Date: Mon, 4 Jun 2001 08:35:53 +0300 (EEST) From: Maxim Sobolev <sobomax@mail-in.net> To: eischen@vigrid.com, bde@zeta.org.au Cc: current@FreeBSD.ORG, deischen@FreeBSD.ORG Subject: Re: sscanf(3) is broken in 5-CURRENT [SIGBUS] Message-ID: <200106040535.f545Zoo56413@mail.uic-in.net> In-Reply-To: <Pine.SUN.3.91.1010603182456.27210A-100000@pcnet1.pcnet.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 3 Jun 2001 18:29:52 -0400 (EDT), Daniel Eischen wrote: > On Mon, 4 Jun 2001, Bruce Evans wrote: > > On Sat, 2 Jun 2001, Maxim Sobolev wrote: > > > > > It seems that something is wrong with sscanf(3) in -current - in > > > some cases it may cause SIGBUS. I failed to reproduce the > > > problem on 4-STABLE, so it is a -current specific bug. Attached > > > please find small showcase that exposes the bug in question > > > and a backtrace after SIGBUS. > [ ... ] > > > > This is because fp->_extra is not initialized by sscanf() (it is stack > > garbage that happened to be 0 when I looked at it). > > Yes, it looks like the change from _up to _extra (to hold _up and other > locking stuff) screwed this up. Here's a fix. My source is a month > out of date (DEVFS won't let my old X applications work, so I need > to upgrade my systems before it becomes mandatory), so I don't know > if it'll apply cleanly to -current sources. Yes, this solved the problem. Please commit it. -Maxim To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200106040535.f545Zoo56413>