tthtlc / compcache

Automatically exported from code.google.com/p/compcache
0 stars 0 forks source link

compcache 0.4 crashes on ARM #2

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
1. use compcache by swapon /dev/compcache0 -p 1or swapon /dev/compcache -p 1
2.
3.

What is the expected output? What do you see instead?
frozen screen

What version of the product are you using? On what operating system?
latest. ubuntu 7.10 , 8.04 alpha 5, debian-eabi 2.6.24, 2.6.23, 2.6.21

Please provide any additional information below.
 this project is not stable, it will lock up any linux machine, SMP or not,
embedded or desktop.

Original issue reported on code.google.com by marblema...@gmail.com on 3 Mar 2008 at 8:01

GoogleCodeExporter commented 9 years ago
its been running for 21 hours now without running x, so i think we can say its x
thats crashing now. It looked like it was frozen to me because my touchscreen,
keyboard stopped working but i was still able to SSH over to the machine so it 
wasent
frozen at all! compcache-0.2 is freezing X sessions when used with xscale and
debian-eabi lenny/sid. I like using the terminal anyways so i dont mind, its 
well
worth the performance increase, any idea why its crashing X? im using icewm and
icewm-experimental.

Original comment by marblema...@gmail.com on 9 Mar 2008 at 6:59

GoogleCodeExporter commented 9 years ago
are you running fedora over windows using vm? do you use linux without 
emulation? any
clues as to why compcache cannot work well with aptitude, apt-get or x? I get
segmentation faults if i use aptitude.

Original comment by marblema...@gmail.com on 9 Mar 2008 at 7:07

GoogleCodeExporter commented 9 years ago
compcache is stable on my system as long as i do not use the 3 above mentioned

Original comment by marblema...@gmail.com on 9 Mar 2008 at 7:07

GoogleCodeExporter commented 9 years ago
your vmware workstation 6 is running on microsoft windows? xp?

Original comment by marblema...@gmail.com on 9 Mar 2008 at 7:08

GoogleCodeExporter commented 9 years ago
heres a new snapshot, its been running for nearly a whole day 

Original comment by marblema...@gmail.com on 9 Mar 2008 at 7:09

Attachments:

GoogleCodeExporter commented 9 years ago
The issue should be changed to something along the lines of "Intel XScale 
X-Server
Crashes when used with Compcache-0.2"

Original comment by marblema...@gmail.com on 9 Mar 2008 at 7:10

GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
Using 64mb's of ram sucks without compcache-0.2, compcache-0.2 is my work area, 
lol!
compcache-0.2 gives me the reaction speed and responsiveness I wanted without
upgradeing the ram BGA (Ball Grid Array) modules. I cant wait till I can use
compcache with x and apt-get aptitude and firefox iceweasel and icehamster!! 
This
project makes 64MB's of ram into enough memory. I LOVE this project, even tho it
freezes I cant stop using it... performance increase is just too good. Thanks 
for
owning the project.The bug only effects ARM Processors i believe since its 
working
great on any x86... is there any way i can change the subject of this issue?  
I'll
try and stop posting unless i find something useful, thanks again for making 
64mb's
of ram into 90 or 128! or maybe even more? Anybody out there using using linux 
and
low ram should use compcache-0.2! so whens compcache-0.3 coming out? I was 
watching
this project befor it went to google code and i tried using the kernel patches 
but it
never worked.. i always wondered  why there was no way to make modules like the 
ones
in compcache-0.2 then just load them and enjoy compressed ram! I'm not going to 
stop
using compcache so if and when i find anything useful I will post it asap! 

Also i was able to run aptitude and update it, also apt-get update worked! I 
believe
X was crashing and as long as i DO NOT RUN X on XScale , Compcache-0.2 is very 
stable
and very fast! Thank you!! :D 

Original comment by marblema...@gmail.com on 9 Mar 2008 at 8:02

GoogleCodeExporter commented 9 years ago
I can still use apt-get update, however i get this error occasionally only when 
i use
compcache-0.2, it doesnt freeze i just get an error:

Get:64 http://nl.arm.mirror.debian.net sid/main 2008-03-09-0829.50.pdiff [8997B]
Get:65 http://nl.arm.mirror.debian.net sid/main 2008-03-09-0829.50.pdiff [8997B]
Get:66 http://armel-debs.applieddata.net sid/main 2008-03-08-2037.26.pdiff [32.6
kB]
Get:67 http://ftp.uk.debian.org testing/main 2008-03-09-0829.50.pdiff [47.4kB] 
*** glibc detected *** /usr/lib/apt/methods/http: free(): invalid next size (fas
t): 0x0001c3e0 ***
*** glibc detected *** apt-get: double free or corruption (!prev): 0x00055bb8 **
*
======= Backtrace: =========
/lib/libc.so.6[0x4030b328]
/lib/ld-linux.so.3[0x40025000]
======= Memory map: ========
00008000-00029000 r-xp 00000000 b3:01 786079     /usr/bin/apt-get
00030000-00031000 rw-p 00020000 b3:01 786079     /usr/bin/apt-get
00031000-000e3000 rwxp 00031000 00:00 0          [heap]
40000000-4001d000 r-xp 00000000 b3:01 1503944    /lib/ld-2.7.so
4001d000-40022000 rw-p 4001d000 00:00 0 
40024000-40026000 rw-p 0001c000 b3:01 1503944    /lib/ld-2.7.so
40037000-400f4000 r-xp 00000000 b3:01 576535     /usr/lib/libapt-pkg-libc6.7-6.s
o.4.6.0
400f4000-400fc000 ---p 000bd000 b3:01 576535     /usr/lib/libapt-pkg-libc6.7-6.s
o.4.6.0
400fc000-400fe000 rw-p 000bd000 b3:01 576535     /usr/lib/libapt-pkg-libc6.7-6.s
o.4.6.0
400fe000-40100000 r-xp 00000000 b3:01 1501562    /lib/libutil-2.7.so
40100000-40107000 ---p 00002000 b3:01 1501562    /lib/libutil-2.7.so
40107000-40108000 r--p 00001000 b3:01 1501562    /lib/libutil-2.7.so
40108000-40109000 rw-p 00002000 b3:01 1501562    /lib/libutil-2.7.so
40109000-401d2000 r-xp 00000000 b3:01 577821     /usr/lib/libstdc++.so.6.0.10
401d2000-401da000 ---p 000c9000 b3:01 577821     /usr/lib/libstdc++.so.6.0.10
401da000-401dc000 r--p 000c9000 b3:01 577821     /usr/lib/libstdc++.so.6.0.10
401dc000-401df000 rw-p 000cb000 b3:01 577821     /usr/lib/libstdc++.so.6.0.10
401df000-401e4000 rw-p 401df000 00:00 0 
401e4000-40286000 r-xp 00000000 b3:01 1503846    /lib/libm-2.7.so
40286000-4028d000 ---p 000a2000 b3:01 1503846    /lib/libm-2.7.so
4028d000-4028e000 r--p 000a1000 b3:01 1503846    /lib/libm-2.7.so
4028e000-4028f000 rw-p 000a2000 b3:01 1503846    /lib/libm-2.7.so
4028f000-4029b000 r-xp 00000000 b3:01 1503774    /lib/libgcc_s.so.1
4029b000-402a2000 ---p 0000c000 b3:01 1503774    /lib/libgcc_s.so.1
402a2000-402a3000 rw-p 0000b000 b3:01 1503774    /lib/libgcc_s.so.1
402a3000-403bb000 r-xp 00000000 b3:01 1503796    /lib/libc-2.7.so
403bb000-403c2000 ---p 00118000 b3:01 1503796    /lib/libc-2.7.so
403c2000-403c3000 r--p 00117000 b3:01 1503796    /lib/libc-2.7.so
403c3000-403c5000 rw-p 00118000 b3:01 1503796    /lib/libc-2.7.so
403c5000-403c8000 rw-p 403c5000 00:00 0 
403c8000-4053d000 r--p 00000000 b3:01 755796     /usr/lib/locale/locale-archive
beb1e000-beb33000 rwxp befeb000 00:00 0          [stack]
/bin/apt.sh: line 3: 11316 Aborted                 apt-get update
Reading package lists... 6%

Original comment by marblema...@gmail.com on 9 Mar 2008 at 8:35

GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
I used ./use_compcache 60000 and was able to apt-get update without any 
problems and
then apt-get upgrade was able to download all the new packages and get up to
prepareing rails, then it segfaulted and everything stopped, ssh terminals died 
no
input would work on the device to restart ssh, but 2 ssh terminals were still 
running
tcpdump and apm -M (apm monitor) as seen here in this screenshot, so its not 
freezing
but its crashing some servcies and its not running out of memory because it has 
over
49mb's of compressed ram free.

Original comment by marblema...@gmail.com on 10 Mar 2008 at 2:40

Attachments:

GoogleCodeExporter commented 9 years ago
error i got with apt-get upgrade then ssh crashed and all input crashed but 
tcpdump
and apm -M still running for some reason so the system isint frozen

/bin/sh: line 1: 17322 Segmentation fault      /usr/sbin/dpkg-preconfigure --apt

Original comment by marblema...@gmail.com on 10 Mar 2008 at 3:06

GoogleCodeExporter commented 9 years ago
Can you also provide following info:
- output of apt-get running under strace.
- backtrace of apt-get when it crashes: run apt-get under gdb and do 'bt' at gdb
prompt       
   (which you will get after apt-get crashes) to get the backtrace.
- after crash, provide contents of:
   - /proc/{meminfo, cpuinfo, tlsfinfo, compcache
   - /var/log/{messages, kernel.log}

Original comment by nitingupta910@gmail.com on 10 Mar 2008 at 9:07

GoogleCodeExporter commented 9 years ago
All seems to work fine in a terminal (except apt-get) as comments above 
suggests.

Original comment by nitingupta910@gmail.com on 10 Mar 2008 at 9:13

GoogleCodeExporter commented 9 years ago
Can you also try to repro this issue on x86 be setting memory to 64M and run 
same
stuff there (apt-get, x-server)? You can set mem to 64M by either passing 
mem=64m
param to kernel at boot time or use VM with RAM=64M.
This will make sure its ARM specific issue. I tried above config myself on my 
x86 and
compcache works without any problem.

(yeah, I use vmware workstation with winxp as host).

Original comment by nitingupta910@gmail.com on 10 Mar 2008 at 9:27

GoogleCodeExporter commented 9 years ago
yes. I just did actually.... used compcache-0.2 on ubuntu hardy heron alpha-6 
using
25% of ram... went to sleep woke up 12 hours later only to find the computer is 
using
all of its compcache on NetworkManager....? then i tried to switch virtual 
desktops
in icewm-session-experimental by using Ctrl+Alt+Left... AS SOON AS I PRESSED 
those
buttons... the machine completely frozen. I tried to ssh, ssh did not work, 
tried to
ping the IP Address, it worked. The machine wont reboot, cant ssh, its crashed 
X....

I'll go post a pair of /var/log/messages and /var/log/kern.log right now from 
that
machine.. i hope theres something there

Original comment by marblema...@gmail.com on 10 Mar 2008 at 7:30

GoogleCodeExporter commented 9 years ago
and the above mentioned i will also add

Original comment by marblema...@gmail.com on 10 Mar 2008 at 7:31

GoogleCodeExporter commented 9 years ago
subject name should probably be changed back to remove the word Armel

Original comment by marblema...@gmail.com on 10 Mar 2008 at 7:31

GoogleCodeExporter commented 9 years ago
debian X crashes on x86 and apt-get crashes on armel and xcrashes on armel

Original comment by marblema...@gmail.com on 10 Mar 2008 at 7:32

GoogleCodeExporter commented 9 years ago
you gotta try linux without windows emulation, its way more amazing

Original comment by marblema...@gmail.com on 10 Mar 2008 at 7:34

GoogleCodeExporter commented 9 years ago
it looks like the x86 is restarting X... painfully slowly

Original comment by marblema...@gmail.com on 10 Mar 2008 at 7:36

GoogleCodeExporter commented 9 years ago
its using up all the CPU on random stuff it wouldent ever need ... i wish i 
could
give you a screenshot but its just so slow now, usually this desktop is 
insanely fast
with 2gbs of ram.  not all the memory is gone 

Original comment by marblema...@gmail.com on 10 Mar 2008 at 7:38

GoogleCodeExporter commented 9 years ago
i cant even move the mouse! this is ridiclious!

Original comment by marblema...@gmail.com on 10 Mar 2008 at 7:40

GoogleCodeExporter commented 9 years ago
> i cant even move the mouse! this is ridiclious!

This is ridiculous!!! Why I can never ever repro this issue *whatever* I do on 
x86.
I will now try test it exhaustively with every possible mem size on physical 
x86 h/w.
I want this problem to happen on my machine. Here I remove all this windows 
stuff....

Original comment by nitingupta910@gmail.com on 10 Mar 2008 at 7:48

GoogleCodeExporter commented 9 years ago
>This is ridiculous!!! Why I can never ever repro this issue *whatever* I do on 
x86.
>I will now try test it exhaustively with every possible mem size on physical 
x86 h/w.
>I want this problem to happen on my machine. Here I remove all this windows 
stuff....

try using native linux and debian based os instead of fedora like ubuntu

Original comment by marblema...@gmail.com on 10 Mar 2008 at 11:12

GoogleCodeExporter commented 9 years ago
windows xp always messes stuff up 

Original comment by marblema...@gmail.com on 10 Mar 2008 at 11:13

GoogleCodeExporter commented 9 years ago
im using Ubuntu 8.04-i386 Alpha-6 x86 on AMD64 with 2gb's of ram using 25% of 
that
ram as compcache (512mb compcache swap drive)

Original comment by marblema...@gmail.com on 10 Mar 2008 at 11:14

GoogleCodeExporter commented 9 years ago
also this bug takes a very long time to become noticable

Original comment by marblema...@gmail.com on 10 Mar 2008 at 11:20

GoogleCodeExporter commented 9 years ago
i left my comp on overnight, i had totally been convinced at that point it was
stable. I was ready to install it on a clients machine because he only has 
512mb's of
ram!! But when I woke up I learned that was not the case unfortunately.

Original comment by marblema...@gmail.com on 10 Mar 2008 at 11:22

GoogleCodeExporter commented 9 years ago
VMware manages memory virtually for linux to run inside of windows, I would 
think
that could make a very big diffrence in the way your linux distro manages random
access memory and compressed random access memory. I think  by going native you
surely will repo this issue.

Original comment by marblema...@gmail.com on 10 Mar 2008 at 11:40

GoogleCodeExporter commented 9 years ago
i turn off swap when i use compcache because i just dont use it, i hope thats 
ok?

Original comment by marblema...@gmail.com on 10 Mar 2008 at 11:56

GoogleCodeExporter commented 9 years ago
when this does work on native linux it's going to be awesome that_is_for_sure! 
I am
trying to repair my drives that were unproperly dismounted in the compcache-0.2 
crash
on my x86 single core AMD64 desktop running Ubuntu-i386 8.04 Alpha-6. when the 
fsck
is finished I'll post those logs/files you spoke of earlier.. I hope i can help

Original comment by marblema...@gmail.com on 11 Mar 2008 at 12:02

GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
Im using the DVD of Ubuntu 8.04 alpha-6 32-bit, linux-kernel-2.6.24-12

if you download the DVD check the md5sum and burn then install it then connect 
to the
net sudo root and:apt-key list ; apt-key update ; apt-get update ; apt-get 
upgrade ;
apt-get dist-upgrade ; apt-get dselect-upgrade ; apt-get -f install ; apt-get
autoclean ; apt-get autoremove ; apt-get clean ; do-release-upgrade ; uname -a

then reboot and build compcache-0.2 (the build sources are pre-installed 
prebuilt) so
you can just run it right away, you dont need to compile a new kernel.
then just run ./use_compcache and wait. That will reproduce this issue. The 
wait part
is crucial and it takes a very long time to happen.

Original comment by marblema...@gmail.com on 11 Mar 2008 at 12:52

GoogleCodeExporter commented 9 years ago
must let the system stay on overnight and maybe longer. Keep top running as i 
did

Original comment by marblema...@gmail.com on 11 Mar 2008 at 12:53

GoogleCodeExporter commented 9 years ago
those steps should reproduce this issue

Original comment by marblema...@gmail.com on 11 Mar 2008 at 12:53

GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
tlsfinfo, compcache were not available in proc after rebooing from the crash.. 
ill
post kern.log messages cpuinfo and meminfo from the crashed x86

Original comment by marblema...@gmail.com on 11 Mar 2008 at 3:05

GoogleCodeExporter commented 9 years ago
> Can you also provide following info:
> - output of apt-get running under strace.
> - backtrace of apt-get when it crashes: run apt-get under gdb and do 'bt' at 
gdb
> prompt       
>    (which you will get after apt-get crashes) to get the backtrace.
> - after crash, provide contents of:
>    - /proc/{meminfo, cpuinfo, tlsfinfo, compcache
>    - /var/log/{messages, kernel.log}

Can you please specify how i should use strace, I have heard of the program 
before
but Im not sure exactly how to use it, I should just install it and read the 
manual. 
how do i use gdb? sorry, Im not familiar with those programs, but ill install 
them
read the manuals and do as much as i can for now! thanks for everything

Original comment by marblema...@gmail.com on 11 Mar 2008 at 7:49

GoogleCodeExporter commented 9 years ago
> Can you please specify how i should use strace

strace <program_name> -o compcache_strace_debug.txt

This will write out what syscalls are being called by given program into given 
output
file (compcache_strace_debug.txt)

> how do i use gdb?
gdb <program_name>

when program crashes, it will drop you to gdb prompt instead of shell. You can 
then
do various debugging things. In this case I only want output of these commands 
on
that gdb prompt:
1. backtrace
2. info locals

I now have Ubuntu and will try to repro it on my machine also.
Thanks for your testing efforts.

Original comment by nitingupta910@gmail.com on 11 Mar 2008 at 8:12

GoogleCodeExporter commented 9 years ago
Are you using Ubuntu x86 DVD of Alpha-6 8.04 Hardy Heron? Because thats the 
ticket.

Original comment by marblema...@gmail.com on 11 Mar 2008 at 8:32

GoogleCodeExporter commented 9 years ago
[deleted comment]