atom feed6 messages in org.freebsd.freebsd-currentRe: mmap() + NFS == &*$%@$!!!
FromSent OnAttachments
Bill PaulNov 28, 1997 8:10 pm 
Karl DenningerNov 28, 1997 9:04 pm 
Tom BartolDec 5, 1997 9:29 am 
Karl DenningerDec 5, 1997 9:36 am 
John S. DysonDec 5, 1997 9:37 am 
John S. DysonDec 5, 1997 9:42 am 
Subject:Re: mmap() + NFS == &*$%@$!!!
From:Karl Denninger (ka@mcs.net)
Date:Dec 5, 1997 9:36:46 am
List:org.freebsd.freebsd-current

Yep. I'm running with 1.41 of that file here as well. Later versions are NOT stable. Frankly, I'm surprised that John or someone else hasn't backed out the commits beyond 1.41 by now - this is an old problem, dating back a few months.

On Fri, Dec 05, 1997 at 09:29:31AM -0800, Tom Bartol wrote:

FYI,

Replacing version 1.44 of /sys/nfs/nfs_bio.c with version 1.41 as described below fixed the "Fatal trap 12: page fault while in kernel mode" error that I had been getting frequently while running pine (my /var/mail is an NFS mounted filesystem). NFS read and write performance were not affected by backing out to version 1.41 (to the best of my ability to measure performance).

On Fri, 28 Nov 1997, Karl Denninger wrote:

Try using "cvs update -r1.41 nfs_bio.c" in your "sys/nfs" directory and rebuild the kernel, then see if the problem goes away.

There are some SERIOUS willies in versions of that file after 1.41.

On Fri, Nov 28, 1997 at 11:11:11PM -0500, Bill Paul wrote:

I did a quick install of the Nov. 28th -current snapshot to test a few NIS+ related things and ran into a problem using mmap() with NFS. The problem is that it's apparently not too hard at all to wedge either a process or the entire system when using mmap() with files on a remote NFS filesystem.