[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH] inline VMX functions


  • To: "Li, Xin B" <xin.b.li@xxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: Keir Fraser <Keir.Fraser@xxxxxxxxxxxx>
  • Date: Mon, 13 Nov 2006 08:22:21 +0000
  • Delivery-date: Mon, 13 Nov 2006 00:22:33 -0800
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AccGzGxsIIVOm6GYQlqxIj8TgPNHxgAMGp1d
  • Thread-topic: [Xen-devel] [PATCH] inline VMX functions

On 13/11/06 2:35 am, "Li, Xin B" <xin.b.li@xxxxxxxxx> wrote:

> inline VMX functions since most of these functions are called just once.
> 
> Signed-off-by: Xin Li <xin.b.li@xxxxxxxxx>

Unless the function call is really hot (in an inner loop for example) I
don't believe that the performance benefit of inlining is measurable (or
maybe even exists at all). Against that there are the disadvantages that
debugging from a backtrace becomes much more of a pain (especially that
vmexit handler, which becomes *really* big!), and there's a tendency for
'inline' to spread around like a 'go faster' magic dust, applied to
functions where it really doesn't help. So I'm moderately inclined not to
take the inlining bits of this patch unless you strongly disagree.

 -- Keir



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.