_ioctl(f.file, fd, cmd, arg);
> ... and we are done with calling ->ioctl(), so
>          fdput(f);
> ... we drop the reference we'd acquired.
>
> Seeing refcount 1 inside ->ioctl() is possible, all right:
>
> CPU1: ioctl(2) resolves fd to struct file *, refcount 2
> CPU2: close(2) rips struct file * from descriptor table and does fput() to drop it;
> 	refcount reaches 1 and fput() is done; no call of ->release() yet.
> CPU1: we get arouund to ->ioctl(), where your trace sees refcount 1
> CPU1: done with ->ioctl(), drop our reference.  *NOW* refcount gets to 0, and
> 	->release() is called.

Thanks for the detail reply, Al

This was my simple program no multithreading just to understand f_counting

int main()
{
ÂÂÂÂÂÂÂ int fd = open("/dev/uinput", O_WRONLY | O_NONBLOCK);
ÂÂÂÂÂÂÂ ioctl(fd, UI_SET_EVBIT, EV_KEY);
ÂÂÂÂÂÂÂ close(fd);
ÂÂÂÂÂÂÂ return 0;
}

   Â  uinput-532 [002] .... 45.312044: SYSC_ioctl: 2  <= 
f_count >ÂÂÂ <After fdget()
ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [002] ....ÂÂÂ 45.312055: SYSC_ioctl: 
2ÂÂÂÂÂÂÂÂÂÂÂ <After fdput()
ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313766: uinput_open: uinput: 
1ÂÂ /* This is from the uinput driver uinput_open()*/

 =>>>>             /* All the above calls happened for the 
open() in userspace*/

ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313783: SYSC_ioctl: 1 /* This 
print is for the trace, i put after fdget */
ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313788: uinput_ioctl_handler: 
uinput: uinput_ioctl_handler, 1 /* This print is from the uinput_ioctl 
driver */

ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313835: SYSC_ioctl: 1 /* This 
print is for the trace, i put after fdput*/
ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [004] ....ÂÂÂ 45.313843: uinput_release: 
uinput:Â 0 /* And this is from the close()Â */


Should fdget not suppose to increment the f_count here, as it is coming 1 ?
This f_count to one is done at the open, but i have no idea how this 
below f_count 2 came before open() for
this simple program.

  Â  uinput-532 [002] .... 45.312044: SYSC_ioctl: 2 <= f_count 
>ÂÂÂ <After fdget()
ÂÂÂÂÂÂÂÂÂ uinput-532ÂÂ [002] ....ÂÂÂ 45.312055: SYSC_ioctl: 
2ÂÂÂÂÂÂÂÂÂÂÂ <After fdput()

-Mukesh

> IOW, in your trace fput() has already been run by close(2); having somebody else
> do that again while we are in ->ioctl() would be a bug (to start with, where
> did they get that struct file * and why wasn't that reference contributing to
> struct file refcount?)
>
> In all cases we only call ->release() once all references gone - both
> the one(s) in descriptor tables and any transient ones acquired by
> fdget(), etc.
>
> I would really like to see a reproducer for the original use-after-free report...