Re: [coLinux-users] Vista SP1 - complete failure

[Henry N. <hen...@ar...>]

Hello,

Josep Maria Antolín Segura schrieb:
> commandline:
> colinux-daemon.exe -t nt @D:\Colinux\debian\ramdisk-mode.txt
> 
> ramdisk-mode.txt:
> kernel=vmlinux
> initrd=initrd.gz
> root=/dev/ram0
> 
> colinux version:
> Linux (none) 2.6.22.18-co-0.7.3 #1 PREEMPT Sat May 24 22:27:30 UTC 2008 i686 unk
> nown
> 
> 
> I don't know if it may be related to being Vista Business, or the AMD chip. I
> believe you'll have to work that out with the developers via debug builds.
> 

thanks, Josep.

The version 0.8.0 with kernel 2.6.17 is a non official build, please use
this not as reference.

The version 20080511-ntconsoleflood is a special build with some more
add ons, and have more debugging inside. So, this version would run
slower. You have seen it as a longer time before problems starts.

Jonathan, debugging is available, if you would try some steps. First
lets configure the simplest start. Perhaps Josep can check, that the
same variant is running without probmlems.

0. All the colinux programs, please run from a CMD.EXE you have opened
as Admin. Please don't use "start.exe" and not a tool "runas.exe".
Please don't think you are an user with admin rights. That is not the
same. Check your admin status with "whoamy /all" or so.

1. You should have installed 0.7.3 final version. We need to verify
exactly the same version. Check the installed driver:
C:\coLinux> colinux-daemon --status-driver
Cooperative Linux Daemon, 0.7.3
Daemon compiled on Sat May 24 22:36:07 2008

checking if the driver is installed
current state: 4 (fully initialized)
current number of monitors: 0
current linux api version: 10
current periphery api version: 20
Driver compiled on: Sat May 24 22:36:07 2008

The installed path can differ. Any different output would make problems.

2. Open your Windows task manager and watch the performance graph.
Splitte the cpu into 2 diagrams to see, if one core is hanging.

Some times later as you sayed, the first cpu core (the left diagram)
would go into 100% of load after coLinux runs somethere in an endless
loop. We need to find the position in Linux kernel boot steps.

3. Read file debugging.txt from your installation, or get it from
http://colinux.svn.sourceforge.net/svnroot/colinux/branches/devel/doc/debugging

4. Open a command prompt, change into colinux directory and run:
C:\coLinux> mode con lines=600 cols=120
C:\coLinux> colinux-debug-daemon.exe -p -d -s prints=31,misc=31
Adjust the font size from sys-menu of this window to see most lines as
you can. Don't worry about XML output needs only to see the text messages.

5. Watch the taskmanager performance graph. This command should not
incrace the cpu load.

6. Run the minimal coLinux ramdisk with special parameters:
colinux-daemon mem=64 kernel=vmlinux initrd=initrd.gz root=/dev/ram0
rootdelay=10

Any signifikant changes in the memory or performance? 10 to 20% more cpu
load is normal after kernel boot, 5-10 MB more memory is normal.

7. Wait for prompt. Do nothing, do not press ENTER before 10 minutes you
have seen not your problem. The colinux-debug output should stop on this
step.

8. Press ENTER and run "halt" to shutdown coLinux. The debugger should
stop with text "string" lines like this:
   <string>prints "colinux: halted, exiting."
   <string>shutting down</string>
   ...
   <string>Pipe broken, exiting</string>
   <string>done freeing</string>
   <string>after free: 0 blocks</string>
   <string>daemon cleanup</string>
   <string>rc=00424400 exit=0</string>

Please copy and send your last lines, if you have more or others.

In every step:
Watch also for NT kernel memory usage after every step, and what the
physical free memory. Would it incrace permanently after one of the steps?

Plese give some coments for the steps. We can rune more debugs later.

-- 
Henry N.