[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <y0m63722gmf.fsf@fche.csb>
Date: Sat, 16 Jan 2010 10:50:48 -0500
From: fche@...hat.com (Frank Ch. Eigler)
To: Jim Keniston <jkenisto@...ibm.com>
Cc: Srikar Dronamraju <srikar@...ux.vnet.ibm.com>,
Peter Zijlstra <peterz@...radead.org>,
Arnaldo Carvalho de Melo <acme@...radead.org>,
Frederic Weisbecker <fweisbec@...il.com>,
LKML <linux-kernel@...r.kernel.org>,
Mark Wielaard <mjw@...hat.com>,
utrace-devel <utrace-devel@...hat.com>
Subject: Re: [RFC] [PATCH 4/7] Uprobes Implementation
Jim Keniston <jkenisto@...ibm.com> writes:
> [...]
> Years ago, we had pre-utrace versions of uprobes where the uprobes
> breakpoint-handler code was dispatched from the die_notifier, before the
> int3 turned into a SIGTRAP. I believe that's what Peter is
> recommending. On my old Pentium M...
> - a pre-utrace uprobe hit cost about 1 usec;
> - a utrace-based uprobe hit cost about 3 usec;
> [...]
> So yeah, learning about the int3 via utrace after the SIGTRAP gets
> created adds some overhead to uprobes. [...]
Was this test comparing likewise fruit? For example, did it account
for factors where other processes were gdb-int3-instrumented or with
lots of kprobes active? Differently multithreaded? Demultiplexing
probes amongst multiple processes?
(It's counterintuitive that the utrace/kernel int3->sigtrap
dispatching code alone should cause thousands of extra instructions.)
- FChE
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists