[Top] [All Lists]

Re: several messages

To: Russell Cattelan <cattelan@xxxxxxxxxxx>, Steve Lord <lord@xxxxxxx>
Subject: Re: several messages
From: Jim Bray <jb@xxxxxxxxx>
Date: Mon, 3 Apr 2000 23:28:53 -0400 (EDT)
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <200004040106.UAA07079@xxxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
On Mon, 3 Apr 2000, Russell Cattelan wrote:

> remake mkfs and send me the new binary.

  May not be necessary: at least I'm getting something sensible now:

program received signal SIGSEGV, Segmentation fault.
0x8049d5e in main (argc=2, argv=0xbffffc74) at xfs_mkfs.c:1195
1195                    if (logfile && *logfile) {
(gdb) x logfile
0x1:    Cannot access memory at address 0x1.

  logfile is a stack variable, and looks like it is getting used before
set. added an '=(char*)0' after the declaration, and we get past that
crash, and on to the other one, which looks the same... yeah, it looks
like this is still blowing the stack somehow in uuid_create. This is
actually in kernel code which is wired with ifdefs so it can be shared
with the utilities (....?); also, it seems pretty clear that the
#ifdef _KERNEL version of uuid_create is being used instead of the user
version, which at first glance might seem more apropos.

  Anyways, I'll send Russell a fresh binary. BTW, in case my static libc
had gotten blown (I don't use it much), I recompiled without -static and
got the same crash.

Jim     http://as220.org/jb

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