On 06/16/2011 12:14 PM, Richard Guenther wrote: > On Thu, Jun 16, 2011 at 7:24 PM, Eric Botcazou <ebotca...@adacore.com> wrote: >>> I think this would be clearer with a name like DEPOSIT_EXPR, >>> similar to the ia64 deposit instruction. >> >> ia64's demise wasn't entirely undeserved then. IMO the descriptive power of >> DEPOSIT_EXPR is almost null. BIT_FIELD_MODIFY_EXPR or something like this. > > It's more like BIT_FIELD_COMPOSE_EXPR which is why I chose BIT_FIELD_EXPR, > similar to how we have COMPLEX_EXPR which composes two scalar values. > I don't mind changing the name though, but maybe to BIT_FIELD_COMPOSE_EXPR > then?
That'd be fine. > The expansion code is ad-hoc, I'm not too familiar with what utilities > we have to do a better job here. I'll have a look at store_bit_field > (though that sounds memory-esque). It probably was, originally. But it's got paths through there to handle writing into a register destination, which will wind up in in the various insv and strict_low_part patterns that the md file supports. r~
