RE: [PATCH v3] Drivers: hv: vmbus: fix the race when querying & updating the percpu list

[Dexuan Cui]

> From: Vitaly Kuznetsov [mailto:vkuznets@xxxxxxxxxx]
> Sent: Wednesday, June 1, 2016 0:27
> To: Dexuan Cui <decui@xxxxxxxxxxxxx>
> Cc: gregkh@xxxxxxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; driverdev-
> devel@xxxxxxxxxxxxxxxxxxxxxx; olaf@xxxxxxxxx; apw@xxxxxxxxxxxxx;
> jasowang@xxxxxxxxxx; KY Srinivasan <kys@xxxxxxxxxxxxx>; Haiyang
> Zhang <haiyangz@xxxxxxxxxxxxx>; rolf.neugebauer@xxxxxxxxxx;
> dave.scott@xxxxxxxxxx; ian.campbell@xxxxxxxxxx
> Subject: Re: [PATCH v3] Drivers: hv: vmbus: fix the race when querying &
> updating the percpu list
> 
> Dexuan Cui <decui@xxxxxxxxxxxxx> writes:
> 
> > There is a rare race when we remove an entry from the global list
> > hv_context.percpu_list[cpu] in hv_process_channel_removal() ->
> > percpu_channel_deq() -> list_del(): at this time, if vmbus_on_event() ->
> > process_chn_event() -> pcpu_relid2channel() is trying to query the list,
> > we can get the kernel fault.
> >
> > Similarly, we also have the issue in the code path: vmbus_process_offer()
> ->
> > percpu_channel_enq().
> >
> > We can resolve the issue by disabling the tasklet when updating the list.
> >
> > The patch also moves vmbus_release_relid() to a later place where
> > the channel has been removed from the per-cpu and the global lists.
> >
> > Reported-by: Rolf Neugebauer <rolf.neugebauer@xxxxxxxxxx>
> > Cc: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
> > Signed-off-by: Dexuan Cui <decui@xxxxxxxxxxxxx>
> 
> Tested 4.7-rc1 with this path applied and kernel always crashes on boot
> (WS2016TP5, 12 CPU SMP guest, Generation 2):
> 
> [    5.464251] hv_vmbus: Hyper-V Host Build:14300-10.0-1-0.1006; Vmbus
> version:4.0
> [    5.471666] hv_vmbus: Unknown GUID: f8e65716-3cb3-4a06-9a60-
> 1889c5cccab5
> [    5.472143] BUG: unable to handle kernel paging request at
> 000000079fff5288
> [    5.477107] IP: [<ffffffffa0004b91>] vmbus_onoffer+0x311/0x570
> [hv_vmbus]
> ...
>   Vitaly

I can't reproduce the panic somehow, but I did find a bug in vmbus_process_offer():

"hv_event_tasklet_disable(channel) and hv_event_tasklet_enable(channel)"
are buggy: the 'channel' parameter should be 'newchannel'.

This was a copy-and-paste bug... Sorry!
Can you fix this and see if the panic will disappear in your side?

Thanks,
-- Dexuan