[ovs-dev] [PATCHv2] connmgr: Check nullptr inside ofmonitor_report()

[Ilya Maximets]

On 2/19/21 1:09 AM, William Tu wrote:
> On Wed, Feb 17, 2021 at 1:09 PM Yifeng Sun <pkusunyifeng at gmail.com> wrote:
>>
>> ovs-vswitchd could crash under these circumstances:
>> 1. When one bridge is being destroyed, ofproto_destroy() is called and
>> connmgr pointer of its ofproto struct is nullified. This ofproto struct is
>> deallocated through 'ovsrcu_postpone(ofproto_destroy_defer__, p);'.
>> 2. Before RCU enters quiesce state to actually free this ofproto struct,
>> revalidator thread calls udpif_revalidator(), which could handle
>> a learn flow and calls ofproto_flow_mod_learn(), it later calls
>> ofmonitor_report() and ofproto struct's connmgr pointer is accessed.
>>
>> The crash stack trace is shown below:
>>
>> 0  ofmonitor_report (mgr=0x0, rule=rule at entry=0x7fa4ac067c30, event=event at entry=NXFME_ADDED,
>>     reason=reason at entry=OFPRR_IDLE_TIMEOUT, abbrev_ofconn=0x0, abbrev_xid=0, old_actions=old_actions at entry=0x0)
>>     at ofproto/connmgr.c:2160
>> 1  0x00007fa4d6803495 in add_flow_finish (ofproto=0x55d9075d4ab0, ofm=<optimized out>, req=req at entry=0x0)
>>     at ofproto/ofproto.c:5221
>> 2  0x00007fa4d68036af in modify_flows_finish (req=0x0, ofm=0x7fa4980753f0, ofproto=0x55d9075d4ab0)
>>     at ofproto/ofproto.c:5823
>> 3  ofproto_flow_mod_finish (ofproto=0x55d9075d4ab0, ofm=ofm at entry=0x7fa4980753f0, req=req at entry=0x0)
>>     at ofproto/ofproto.c:8088
>> 4  0x00007fa4d680372d in ofproto_flow_mod_learn_finish (ofm=ofm at entry=0x7fa4980753f0,
>>     orig_ofproto=orig_ofproto at entry=0x0) at ofproto/ofproto.c:5439
>> 5  0x00007fa4d68072f9 in ofproto_flow_mod_learn (ofm=0x7fa4980753f0, keep_ref=keep_ref at entry=true,
>>     limit=<optimized out>, below_limitp=below_limitp at entry=0x0) at ofproto/ofproto.c:5499
>> 6  0x00007fa4d6835d33 in xlate_push_stats_entry (entry=0x7fa498012448, stats=stats at entry=0x7fa4d2701a10,
>>     offloaded=offloaded at entry=false) at ofproto/ofproto-dpif-xlate-cache.c:127
>> 7  0x00007fa4d6835e3a in xlate_push_stats (xcache=<optimized out>, stats=stats at entry=0x7fa4d2701a10,
>>     offloaded=offloaded at entry=false) at ofproto/ofproto-dpif-xlate-cache.c:181
>> 8  0x00007fa4d6822046 in revalidate_ukey (udpif=udpif at entry=0x55d90760b240, ukey=ukey at entry=0x7fa4b0191660,
>>     stats=stats at entry=0x7fa4d2705118, odp_actions=odp_actions at entry=0x7fa4d2701b50,
>>     reval_seq=reval_seq at entry=5655486242, recircs=recircs at entry=0x7fa4d2701b40, offloaded=false)
>>     at ofproto/ofproto-dpif-upcall.c:2294
>> 9  0x00007fa4d6825aee in revalidate (revalidator=0x55d90769dd00) at ofproto/ofproto-dpif-upcall.c:2683
>> 10 0x00007fa4d6825cf3 in udpif_revalidator (arg=0x55d90769dd00) at ofproto/ofproto-dpif-upcall.c:936
>> 11 0x00007fa4d6259c9f in ovsthread_wrapper (aux_=<optimized out>) at lib/ovs-thread.c:423
>> 12 0x00007fa4d582cea5 in start_thread () from /usr/lib64/libpthread.so.0
>> 13 0x00007fa4d504b96d in clone () from /usr/lib64/libc.so.6
>>
>> At the time of crash, the involved ofproto was already deallocated:
>>
>> (gdb) print *ofproto
>> $1 = ..., name = 0x55d907602820 "nsx-managed", ..., ports = {...,
>>     one = 0x0, mask = 63, n = 0}, ..., connmgr = 0x0, ...
>>
>> This patch fixes it.
>>
>> VMware-BZ: #2700626
>> Signed-off-by: Yifeng Sun <pkusunyifeng at gmail.com>
>> ---
>> v1->v2: Add check for ofmonitor_flush, thanks William.
>>
> 
> LGTM, thanks.
> Acked-by: William Tu < u9012063 at gmail.com>

Thanks!
Applied to master and backported down to 2.12.

Best regards, Ilya Maximets.