[Top] [All Lists]

Re: The mmap() problem is back

To: linux-xfs@xxxxxxxxxxx
Subject: Re: The mmap() problem is back
From: Wessel Dankers <wsl@xxxxxxxxxxxx>
Date: Sat, 3 Aug 2002 16:24:05 +0200
In-reply-to: <9862.1028384103@ocs3.intra.ocs.com.au>
Mail-followup-to: linux-xfs@xxxxxxxxxxx
References: <20020803132927.GA576@fruit.eu.org> <9862.1028384103@ocs3.intra.ocs.com.au>
Sender: owner-linux-xfs@xxxxxxxxxxx
User-agent: Mutt/1.4i
On 2002-08-04 00:15:03+1000, Keith Owens wrote:
> I don't see this problem, running xfs 2.4.19.  mapcheck finds no
> errors, the test case from
> http://marc.theaimsgroup.com/?l=linux-xfs&m=102408887618294&w=2 passes.
> However I am running gcc 2.96, not 3.1.  Can you supply a small test
> case that always fails?

It's the compiler itself that fails, once the source file is corrected it
produces a correct executable. A simple test case for me is:

% rm -f foo
% dmesg >foo
% ls -l foo
-rw-r-----    1 wsl      wsl          8557 Aug  3 16:18 foo
% mapcheck foo
1 files scanned 1 files fixed 0 errors
% mapcheck foo
1 files scanned 0 files fixed 0 errors

The bug crops up when the file is created. It also doesn't happen 100% of
the time: when I created a "bar" file to test it again, it passed. I'm not
able to determine what triggers the problem.

Kind regards,

Wessel Dankers <wsl@xxxxxxxxxxxx>

You've been infected by the Telescoping Hubble virus.

<Prev in Thread] Current Thread [Next in Thread>