search

kowr / 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 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
yeah compcache-0.2 works great on my x86's both with 2GB's of ram, no crashing 
at all.

Original comment by marblema...@gmail.com on 6 Mar 2008 at 7:22

GoogleCodeExporter commented 9 years ago 
is there any way to change the subject of this issue to PXA270 Unstable -- 
frozen in
the mean time while i test it? it seems stable to me on x86

Original comment by marblema...@gmail.com on 6 Mar 2008 at 7:23

GoogleCodeExporter commented 9 years ago 
im also going to try this on my PXA255 with 128mbs of ram. What would be the 
optimal
setting for that configuration? %10?

Original comment by marblema...@gmail.com on 6 Mar 2008 at 7:40

GoogleCodeExporter commented 9 years ago 
> im also going to try this on my PXA255 with 128mbs of ram. What would be the 
optimal
> setting for that configuration? %10?

Yes 10% should work well :)

Original comment by nitingupta910@gmail.com on 6 Mar 2008 at 7:57

GoogleCodeExporter commented 9 years ago 
ok the sources are almost done im going to do a new build of the latest 
compcache,
since this is a stability test should i use less than use_compcache.sh 10240 , 
say 
use_compcache.sh 8xxx somewhere in the 8mb range, should be stable, correct?

Original comment by marblema...@gmail.com on 6 Mar 2008 at 8:26

GoogleCodeExporter commented 9 years ago 
Yes. compcache in range 6-10M should be stable on 64M RAM system.
I suggest you start your test with 8M compcache.

Original comment by nitingupta910@gmail.com on 6 Mar 2008 at 8:34

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
Well this blows. I get all sorts of segmentation faults when using 
compcache-0.2 with
apt-get and also frozen soon after. Heres the logs. i hope this works eventually
because the performance increase is amazing... untill you freeze...

This was a brand new total rebuild. I put ALOT of work building EVERYTHING over 
from
scratch Just to get Frozen AGAIN...............................

I pray this helps

Original comment by marblema...@gmail.com on 6 Mar 2008 at 11:31

Attachments:

GoogleCodeExporter commented 9 years ago 
i was using ./use_compcache 8192just fyi

Original comment by marblema...@gmail.com on 6 Mar 2008 at 11:43

GoogleCodeExporter commented 9 years ago 
i was using ./use_compcache 8192 just fyi. also ive ran out of memory before 
without
using compcache and without using swap. My system has never frozen from using 
up all
of the ram it gets very sluggish but it never freezes up, it always just hangs 
for a
long time.. Compcaache is freezing my system.

Original comment by marblema...@gmail.com on 6 Mar 2008 at 11:46

GoogleCodeExporter commented 9 years ago 
I still don't see much in logs related to system freeze. I will now try to 
reproduce
this issue on similar configuration.

Original comment by nitingupta910@gmail.com on 7 Mar 2008 at 10:06

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 
if there is anything else i can do, dont hesitate to ask. Ill do anything to 
get this
compcache working

i do not mind testing and crashing since i always run bleeding edge software im 
used
to testing new things and dealing with issues that later get worked out.

Original comment by marblema...@gmail.com on 7 Mar 2008 at 4:18

GoogleCodeExporter commented 9 years ago 
noticed 3 updates on svn, Im rebuilding and retesting and re-uploading a new 
log set.
This time I will just let it sit and freeze and then upload the log, hopefully 
it
will contain useful information for debugging compcache

Original comment by marblema...@gmail.com on 8 Mar 2008 at 7:18

GoogleCodeExporter commented 9 years ago 
Still there are no changes that fixes any bug. I'm still trying to root-cause 
the issue.

Original comment by nitingupta910@gmail.com on 8 Mar 2008 at 7:33

GoogleCodeExporter commented 9 years ago 
heres what i got so far... I will continue to try and make myself useful/helpful

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

Attachments:

GoogleCodeExporter commented 9 years ago 
Do you get crash/freeze after a long time or immediately after loading 
compcache module?

Original comment by nitingupta910@gmail.com on 8 Mar 2008 at 9:06

GoogleCodeExporter commented 9 years ago 
You are using PXA2xx processor which is based on ARM. Can you try to repro this 
issue
on x86 machine? For this, you can simply create a virtual m/c and set RAM=64m 
to get
similar configuration. If it works fine on x86, then I will get some ARM 
emulator to
focus on arch specific issue.

Original comment by nitingupta910@gmail.com on 8 Mar 2008 at 9:10

GoogleCodeExporter commented 9 years ago 
Actually i just think i discoverd whats going on here. Im ssh'ed to the machine 
when
its "frozen" and its still reacting to input so its not actually frozen! I 
think its
just crashing xorg-xserver x11. So it appears to be that the device locks up and
doesnt accept any input from the keyboard (which is very bizzare because this 
device
is designed for linux only!) so yeah i guess its not actually crashing at all.

Original comment by marblema...@gmail.com on 8 Mar 2008 at 9:43

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
Not completely sure if this non-responsiveness is due to new packages from the
debian-unstable feed.. So im going to have to retry on a stable rootfs. lenny 
might
not be the best platform for testing, sid may be ok but i think im just going 
to use
a very basic stable branch debian etch rootfs fully updated, tomorrow just for
testing compcache-0.2 svn.

also im leaving it on overnight with the ssh terminal on the x86 desktop so ill 
know
whats up when i wake up hopefully.

Original comment by marblema...@gmail.com on 8 Mar 2008 at 9:50

GoogleCodeExporter commented 9 years ago 
Not 100% sure if this non-responsiveness is due to new packages from the
debian-unstable feed.. So im going to have to retry on a stable rootfs. lenny 
might
not be the best platform for testing, sid may be ok but i think im just going 
to use
a very basic stable branch debian etch rootfs fully updated, tomorrow just for
testing compcache-0.2 svn.

also im leaving it on overnight with the ssh terminal on the x86 desktop so ill 
know
whats up when i wake up hopefully.

I just also noticed when the icewm-session locked up and the keyboard stopped 
working
I was still able to get the system to suspend by pressing the power button, so i
should be able to tell if it is frozen by pressing that button after the crash 
or
"crash"occurs. The new 0.2 is a lot more stable on my system i think it might 
even
last overnight sitting idle. Its when i start using apt-get that i start having
problems i think. It would make more sense that compcache is just locking up
x-server, ill upload logs when i wake up and hopefully you can tell me whats 
going on?

Original comment by marblema...@gmail.com on 8 Mar 2008 at 10:01

GoogleCodeExporter commented 9 years ago 
Maybe this is the cause of system-freeze when its left idle:

I went through suspend code from http://suspend.sourceforge.net/intro.shtml
If this code tries to suspend to compcache swap (which is visible to userspace 
as
"yet another swap"), it will surely make system appear hung. For e.g:
reset_signature():suspend.c does while(1); if it fails to reset swap signature 
- and
in case of compcache it will surely fail here!

Original comment by nitingupta910@gmail.com on 8 Mar 2008 at 11:37

GoogleCodeExporter commented 9 years ago 
So, for testing puposes, make sure this suspend thing is completely disabled.

Original comment by nitingupta910@gmail.com on 8 Mar 2008 at 11:39

GoogleCodeExporter commented 9 years ago 
suspend works fine while using compcache i can suspend and wake up without 
issue. I
left the device on overnight using compcache, Its still working. I doubt its 
actually
crashing, its just not responding to ssh or input for some crazy ass reason. Im 
going
to restore and retry this using debian etch. The suspend fearture was added by
applieddatasystems.net, it works flawlessly, I am unaware of any way to disable 
it.
The system doesnt suspend unless I tell it to, It stays on with the AC adaptor 
and if
its about to run out of batteries then it suspends to save battery life.. So 
its not
suspend thats causing the problems, could just be lenny/sid on debian-eabi... I 
got
nothing else to work on but this so ill keep you posted on my findings.

Original comment by marblema...@gmail.com on 8 Mar 2008 at 6:17

GoogleCodeExporter commented 9 years ago 
Applied Data Systems Suspend is very stable however
(http://www.linuxdevices.com/news/NS8333074503.html) 

Sleep mode -- the entire board, including the StrongARM processor, required 
less than
10 milliwatts (0.01W) total power.

If you know any way to completely disable this fully functional feature, I'll 
disable
it for kicks. I highly doubt it will do anything for compcache however, since 
Ive
already confirmed that suspend works with compcache flawlessly. If you have any
suggestions i mean any suggestions for me, it'll be helpful. Im not going to 
stop
trying to get this working tho..i really love the performance i get with
compcache-0.2, Just wish i could figure out whats really going on...

Original comment by marblema...@gmail.com on 8 Mar 2008 at 6:38

GoogleCodeExporter commented 9 years ago 

"Idle" vs. "sleep"

ADS says previous power management schemes for Linux on StrongARM have been 
limited
to the "on", "off", and "idle" states. According to ADS, the idle and sleep 
power
management modes differ from each other as follows . . .

    * Idle mode -- a non-operational state initiated by the kernel due to a lack of
system activity, where the CPU is running at a slower rate. Peripherals are 
typically
unaffected. The system continues to process interrupts.

    * Sleep mode -- all system peripherals are powered off. Only the system's SDRAM
memory, the processor's internal real time clock, and certain of the 
processor's I/O
inputs remain powered on. This state is initiated by a call to a sleep function 
by
the application or by an inactivity timeout. Device drivers are notified of
transition to this state by a call to their respective power management 
functions.

Original comment by marblema...@gmail.com on 8 Mar 2008 at 7:34

GoogleCodeExporter commented 9 years ago 
Off topic question, dynamic swapspace seems alot faster than just using a 
regular
swapfile/partition... Is there any reason for this? could it be that several 
smaller
swapfiles can work faster than one huge one? or could it be the dynamic 
swapspace
manager its self?

Original comment by marblema...@gmail.com on 8 Mar 2008 at 7:36

GoogleCodeExporter commented 9 years ago 
The more I use this, the more i think its not crashing at all.. Its just kinda
"playing dead" ... But i can still SSH to the device, so it cant be frozen.

Original comment by marblema...@gmail.com on 8 Mar 2008 at 9:20

GoogleCodeExporter commented 9 years ago 
i dont know what this means but i tried running aptitude while using 
use_compcache.sh
10240, i got an error 

*** glibc detected *** /bin/login: corrupted double-linked list: 0x00026ba8 ***

then it logs me out.
this is the first time ive ever seen this error, using debian-unstable lenny sid
unstable repos's, not testing repos. I should use a stable branch :/

Original comment by marblema...@gmail.com on 8 Mar 2008 at 9:28

GoogleCodeExporter commented 9 years ago 
I downloaded a brand new 2008 rootfs tarball for debian stable etch from
applieddatasystems, im going to just extract it to my SDHC card and boot it to 
test
compcache with. Should be less headache prone

Original comment by marblema...@gmail.com on 8 Mar 2008 at 9:32

GoogleCodeExporter commented 9 years ago 
This has been running alot of stuff for 58 minutes, take a look here at the
screenshot. Its running pretty stable. I was even able to install and run 
tcpdump.
heres my screenshot of my 64mb ram system running with compcache and a 64mb 
swapfile
on the root of the 16GB SDHC card. running debian-eabi lenny/sid 2.6.24-yonggun

Original comment by marblema...@gmail.com on 8 Mar 2008 at 10:12

Attachments:

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
here is an additional screen shot with the output of /proc/compcache and 
/proc/swaps
to prove its running compcache-0.2

Original comment by marblema...@gmail.com on 8 Mar 2008 at 10:18

Attachments:

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
its getting pretty stable, and WAY faster than using swap. WOW!!!

Original comment by marblema...@gmail.com on 8 Mar 2008 at 10:22

GoogleCodeExporter commented 9 years ago 
seems if i stay in a regular terminal and do not start x and do not do any 
apt-get
dist-upgrade or apt-get upgrade it stays stable without locking up at all.. 
heres a
screen shot running for 2 and a half hours so far making sources of linux 
2.6.24

Original comment by marblema...@gmail.com on 8 Mar 2008 at 11:43

Attachments:

GoogleCodeExporter commented 9 years ago 
[deleted comment]
GoogleCodeExporter commented 9 years ago 
i think we should change the subject of this issue to Armel debian x-server 
crashes
with compcache-0.2 because in the terminal its stable. Its just crashing
icewm-session-experimental and icewm-session. By not starting X and using SSH 
through
x on another computer it runs stable it seems as long as i dont run aptitude 
update
or apt-get update or any apt-get string for that matter. heres another screen 
shot of
the machine running for over 3 hours now stable

Original comment by marblema...@gmail.com on 9 Mar 2008 at 12:56

Attachments:

GoogleCodeExporter commented 9 years ago 
still running at 6 hours

Original comment by marblema...@gmail.com on 9 Mar 2008 at 3:24

Attachments:

GoogleCodeExporter commented 9 years ago 
Ill let it keep running for a few days... see what happenes

Original comment by marblema...@gmail.com on 9 Mar 2008 at 3:25

GoogleCodeExporter commented 9 years ago 
heres another screen shot showing the output of /proc/cpuinfo. Id not mind 
seeing
yours just cor curiosity. What arch and cpu types are you using with compcache?

Original comment by marblema...@gmail.com on 9 Mar 2008 at 3:29

Attachments:

GoogleCodeExporter commented 9 years ago 
> What arch and cpu types are you using with compcache?

Here's my /proc/cpuinfo:

processor   : 0
vendor_id   : GenuineIntel
cpu family  : 15
model       : 3
model name  : Intel(R) Pentium(R) 4 CPU 2.80GHz
stepping    : 8
cpu MHz     : 2787.167
cache size  : 1024 KB
fdiv_bug    : no
hlt_bug     : no
f00f_bug    : no
coma_bug    : no
fpu     : yes
fpu_exception   : yes
cpuid level : 5
wp      : yes
flags       : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 
clflush
dts acpi mmx fxsr sse sse2 ss constant_tsc up pebs bts sync_rdtsc pni ds_cpl
bogomips    : 5606.02
clflush size    : 64

This is basically Linux VM(Fedora7) running on VMware Workstation 6.

Original comment by nitingupta910@gmail.com on 9 Mar 2008 at 8:50