2008/3/11, Zdenek Kabelac <[EMAIL PROTECTED]>:
> 2008/3/9, Zdenek Kabelac <[EMAIL PROTECTED]>:
>
> > 2008/3/7, Zdenek Kabelac <[EMAIL PROTECTED]>:
>  >
>  > > 2008/3/5, Zdenek Kabelac <[EMAIL PROTECTED]>:
>  >  >
>  >  > > 2008/3/5, Avi Kivity <[EMAIL PROTECTED]>:
>  >  >  >
>  >  >  > > Andi Kleen wrote:
>  >  >  >  >  > Avi Kivity <[EMAIL PROTECTED]> writes:
>  >  >  >  >  >
>  >  >  >  >  >> Most likely movs emulation is broken for long counts.  Please 
> post a
>  >  >  >  >  >> disassembly of copy_user_generic_string to make sure we're 
> looking at
>  >  >  >  >  >> the same code.
>  >  >  >  >  >>
>  >  >  >  >  >
>  >  >  >  >  > Be careful -- this code is patched at runtime and what you
>  >  >  >  >  > see in the vmlinux is not necessarily the same that is executed
>  >  >  >  >  >
>  >  >  >  >  >
>  >  >  >  >
>  >  >  >  >
>  >  >  >  > If the disassembled instruction isn't marked as an alternative in 
> the
>  >  >  >  >  source, then it can't be patched, right?
>  >  >  >  >
>  >  >
>  >  >
>  >  > Hello
>  >  >
>  >  >  Any progress on this - It looks like I get this bug quite often when I 
> test
>  >  >  device-mapper code.
>  >  >
>  >
>  >
>  >
>  > Hello
>  >
>  >  I've made some more experiments and noticed few more things:
>  >
>  >  a) - it is just enough to run parallel loop with cat LVM partition
>  >  >/dev/null and dmsetup status
>  >
>  >  b) when I insert for() loop for zeroing allocated memory in the
>  >  dm-ioctl copy_params function my loop start once the memory crosses
>  >  exactly 4KB boundary (visible from register content)
>  >
>  >  c) in my trace log I could usually always see this pattern:
>  >  [  160.634897]  [<ffffffff812ee5ba>] preempt_schedule_irq+0x5a/0xa0
>  >  [  160.634897]  [<ffffffff8100cf46>] retint_kernel+0x26/0x30
>  >
>  >  from the look in the arch/x86/kernel/entry64.s I could really see
>  >  there is some potentiality for internal loop that may call
>  >  preempt_schedule_irq in upon some check in  exit_intr - but having
>  >  actually now idea what's this all about...
>  >
>  >  I've put there just some extra dump_stack trace in the
>  >  preempt_schedule_irq - and it's really being printed - but quite
>  >  slowly actually considering process eats 100% CPU
>  >
>  >  So anyone has any idea what might be wrong ?
>
>
>
> Hello
>
>  I've some more news here - it looks I've found working setup on my C2D.
>
>  All I need to do is compile my 64bit kernel with optimization for space.
>  This will magical start to work - at least in this case.
>
>  I'll now probably slowly try to figure out which directory with -Os
>  compilation makes the difference.
>
>  Also I've noticed that standard Debian 2.6.24-4-686 kernel loops in
>  Qemu, but 486 version doesn't.

Argh - being stupid here - it looks like these 'working' kernels were
not SMP actually.
As long as the SMP is used - I still get the busy loop :(
Now being clueless....

Zdenek

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
kvm-devel mailing list
kvm-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/kvm-devel

Reply via email to