This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: Likely problem in stap_compile.exp
- From: Mark Wielaard <mjw at redhat dot com>
- To: wenji dot huang at oracle dot com
- Cc: SystemTAP <systemtap at sources dot redhat dot com>
- Date: Tue, 01 Sep 2009 11:40:29 +0200
- Subject: Re: Likely problem in stap_compile.exp
- References: <4A9CE794.6060101@oracle.com>
Hi Wenji,
On Tue, 2009-09-01 at 17:21 +0800, Wenji Huang wrote:
> Seems there's problem in testsuite/lib/stap_compile.exp when
> matching the output with expected results since commit
> a4433a9da1f21c6536266531308ba231e5efae81.
>
> This is one piece of systemtap.base/marker.exp test log.
>
> running stap -v -p4 -e "probe kernel.mark(\"kvm_trace_entryexit\") { }"
> Pass 1: parsed user script and 59 library script(s) in
> 630usr/110sys/738real ms.^M
> WARNING: side-effect-free probe 'probe_1805': keyword at <input>:1:1^M
> source: probe kernel.mark("kvm_trace_entryexit") { }^M
> ^^M
> Pass 2: analyzed script: 1 probe(s), 0 function(s), 1 embed(s), 0
> global(s) in 20usr/0sys/33real ms.^M
> Pass 3: translated to C into
> "/tmp/stapxxfUqx/stap_98582564a144af32100025c9d1396921_789.c" in
> 0usr/0sys/0real ms.^M
> FAIL: K_MARKER02_compile failed, but got no error messages
> PASS: K_MARKER02 compilation succeeded
>
> It's a little odd to say FAIL and PASS together.
They are slightly different tests, the first (with the _compile
addition) tests whether the exit code of stap, plus the error messages
seen correspond to the exit code. The second (without the _compile
postfix) only looks at the error messages themselves to see if they are
expected. But I guess we could just "return" after the first fail.
It looks like in your case either the testcase didn't match the full
output somehow, or stap crashed after Pass 3 (both since the Pass 4
output is missing). Could you run the stap command by hand and check the
exit code? (echo $?)
> In fact, it works fine on FC11.
On what setup do you see it fail?
Thanks,
Mark