Date: Fri, 2 May 2003 09:46:20 +0200 From: Vincent Jardin <vjardin@wanadoo.fr> To: Mike Silbersack <silby@silby.com>, freebsd-net@freebsd.org Subject: Re: More mbuf INVARIANTS code, comments needed Message-ID: <200305020946.20514.vjardin@wanadoo.fr> In-Reply-To: <20030502010545.U610@odysseus.silby.com> References: <20030502010545.U610@odysseus.silby.com>
next in thread | previous in thread | raw e-mail | index | archive | help
It is a good idea. I do not see any problems with your patch. An esthetic comment: I would prefer to see other trivial hexadecimal values like: - 0xd0 (as in "Duh", used by stdlib/malloc()) - or 0xdeadc0de (used by kern_malloc.c:#define WEIRD_ADDR 0xdeadc0de) - or 0xdead0137, 0xdead0138, 0xdead0139, ... According to me, these values are easier to analyse when you get a panic = or=20 when you dump the memory. Regards, Vincent On Friday 02 May 2003 08:08, Mike Silbersack wrote: > Now that I have the double-free code in (under INVARIANTS), I'm > considering the attached patch as well; it fills the m_data, m_next, an= d > m_nextpkt fields with non-NULL garbage in hopes that any uses after fre= e > will be immediately fatal. > > Does anyone see problems with this, and/or other simple checks that cou= ld > be added cheaply? > > Thanks, > > Mike "Silby" Silbersack
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200305020946.20514.vjardin>