Date: Wed, 16 Mar 2005 18:45:09 +0000 From: Kris Kennaway <kris@FreeBSD.org> To: Maxime Henrion <mux@FreeBSD.org> Cc: cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/dev/fxp if_fxp.c Message-ID: <20050316184509.GD91771@hub.freebsd.org> In-Reply-To: <20050316184056.GP31320@elvis.mu.org> References: <200503161639.j2GGd4g1083863@repoman.freebsd.org> <20050316181535.GA91771@hub.freebsd.org> <20050316184056.GP31320@elvis.mu.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Mar 16, 2005 at 07:40:56PM +0100, Maxime Henrion wrote: > Kris Kennaway wrote: > > On Wed, Mar 16, 2005 at 04:39:04PM +0000, Maxime Henrion wrote: > > > mux 2005-03-16 16:39:04 UTC > > > > > > FreeBSD src repository > > > > > > Modified files: > > > sys/dev/fxp if_fxp.c > > > Log: > > > Unload and destroy the TX DMA maps before destroying the DMA tag > > > they're attached to, not after. > > > > > > Spotted by: Coverity via sam > > > > Is this likely to have caused panics at runtime? A few people > > (including me) have seen dma-related or use-after-free panics from > > fxp. > > This won't fix any use-after-free panics at runtime, since the faulty > code is in fxp_release() which will only get called from fxp_attach() in > error scenarios and from fxp_detach(). > > I've never seen reports about use-after-free from fxp(4), I'd like to > know about those, even if the odds of those not being false positives is > small... I think I mailed current about it a few months ago (but can't access my email archive until the weekend), and a few other people have reported similar-looking panics too. I've only seen it 2-3 times in total. Kris -- In God we Trust -- all others must submit an X.509 certificate. -- Charles Forsythe <forsythe@alum.mit.edu>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050316184509.GD91771>