This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: BUG [preempt-rt] stapio:4521 task might have lost a preemption check!
- From: fche at redhat dot com (Frank Ch. Eigler)
- To: Kiran <kiran at linux dot vnet dot ibm dot com>
- Cc: systemtap at sources dot redhat dot com, sripathik at in dot ibm dot com
- Date: Tue, 16 Jun 2009 12:17:44 -0400
- Subject: Re: BUG [preempt-rt] stapio:4521 task might have lost a preemption check!
- References: <1245149554.14651.36.camel@kiran-laptop>
Kiran <kiran@linux.vnet.ibm.com> writes:
> BUG: stapio:4521 task might have lost a preemption check!
> Call Trace:
> [c0000000e0ba3930] [c0000000000122e8] .show_stack+0x94/0x198 (unreliable)
> [c0000000e0ba39e0] [c000000000012414] .dump_stack+0x28/0x3c
> [c0000000e0ba3a60] [c0000000000a688c] .preempt_enable_no_resched+0x68/0x84
> [c0000000e0ba3ae0] [d0000000017368d0] .enter_be_probe+0x278/0x29c [stap_828d9711d4343f96e62db9359ace031f_8344]
> [...]
>From just looking at the rt code, this error appears to occur if we
call preempt_enable_no_resched() without a corresponding
preempt_disable(). This should not arise for a begin/end probe that
the stack trace suggests your script had.
You could check out the output of "stap -p3 YOURSCRIPT.stp" to see
whether the preempt calls are matching up properly. Or you could send
us the script and perhaps the vitals suggested in the wiki/HowToReportBugs.
- FChE