From owner-freebsd-hackers Thu Jun 19 10:04:58 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id KAA05969 for hackers-outgoing; Thu, 19 Jun 1997 10:04:58 -0700 (PDT) Received: from ui-gate.utell.co.uk (root@ui-gate.utell.co.uk [194.200.4.253]) by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id KAA05960; Thu, 19 Jun 1997 10:04:50 -0700 (PDT) Received: from utell.co.uk (shift.utell.net [97.3.0.21]) by ui-gate.utell.co.uk (8.7.6/8.7.3) with ESMTP id SAA05474; Thu, 19 Jun 1997 18:04:36 +0100 (BST) Received: (from brian@localhost) by utell.co.uk (8.8.5/8.8.5) id SAA24424; Thu, 19 Jun 1997 18:04:36 +0100 (BST) Date: Thu, 19 Jun 1997 18:04:36 +0100 (BST) Message-Id: <199706191704.SAA24424@utell.co.uk> Mime-Version: 1.0 X-Newsreader: knews 0.9.8 Reply-To: brian@awfulhak.org, brian@utell.co.uk Organization: Awfulhak Ltd. References: <33956BBB.1BB@nospam.com> <5n7bk7$osi@uriah.heep.sax.de> <339863C7.1DF5@nospam.com> <5nbrkn$2h6@uriah.heep.sax.de> <339D9548.7B07@nospam.com> <5ob5e7$31j@ui-gate.utell.co.uk> From: brian@utell.co.uk (Brian Somers) Subject: Re: xperfmon++ problem (again) Newsgroups: comp.unix.bsd.freebsd.misc Cc: jmz@freebsd.org, freebsd-hackers@freebsd.org Content-Type: text/plain; charset=us-ascii Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk [cc'd to jmz@FreeBSD.org (the XFree86 maintainer for FreeBSD)] [cc'd to freebsd-hackers@FreeBSD.org] [posted to comp.unix.bsd.freebsd.misc] In article <5ob5e7$31j@ui-gate.utell.co.uk>, brian@shift.utell.net (Brian Somers) writes: > In article <339D9548.7B07@nospam.com>, > Marvin Foster writes: >> J Wunsch wrote: >>> >>> Marvin Foster wrote: >>> >>> > Sorry. I was doing this from memory. The message was >>> > >>> > Error: Cannot perform realloc >>> >>> Well, i was checking all ``Cannot'' messages in the source, and none >>> didn't come close to what you've mentioned. >>> >>> Of course, there's a chance that it's a message from some X11 library. >>> Sorry for not being more helpful, but all i can say is that it works >>> for me (and used to work all the time). >>> >>> -- >>> cheers, J"org >>> >>> joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE >>> Never trust an operating system you don't have sources for. ;-) >> >> I did the following: >> >> [pozzo:/usr/home/mefoster] ldd /usr/X11R6/bin/xperfmon++ >> /usr/X11R6/bin/xperfmon++: >> -lXaw.6 => /usr/X11R6/lib/libXaw.so.6.1 (0x8022000) >> -lXt.6 => /usr/X11R6/lib/libXt.so.6.0 (0x8058000) >> -lXmu.6 => /usr/X11R6/lib/libXmu.so.6.0 (0x8099000) >> -lX11.6 => /usr/X11R6/lib/libX11.so.6.1 (0x80a9000) >> -lkvm.2 => /usr/lib/libkvm.so.2.0 (0x813f000) >> -lc.3 => /usr/lib/libc.so.3.0 (0x8143000) >> [pozzo:/usr/home/mefoster] >> [pozzo:/usr/home/mefoster] strings /usr/X11R6/lib/libXt.so.6.0 | grep >> realloc >> realloc >> _realloc >> [pozzo:/usr/home/mefoster] strings /usr/X11R6/lib/libXt.so.6.0 | grep >> perform >> Cannot perform %s >> [pozzo:/usr/home/mefoster] >> >> I'm sort of new to FreeBSD, and not very expert at unix debugging, but >> I will keep looking into it and submit a problem report if I find >> anything. > > Ouch. I just built XFree86-3.3, and I'm now getting this from knews. > I rebuilt knews too - no change. > > It happens just after I press the post key !!!!!! > > This is on a RELENG_2_2 machine built June 15. Even setting > /etc/malloc.conf to RAD doesn't yield any results - somethings > lying - realloc didn't fail ! I'll post any findings. > > Here goes again *cringe* Well, xc/lib/Xt/Alloc.c does the realloc (in function XtRealloc). I changed the diagnostic so that it says how much we're allocating, and it comes out as zero ! Ah ha ! shift $ cd /tmp shift $ cat >x.c < #include main() { char *x = (char *)malloc(0); printf("%p\n", x); x = (char *)realloc(x,0); printf("%p\n", x); return 0; } eof shift $ cc -o x x.c shift $ ./x 0x4030 0x0 shift $ Now this is *really* bad news AFAIK. X provides the following in FreeBSD.cf: #define Malloc0ReturnsNull NO but I don't know of any provisions for malloc returning a pointer and realloc returning NULL :( We could always do a simple patch to this file, but is this the right way to go ? -- Brian Don't _EVER_ lose your sense of humour !