Re: [PATCH 3/3] xenoprof: Make the escape code consistent across 32 and 64-bit xen

[George Dunlap <george.dunlap@citrix.com>]

On Mon, 2012-01-23 at 17:31 +0000, Keir Fraser wrote:
> > I really prefer to keep patches with no functional change separate from
> > those with a pretty major functional change; even if the major
> > functional change is only one line. :-)
> 
> The one-line major functional change didn't work. Obviously the other
> changes to make it build must be part of the same logical patch.

Perhaps "functional" change isn't what I meant, so much as
"guest-visible behavior change".

If I were writing this patch series from the beginning, I'd have one
patch which "fixed" the various internal functions to use 64-bits for
EIP, since even in 32-bit that value in the struct can be 64-bit.  That
patch by itself could have a bug in it, and so as a reviewer /
archaeologist, *I* would have preferred it did just that one thing, so I
didn't have to figure out what each individual line was supposed to be
doing.  Then the guest-visible functional change would have been its own
patch.

In any case, I'll send a combined patch tomorrow, since that seems
preferred.

 -George