On Wed, Jan 06, 2021 at 11:45:50AM -0500, Alan Stern wrote:

On Wed, Jan 06, 2021 at 01:10:05PM +0300, Dan Carpenter wrote:

quoted

Syzbot discovered that the probe error handling doesn't clean up the
resources allocated in zr364xx_board_init().  There are several
related bugs in this code so I have re-written the error handling.

Dan:

I recently sent in a patch for a similar problem in the gspca driver
(commit e469d0b09a19 "media: gspca: Fix memory leak in probe").  It
seems there may be similar issues in that driver: one single function
call tries to undo an indeterminate number of initializations.

I don't know enough about these subsystems to evaluate this.  Can you
take a look at it?

The probe error handling in gspca_dev_probe2() is fine now.  All those
functions are no-ops when they haven't been allocated/registered.

regards,
dan carpenter