Re: [PATCH 2/2] usb: host: xhci: Handle the right timeout command

[Mathias Nyman]

On 27.12.2016 05:07, Baolin Wang wrote:
> Hi,
>
> On 21 December 2016 at 21:00, Mathias Nyman
> <mathias.nyman@xxxxxxxxxxxxxxx> wrote:
> > On 21.12.2016 04:22, Baolin Wang wrote:
> > > Hi Mathias,
> > >
> > > On 20 December 2016 at 23:13, Mathias Nyman
> > > <mathias.nyman@xxxxxxxxxxxxxxx> wrote:
> > > > On 20.12.2016 09:30, Baolin Wang wrote:
> > > > ...
> > > >
> > > > Alright, I gathered all current work related to xhci races and timeouts
> > > > and put them into a branch:
> > > >
> > > > git://git.kernel.org/pub/scm/linux/kernel/git/mnyman/xhci.git
> > > > timeout_race_fixes
> > > >
> > > > Its based on 4.9
> > > > It includes a few other patches just to avoid conflicts and  make my life
> > > > easier
> > > >
> > > > Interesting patches are:
> > > >
> > > > ee4eb91 xhci: remove unnecessary check for pending timer
> > > > 0cba67d xhci: detect stop endpoint race using pending timer instead of
> > > > counter.
> > > > 4f2535f xhci: Handle command completion and timeout race
> > > > b9d00d7 usb: host: xhci: Fix possible wild pointer when handling abort
> > > > command
> > > > 529a5a0 usb: xhci: fix possible wild pointer
> > > > 4766555 xhci: Fix race related to abort operation
> > > > de834a3 xhci: Use delayed_work instead of timer for command timeout
> > > > 69973b8 Linux 4.9
> > > >
> > > > The fixes for command queue races will go to usb-linus and stable, the
> > > > reworks for stop ep watchdog timer will go to usb-next.
> > >
> > >
> > > How about applying below patch in your 'timeout_race_fixes' branch?
> > > [PATCH] usb: host: xhci: Clean up commands when stop endpoint command is
> > > timeout
> > > https://lkml.org/lkml/2016/12/13/94
> >
> > usb_hc_died() should eventyally end up calling xhci_mem_cleanup()
> > which will cleanup the command queue. But I need to look into that
>
> usb_hc_died() did not call xhci_mem_cleanup() to clean up command
> queue, it just disconnects all children devices attached on the dying
> hub. I did not find where it will clean up the command queue when
> issuing usb_hc_died(). Also before issuing usb_hc_died() in
> xhci_handle_command_timeout(), we will call
> xhci_cleanup_command_queue(). I think it should same as in
> xhci_stop_endpoint_command_watchdog().

You're right, it doesn't call xhci_mem_cleanup.
Need to look at this after getting first series of fixes to usb-linus

-Mathias