gnome-system-monitor showing impossible amounts of memory

ProblemType: Bug
Architecture: amd64
Date: Fri Dec 7 14:49:04 2007
DistroRelease: Ubuntu 7.10
ExecutablePath: /usr/bin/gnome-system-monitor
NonfreeKernelModules: fglrx
Package: gnome-system-monitor 2.20.1-0ubuntu1
PackageArchitecture: amd64
ProcCmdline: gnome-system-monitor
ProcCwd: /home/sauron
ProcEnviron:
 PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games
 LANG=en_IN
 SHELL=/bin/bash
SourcePackage: gnome-system-monitor
Uname: Linux Mordor 2.6.22-14-generic #1 SMP Sun Oct 14 21:45:15 GMT 2007 x86_64 GNU/Linux

thanks for your report, Does the same happens with another user? Can you tell us a few steps to trigger the bug? thanks.

Just open the GNOME System monitor... click on "Memory" a few times. The other user in the forum has also been using 64 bit Ubuntu, I do not know whether this bug occurs in 32 bit.

do you use xgl? does it happen with a normal xorg server?

How do I find out?

Thanks for your comments. This does not appear to be a bug report as such. We appreciate the difficulties you are facing, but it would make more sense to raise your question in the support tracker. http://launchpad.net/support

I have the same problem on two quad core machines. I did not have this problem on my previous install on either machine. I reinstalled because I messed up the compiz install and could log in, but had a blank screen. I now have a working compiz/fglx install. The only difference between the installs is this time I installed xserver-xgl versus the default. Neither machine did this before, now they both do.

Confirmed on a Core2 Duo PC with Gutsy amd64

2.6.22-14-generic #1 SMP Tue Feb 12 02:46:46 UTC 2008 x86_64 GNU/Linux

Could this be an easter egg?

http://66.102.9.104/search?q=cache:xIINkbxei0sJ:www.cameronmoll.com/archives/2007/04/surviving_the_allday_tech_interview/+distance+17179869180&hl=nl&ct=clnk&cd=1

<quote>
13 David Yeiser ~ 04 April 2007

The moon one is actually a lot less folds than I thought when I first heard this riddle.

Assume the phone book to be 2” thick:

1 fold (1f) = 2 x previous thickness.

1f = 4”
2f = 8”
3f = 16”
4f = 32”
5f = 64”
…
33f = 17179869180” = 271,147 miles

Distance from the earth to the moon = 238,906 miles.

It’s based on geometric growth, so you can produce a formula that would give you the height of the resulting folds with the variables being the thickness of the phone book and the number of folds.

t0 = original thickness of phone book
f = number of folds
h = height of phone book after folds

EQ:
h = t0*(2^f)

So for 33 folds:
h = (2)*(2^33) = 17179869180”

Okay, I’m done now. Feel free to delete this comment if it’s too far off-topic!
</quote>

I have this same thing happening to a few of my (I think 32 bit) applications. Note that a small experiment with the calculator shows that the value given (17179869180.0 GB) is exactly ((2^64-2^32)/(1024^3)) as in, it's what happens when you do memory&0xFFFFFFFF, and memory happens to be a 64 bit hex number, with the upper 32 bits all being ones.

If you get an application to use 100M of ram, it'll show it as 17179869180.1, etc. It's just a display-scale error.

Seems like bug is related with
return value of X_XResQueryClientPixmapBytes query

the package obtains xserver usage value for gnome-system-monitor is libxres

changing line 224 of XRes.c
from
*bytes = (rep.bytes_overflow * 4294967295) + rep.bytes;
to
*bytes = rep.bytes_overflow ? ( -1 * rep.bytes ) : rep.bytes;

seem to give more reasonable results