This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
RE: Test result of systemtap-20060107
- From: "Mao, Bibo" <bibo dot mao at intel dot com>
- To: <systemtap at sources dot redhat dot com>
- Date: Mon, 9 Jan 2006 16:44:06 +0800
- Subject: RE: Test result of systemtap-20060107
I think that it is because maxactive value is 0 when registering kretprobe, it is to say that without CONFIG_PREEMPT option, maxactive will be equal to NR_CPUS. And schedule() function is something like recursive, sometimes maybe schedule(A->B), (B->C), ....(X->A), so when schedule function return, this function may be called for many times, and its free_rp_inst is limited to maxactive value, sometimes kretprobe will be missed.
Regards
Bibo,mao
>-----Original Message-----
>From: systemtap-owner@sourceware.org [mailto:systemtap-owner@sourceware.org]
>On Behalf Of Mao, Bibo
>Sent: 2006年1月9日 15:46
>To: systemtap@sources.redhat.com
>Subject: Test result of systemtap-20060107
>
> The recently snapshot can run well in my IA64 box by make check
>command. But when I run some test cases, there are information abnormal.
>About running script buildok/five.stp in SMP enviroment, times of
>"arrival" information is larger than that of "departure" log
>information, and after long time, there is occasionally "departure"
>information.
> And when this script exits, there will be such information like
>this:
> WARNING: Number of errors: 0, skipped probes: 36211
>In my EM64T smp machine, there is also this phenomenon.
>
>Regards
>Bibo,mao