Re: [PATCH 6/7] xen/guest_access: Consolidate guest access helpers in xen/guest_access.h

[Ian Jackson <ian.jackson@xxxxxxxxxx>]

Hi.  My attention was drawn to this thread.

As I understand it, everyone is agreed that deduplicating the
implementation is good (I also agree).  The debate is only between:

1. Put it in xen/ until an arch comes along that needs something
  different, at which point maybe introduce an asm-generic-style
  thing with default implementations.

2. Say, now, that this is a default implementation and it should go in
   asm-generic.

My starting point is that Julien, as the primary author of this
cleanup, should be given leeway on a matter of taste like this.
(There are as I understand it no wider implications.)

Also, ISTM that it can be argued that introducing a new abstraction is
an additional piece of work.  Doing that is certainly not hampered by
Julien's change.  So that would be another reason to take Julien's
patch as-is.

On the merits, I don't have anything to add to the arguments already
presented.  I am considerably more persuaded by Julien's arguments
than Jan's.

So on all levels I think this commit should go in, unless there are
other concerns that have not been discussed here ?

Thanks,
Ian.