This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Avoid undefined behavior in __FD_MASK
- From: OndÅej BÃlka <neleai at seznam dot cz>
- To: Florian Weimer <fweimer at redhat dot com>
- Cc: Roland McGrath <roland at hack dot frob dot com>, libc-alpha <libc-alpha at sourceware dot org>
- Date: Tue, 11 Jun 2013 18:40:40 +0200
- Subject: Re: [PATCH] Avoid undefined behavior in __FD_MASK
- References: <51651881 dot 2020906 at redhat dot com> <20130411203232 dot 72CA82C092 at topped-with-meat dot com> <5167C5ED dot 4010902 at redhat dot com>
On Fri, Apr 12, 2013 at 10:29:33AM +0200, Florian Weimer wrote:
> On 04/11/2013 10:32 PM, Roland McGrath wrote:
> >>__fd_mask
> >>should really be an unsigned type, but changing that is more risky.
> >
> >That really is the right fix (and then your change is unnecessary).
> >What risks do you imagine with the type change?
>
> The fd_bits member used to be part of the public API in the 90s, and
> someone might point a long * variable to it. There's also a
> comment:
>
> /* The fd_set member is required to be an array of longs. */
> typedef long int __fd_mask;
>
> If someone still has the discussion to Austin group issue
> bwg2001-005 (which led to the removal of the fd_bits member), it
> might shed some light onto this.
>
> But I'm probably overly conservative here.
>
Besides select.h files I can find fd_mask only in
sunrpc/svc.c, how would unsigned change this.
I am ok with this patch.