From: George Smith
Subject: cmucl *** Sigsegv in page not marked as write protected
Date: 
Message-ID: <m2hf1w59v5.fsf@bloomfield.phil1.uni-potsdam.de>
I'm running cmucl on red hat linux (kernal-2.2.5-22) . The cmucl
version is from the debian package (2.4.20). I'm not sure how that
translates into cmucl release numbers. In any case, I get the
following on startup:

	CMU Common Lisp release x86-linux 2.4.20  1 June 2000 build
	710

Sometimes I get memory allocation errors starting like this:

	munmap: Cannot allocate memory
	Error in allocating memory

Then come hints on how to allocate more memory:

--------------------------------------------------------------
[...]
Now you have two choices:
 - Accept this and lift the kernel and other limits by doing:
 as root:
 echo 1 > /proc/sys/vm/overcommit_memory
 as the user:
 ulimit -d unlimited 
 ulimit -v unlimited 
 ulimit -m unlimited 

  - Try to use the lazy-allocation routines. They are pretty
    experimental

 and might interact badly with some kernels. To do this start lisp
 with the "-lazy" flag, like:
 lisp -lazy
[...]
---------------------------------------------------------------

After a number of such messages in short succession I get rapid fire:

	*** Sigsegv in page not marked as write protected: 49c86000 1
            460 1024

Up to now, I always start cmucl with 

	lisp -lazy

When I try the other suggestion, ulimit reacts fine with -d and -m,
but with -v I get the following:

	$ ulimit -v unlimited
	ulimit: cannot raise limit: Invalid argument

and cmucl (now without -lazy) won't start, complaining about an "Error
in allocating memory".

The machine has plenty of physical memory (512 MB). Any suggestions on
how I could make use of it?

Thanks for any help,

George
From: Jochen Schmidt
Subject: Re: cmucl *** Sigsegv in page not marked as write protected
Date: 
Message-ID: <96h1vb$kf2mb$1@ID-22205.news.dfncis.de>
George Smith wrote:

> I'm running cmucl on red hat linux (kernal-2.2.5-22) . The cmucl
> version is from the debian package (2.4.20). I'm not sure how that
> translates into cmucl release numbers. In any case, I get the
> following on startup:

All your problems will disappear when you use a newer Kernel.
2.2.5 is rather old and buggy...

Regards,
Jochen
From: Jochen Schmidt
Subject: Re: cmucl *** Sigsegv in page not marked as write protected
Date: 
Message-ID: <96h2fl$kf2mb$2@ID-22205.news.dfncis.de>
Jochen Schmidt wrote:

> George Smith wrote:
> 
>> I'm running cmucl on red hat linux (kernal-2.2.5-22) . The cmucl
>> version is from the debian package (2.4.20). I'm not sure how that
>> translates into cmucl release numbers. In any case, I get the
>> following on startup:
> 
> All your problems will disappear when you use a newer Kernel.
> 2.2.5 is rather old and buggy...

I use 2.2.16 and if I remember right the Memory problems disappeared 
somewhere around 2.2.12.

If you update, then CMUCL runs fine without "lazy" Option.

Regards,
Jochen
 
From: George Smith
Subject: Re: cmucl *** Sigsegv in page not marked as write protected
Date: 
Message-ID: <m2u25uby3x.fsf@bloomfield.phil1.uni-potsdam.de>
Jochen Schmidt <ยทยทยท@dataheaven.de> writes:

> Jochen Schmidt wrote:
> 
> > George Smith wrote:
> > 
> >> I'm running cmucl on red hat linux (kernal-2.2.5-22) . The cmucl
> >> version is from the debian package (2.4.20). I'm not sure how that
> >> translates into cmucl release numbers. In any case, I get the
> >> following on startup:
> > 
> > All your problems will disappear when you use a newer Kernel.
> > 2.2.5 is rather old and buggy...
> 
> I use 2.2.16 and if I remember right the Memory problems disappeared 
> somewhere around 2.2.12.
> 
> If you update, then CMUCL runs fine without "lazy" Option.

My apologies for starting to get away from common lisp with my follow
up question, but maybe someone here using cmucl on linux has been
through something similar before. If not, I'll repost to a linux group
and hope that I can understand the responses there; linux isn't my
native language ;-)

I downloaded the and installed following rpm:

	kernel-2.2.17-14.i686.rpm

My previous lilo.conf file looked like this:


boot=/dev/sda
map=/boot/map
install=/boot/boot.b
prompt
timeout=50
image=/boot/vmlinuz-2.2.5-22smp
        label=linux
        root=/dev/sda8
        initrd=/boot/initrd-2.2.5-22smp.img
        read-only

After making an initrd image like this:

	/sbin/mkinitrd initrd-2.2.17-14.img 2.2.17-14

I added the following (as the first image)

image=/boot/vmlinuz-2.2.17-14
        label=linux-new
        root=/dev/sda8
        initrd=/boot/initrd-2.2.17-14.img
        read-only

I updated the symbolic links that I found in /boot as follows:

System.map -> System.map-2.2.17-14
module-info -> module-info-2.2.17-14
vmlinuz -> vmlinuz-2.2.17-14

I ran /sbin/lilo and got no error messages.

Now, when I reboot, I get the following:

kernel panic: VFS: unable to mount root fs on 08:08

I got that message in previous attempts without the initrd image and
the relevant line in linux.conf

Does anyone know where I went wrong?

Thanks for any help,

George
comp.lang.lisp / 2001 / feb