Success!

Success!

[Ken Causey]

I just wanted to say thanks to everyone for all of their work and for
the help I've recently received from the list.  I especially want to
thank Neil Brown who has provided significant help.  I finally finished
setting up and testing a Debian testing SCSI software RAID system which
utilizes 3 drives, RAID1 and RAID5, and now seems to be bullet proof.  I
can lose any single drive and the system runs and boots happily.  Soon I
will be building another system very similar but with IDE.  After that I
might put together a little howto since I was never found one that had
all the info I needed.  Of course even if mine is perfect it will be out
of date in 3 to 6 months, but anyway...

Thanks!

Ken Causey

Success!

[Gregory Haskins]

[-- Attachment #1: Type: text/plain, Size: 1353 bytes --]


Im pleased to report that with the addition of the PIIX3 fix I posted earlier coupled with the reversion of 4423, we can now put SUSE Linux Enterprise Desktop on the list of booting guests.  This is verified with 64 bit on 64 bit on Intel only.  The GRUB segfault that I was seeing in 12/13 is fixed also once moving to the KVM-16 codebase, so the whole thing is working appropriately now.  Thanks for all the help guys!  

Attached, please find a tarball and a quilt series on the patches we are using to run this on SLEx (I excluded the patches obtained from the tree).  I dont expect most of them to be adopted into the tree.  I am just sharing them in case someone else wants to run on SUSE or has similar problems.   One of the patches adds a rudimentary trace facility that printks VMEXIT activity (mostly only supporting Intel, however).  It also has an optional mode that allows you to turn on single-step breakpoints so that you can watch the RIP for each op-code.  Note however that turning single-step on seems to change the behavior (takes unexpected triplefaults) so its not fully baked.  The non-singlestep mode was quite helpful in tracking down the real-mode problems I was having with the 4423 BIOS.

We will have to figure out what the problem in the BIOS is w.r.t. 4423 before this will work with the code in /trunk, however.

-Greg

[-- Attachment #2: patches.tar.bz2 --]
[-- Type: application/octet-stream, Size: 51200 bytes --]

[-- Attachment #3: Type: text/plain, Size: 345 bytes --]

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

[-- Attachment #4: Type: text/plain, Size: 186 bytes --]

_______________________________________________
kvm-devel mailing list
kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/kvm-devel

Re: Success!

[Avi Kivity]

Gregory Haskins wrote:
> Im pleased to report that with the addition of the PIIX3 fix I posted earlier coupled with the reversion of 4423, we can now put SUSE Linux Enterprise Desktop on the list of booting guests.  This is verified with 64 bit on 64 bit on Intel only.  The GRUB segfault that I was seeing in 12/13 is fixed also once moving to the KVM-16 codebase, so the whole thing is working appropriately now.  Thanks for all the help guys!  
>   

Great that it's working!  It's nice to have more Linux distros as guests.

> Attached, please find a tarball and a quilt series on the patches we are using to run this on SLEx (I excluded the patches obtained from the tree).  I dont expect most of them to be adopted into the tree.  I am just sharing them in case someone else wants to run on SUSE or has similar problems.   One of the patches adds a rudimentary trace facility that printks VMEXIT activity (mostly only supporting Intel, however).  It also has an optional mode that allows you to turn on single-step breakpoints so that you can watch the RIP for each op-code.  Note however that turning single-step on seems to change the behavior (takes unexpected triplefaults) so its not fully baked.  The non-singlestep mode was quite helpful in tracking down the real-mode problems I was having with the 4423 BIOS.
>
>   

There is support for singlestep and hardware breakpoints as a gdbserver 
stub; you can debug vmlinux in a guest remotely.  It's been a long while 
since I last tested it, so maybe it needs some attention.


> We will have to figure out what the problem in the BIOS is w.r.t. 4423 before this will work with the code in /trunk, however.
>   

I'll take a look.


-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Success!

[Avi Kivity]

Avi Kivity wrote:
>   
>> We will have to figure out what the problem in the BIOS is w.r.t. 4423 before this will work with the code in /trunk, however.
>>   
>>     
>
> I'll take a look.
>   

I went through a Windows XP install cycle; it worked but no acpi.  I 
don't know what fixed this; maybe it was the non-pae fix.

Anyway I've reverted r4423; let's pretend it never happened.

-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: Success!

[Avi Kivity]

Avi Kivity wrote:
> Avi Kivity wrote:
>>  
>>> We will have to figure out what the problem in the BIOS is w.r.t. 
>>> 4423 before this will work with the code in /trunk, however.
>>>       
>>
>> I'll take a look.
>>   
>
> I went through a Windows XP install cycle; it worked but no acpi.  I 
> don't know what fixed this; maybe it was the non-pae fix.
>
> Anyway I've reverted r4423; let's pretend it never happened.
>

Actually that reverted the bios to qemu-0.8.2. Silly me.  Anyway I 
hacked around it and now Linux boots with acpi and ide dma.  I'll try 
Windows too.

-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV