Re: [PATCH] Tree SRA and atomicity/volatility

["Richard Guenther" <richard.guenther@gmail.com>]

On 1/24/07, Richard Kenner <kenner@vlsi1.ultra.nyu.edu> wrote:
> > We don't have defined what TYPE_VOLATILE on a structure type means.
>
> It means exactly the same as defined everywhere else: do as little as
> possible (in a pragmatic sense) in changing the accesses to objects of
> that type from the way they were originally written.
>
> > And it certainly won't work the way the Ada frontend seems to thing for
> > random structures and/or targets.  I'm just requesting that we define what
> > it means and that the Ada frontend maybe re-considers the way it
> > want's to guarantee its 'atomicity' constraints on whole-aggregate access.
>
> If there's a problem here, please be specific as to what it is because
> I'm having trouble understanding this point.

I would suggest specifying TYPE_VOLATILE on a structure type as matching
behavior with TYPE_VOLATILE on a bitfield with the same machine mode.
So, if you can set TYPE_MODE of the structure to SImode (as in Erics original
example) volatile on the strucutre should behave as if it were an SImode
bitfield (or integer).  So if you can only use BLKmode on the
structure it's either
invalid or undefined what happens.

This would of course force more "whole aggregate" access than appearantly
required by the language, for example forcing read-modify-write cycles for
structure component access rather than just forcing whole aggregate access
for whole aggregate copy (which doesn't sound too useful on its own).

Richard.