+1 this seems OK to me. From: vpp-dev@lists.fd.io <vpp-dev@lists.fd.io> On Behalf Of Damjan Marion via Lists.Fd.Io Sent: Thursday, March 12, 2020 6:16 AM To: Lijian Zhang <lijian.zh...@arm.com> Cc: vpp-dev@lists.fd.io Subject: Re: [vpp-dev] Ignore SIGPROF signal in VPP
> On 12 Mar 2020, at 10:49, Lijian Zhang > <lijian.zh...@arm.com<mailto:lijian.zh...@arm.com>> wrote: > > Hi Maintainers, > We are profiling VPP with MAP (a software profile suite on Arm CPUs, see > details > inhttps://www.arm.com/products/development-tools/server-and-hpc/forge/map) on > Arm CPUs. > > The MAP sampler runs inside target process as a library because it does a lot > more things that require access to the program's memory like matching up > OpenMP threads etc. So it is a lot more invasive and uses the SIGPROF signal > to control the sample rate. > > VPP will receive SIGPROF signal because MAP uses SIGPROF signal to drive its > sampler to do profiling on VPP. However, the default action of SIGPROF signal > handler in VPP such as unix_signal_handler() is process termination. To > profile VPP with MAP, We need to change VPP signal handler to ignore SIGPROF > signal. > > Can we upstream a patch to simply ignore the SIGPROF in VPP? I think so, please submit patch and unless somebody raises his concern here I will merge it... > > diff --git a/src/vlib/unix/main.c b/src/vlib/unix/main.c > index e40a462..6138a6f 100755 > --- a/src/vlib/unix/main.c > +++ b/src/vlib/unix/main.c > @@ -218,6 +218,7 @@ > /* ignore SIGPIPE, SIGCHLD */ > case SIGPIPE: > case SIGCHLD: > + case SIGPROF: > sa.sa_sigaction = (void *) SIG_IGN; > break; — Damjan
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#15749): https://lists.fd.io/g/vpp-dev/message/15749 Mute This Topic: https://lists.fd.io/mt/71898718/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
