Date: Mon, 19 May 1997 11:14:19 +0800 From: Peter Wemm <peter@spinner.DIALix.COM> To: "Justin T. Gibbs" <gibbs@plutotech.com> Cc: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>, freebsd-bugs@hub.freebsd.org Subject: Re: kern/3618: getsockname and getpeername may cause trap 12 Message-ID: <199705190314.LAA18274@spinner.DIALix.COM> In-Reply-To: Your message of "Sun, 18 May 1997 14:48:41 CST." <199705181950.NAA19710@pluto.plutotech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
"Justin T. Gibbs" wrote:
> > <<On Sun, 18 May 1997 15:08:52 +0200 (CEST), Tor Egge <Tor.Egge@idt.ntnu.no
>
> >said:
> >
> > > During the getsockname() or getpeername() call, a network interrupt
> > > might reset the connection, causing the socket to no longer have a pcb.
> > > If this happens at the wrong time, the system gets a trap 12.
> >
> > This fix looks OK. You have commit privs...
>
> I think that the splnet() and perhaps the other initializers should be
> broken appart from the declarations since, as I recall, he compiler can
> re-order the initializations if it doesn't see any interdependancies.
> It would also make it clearer exactly which actions need splnet protection.
Also, from a fine grained SMP locking perspective, we're going to need to
get at them too.
Anyway, I'd prefer if we didn't add any more of these type of declarations
if possible since they are going to have to be revisited soon anyway for
smp locking.
> --
> Justin T. Gibbs
> ===========================================
> FreeBSD: Turning PCs into workstations
> ===========================================
Cheers,
-Peter
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199705190314.LAA18274>