Date: Thu, 27 Jul 2000 17:47:14 -0400 (EDT) From: Andrew Gallatin <gallatin@cs.duke.edu> To: Mark Abene <phiber@radicalmedia.com> Cc: "David O'Brien" <obrien@FreeBSD.ORG>, freebsd-alpha@FreeBSD.ORG Subject: Re: select problems ( was Re: weird XFree86 3.3.6 behavior...) Message-ID: <14720.44407.464338.134415@grasshopper.cs.duke.edu> In-Reply-To: <20000727174257.A19445@radicalmedia.com> References: <20000727022930.A9638@radicalmedia.com> <20000727000033.A44502@dragon.nuxi.com> <20000727035128.A9982@radicalmedia.com> <20000727164833.A18644@radicalmedia.com> <14720.41641.666813.600170@grasshopper.cs.duke.edu> <20000727174257.A19445@radicalmedia.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Mark Abene writes: > On Thu, Jul 27, 2000 at 05:06:55PM -0400, Andrew Gallatin wrote: > > They're using real fd_sets as args to select, however it appears they are > doing mask arithmetic with ints. This may be the problem, since an fd_set > on Alpha is 64 bits. I'll try a little experiment... That sounds suspicious.. > > What was the nature of this bug you mention? And yes, I built X from the See the CVS logs for Peter's commit to sys/types.h and kern/sys_generic.c on Feb. 20th: Fix select(2) for the Alpha. (!!) It was never returning true for fd's in the range of 32-63, 96-127 etc. The first problem was the FD_*() macros were shifting a 32 bit integer "1" left by more than 32 bits. The same problem happened in selscan(). ffs() also takes an int argument and causes failure. For cases where int == long (ie: the usual case for x86, but not always as gcc can have long being a 64 bit quantity) ffs() could be used. Drew To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?14720.44407.464338.134415>