Re: paravirtualization status

[Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org>]

Omar Khan wrote:
> hi,
>    What is the status of paravirtualization? 

It's basically stalled.

However, I'd like to restart this, on a much bigger scale.  Using 
paravirt_ops, one can use direct paging (like Xen) instead of shadow 
paging, and achieve better performance.  In fact, it should be faster 
than Xen as, on one hand, syscalls are fast (esp. with 64-bit guests), 
and on the other hand, there's very little mmu overhead.

> Also when Ingo released his
> paravirtualization patch and some results Avi noted that :
>
> "Very impressive!  The gain probably comes not only from avoiding the 
> vmentry/vmexit, but also from avoiding the flushing of the global page 
> tlb entries." [http://thread.gmane.org/gmane.linux.kernel/481084]
>
> can someone please explain briefly what the "global page tlb entries" are? 
>   

When the kernel switches context to another process, it flushes out the 
tlb (a cache of translations of virtual addresses to physical 
addresses).  An optimization is not to flush cached translations of 
kernel virtual addresses, since these are the same for all processes in 
Linux.  Such translations are marked as 'global', indicating to the cpu 
that they need not be flushed on a context switch.


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


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/