This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: conditional probe semantics
- From: "Frank Ch. Eigler" <fche at redhat dot com>
- To: Dave Nomura <dcnltc at us dot ibm dot com>
- Cc: systemtap at sourceware dot org
- Date: Fri, 13 Jun 2008 21:28:12 -0400
- Subject: Re: conditional probe semantics
- References: <485317A2.3030201@us.ibm.com>
Hi -
On Fri, Jun 13, 2008 at 05:58:10PM -0700, Dave Nomura wrote:
> probe process(...)..... if (condition) { ... }
> and
> probe process(...).... { if (condition) { ... } }
The former is considered a necessary but perhaps not sufficient
condition for the probe to potentially fire.
It hints that "condition" will typically be false, so that actually
removing the underlying e.g. kprobe should be considered by the
runtime. Once the condition returns to true, the e.g. kprobe should
be reinserted as soon as possible, though some events may be missed in
the interim.
The above is in any case how it will probably work. Right now, they
are functionally equivalent, except that the necessary semantic
restrictions of the condition expression are already enforced.
- FChE