This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: [PATCH -tip v5 0/7] tracing: kprobe-based event tracer and x86 instruction decoder
- From: Ingo Molnar <mingo at elte dot hu>
- To: Masami Hiramatsu <mhiramat at redhat dot com>
- Cc: Steven Rostedt <rostedt at goodmis dot org>, lkml <linux-kernel at vger dot kernel dot org>, Avi Kivity <avi at redhat dot com>, "H. Peter Anvin" <hpa at zytor dot com>, Frederic Weisbecker <fweisbec at gmail dot com>, Ananth N Mavinakayanahalli <ananth at in dot ibm dot com>, Andrew Morton <akpm at linux-foundation dot org>, Andi Kleen <andi at firstfloor dot org>, Jim Keniston <jkenisto at us dot ibm dot com>, "K.Prasad" <prasad at linux dot vnet dot ibm dot com>, KOSAKI Motohiro <kosaki dot motohiro at jp dot fujitsu dot com>, systemtap <systemtap at sources dot redhat dot com>, kvm <kvm at vger dot kernel dot org>, Tom Zanussi <zanussi at comcast dot net>
- Date: Mon, 11 May 2009 23:47:12 +0200
- Subject: Re: [PATCH -tip v5 0/7] tracing: kprobe-based event tracer and x86 instruction decoder
- References: <20090509004829.5505.38720.stgit@localhost.localdomain> <20090509044302.GH8007@elte.hu> <4A0838E6.2000309@redhat.com> <alpine.DEB.2.00.0905111054490.2492@gandalf.stny.rr.com> <4A088523.6080003@redhat.com>
* Masami Hiramatsu <mhiramat@redhat.com> wrote:
> Steven Rostedt wrote:
> > On Mon, 11 May 2009, Masami Hiramatsu wrote:
> >>> Two high-level comments:
> >>>
> >>> - There's no self-test - would it be possible to add one? See
> >>> trace_selftest* in kernel/trace/
> >> I'm not so sure. Currently, it seems that those self-tests are
> >> only for tracers which define new event-entry on ring-buffer.
> >> Since this tracer just use ftrace_bprintk, it might need
> >> another kind of selftest. e.g. comparing outputs with
> >> expected patterns.
> >> In that case, would it be better to make a user-space self test
> >> including filters and tracepoints?
> >
> > Or have the workings in the selftest in kernel. As if a user started it.
> > It does not need to write to the ring buffer, that is just what I did. The
> > event selftests don't check if anything was written to the ring buffer,
> > they just make sure that the tests don't crash the system.
>
> Would you mean that it is enough to enable some probes and just
> see what happened at boot time?
> That's so easy to add.
Yes, that's the idea!
Try to think of regressions/crashes/misbehavior you generally
trigger while you developed kprobes, and try to add a reasonable set
of probes that test the code from those angles.
It doesnt have to be a full, complex test-suite, but even just 80%
of coverage of functionality keeps 4/5th of all regressions out of
the kernel at a very early stage ...
Ingo