Re: [coLinux-users] CoLinux does not start

[Jurgis <ju...@em...>]

Henry,

Thank you for such a thorough descriptions of all the steps!

Here are the results:
http://jurgis.emails.lv/colinux/02/

Jurgis

Henry Nestler wrote:
> Am 22.03.2009 23:15, schrieb Henry Nestler:
>> Am 22.03.2009 08:16, schrieb Jurgis:
>>> Henry,
>>>
>>> Thank you for the provided information!
>>>
>>> I have run following commands and attached screenshots and debug.xml 
>>> file.
>>>    >   colinux-daemon kernel=vmlinux initrd=initrd.gz root=/dev/ram0 
>>> -v 1
>>>    >   colinux-debug-daemon.exe -d -p -s prints=31,misc=31 -f debug.xml
>>>
>>> I'm still using devel-coLinux-20090319.exe
>>>
>>> When running colinux-daemon.exe it does not end normally, I have to 
>>> click
>>> on the cross at the right upper corner to close it.
>>
>> Sorry, the debug have not more helpfully informations.
>>
>> I think, it's a problem with timer. The next boot step after
>> "pidhash_init" (your last line) will setup the timer source for coLinux
>> kernel. That is very hard to debug. I feel this PC has one of the
>> special timer chip, that we don't handle right.
>>
>> Please view the older mail:
>> http://www.nabble.com/forum/ViewPost.jtp?post=18019511&framed=y
>>
>> Boot with a Live CD, for example Knoppix or Gparted and watch for such
>> kernel boot message: "MP-BIOS bug: 8254 timer not connected to IO-APIC"
>>
>
> It would be nice, if you can send me kernel boot messages from such 
> live cd.
> Run "dmesg >/mesg.txt" to store the messages in file, and please the 
> list of interrupts under this native Linux ("cat /proc/interrupts 
> >interrupts.txt")
>
> And under Windows please send me the list of interrupts. You will find 
> this by running
> "C:\Program files\Shared files\Microsoft Shared\MsInfo\msinfo32.exe" 
> and selecting Hardwareresource IRQs. The Directory can be localized in 
> your language.
>
> To solve the problem we need to locatisate the stopping function 
> inside coLinux kernel. But this needs many more developing time and 
> special builds with more outputs between the steps to find the 
> stopper. I don't know how we can do this over a mail distance. That 
> would cost time for your testing and my coding.
>