From mboxrd@z Thu Jan 1 00:00:00 1970 From: Joel Soete Subject: Re: [parisc-linux] Latest palinux crash Date: Thu, 21 Jul 2005 19:48:23 +0000 Message-ID: <42DFFC07.5080607@tiscali.be> References: <20050720122227.GK10156@parcelfarce.linux.theplanet.co.uk> <20050721043626.GY30182@systemhalted.org> <42DF883E.1020607@tiscali.be> <20050721124713.GB27330@colo.lackof.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Cc: Carlos O'Donell , Matthew Wilcox , parisc-linux@parisc-linux.org To: Grant Grundler Return-Path: In-Reply-To: <20050721124713.GB27330@colo.lackof.org> List-Id: parisc-linux developers list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: parisc-linux-bounces@lists.parisc-linux.org Grant Grundler wrote: > On Thu, Jul 21, 2005 at 11:34:22AM +0000, Joel Soete wrote: > >>This kernel seems to be very more stable (passed with success some stress >>test of mine during severall days when it would crash in few ours or less >>when compile with gcc-3.3). > > > Joel, > This is good news. Thanks for trying this out. > > >>... >>but pressing TOC button launched well a panic, and unfortunately pim >>analisys didn't help at all :-? >>(fyi: >>) > > > In 026855.html you wrote you didn't save the System.map. > Yes my script analysis is based on the famous dump_analyser.sh so still grab the label into System.map but insn match well the address ;-) (today I learn more about addr2line and thought that it would help to get rid of this pb :-) > >>So am I also looking for way to help: lkcd? instrumenting the kernel like >>kprobes/dprobes? (but not yet hppa support) > > > kprobes will need more support for runtime patching of the code. > It's not trivial to do that given the VIVT caches on parisc. (Sorry: VIVT?) > It's certainly possible though since I know HPUX does that. > But it also means changing the kernel text to RW or using > absolute stores/icache flushing to modified the kernel instructions. > well so not sure that kprobe was already for p-l but I find this paper on gcc fnct instrumentation very attractive and looking for something like this for the kernel: in paper the link to kprobes is dead but resent me to dprobes on sf.net? > > >>Unfortunately, I definetily don't have deep enough knowledge of linux >>kernel to help more, sorry. > > > I only know enough to be dangerous to be people who know less. :^) > > grant > Thanks, Joel _______________________________________________ parisc-linux mailing list parisc-linux@lists.parisc-linux.org http://lists.parisc-linux.org/mailman/listinfo/parisc-linux