[Libre-soc-bugs] [Bug 558] gcc SV intrinsics concept

bugzilla-daemon at libre-soc.org bugzilla-daemon at libre-soc.org
Mon Jan 11 23:06:17 GMT 2021


--- Comment #43 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Jacob Lifshay from comment #41)
> (In reply to Jacob Lifshay from comment #40)

> > the cr logic ops and the cr from int and cr to int ops are generated
> > automatically by the instruction selector and by the register allocator.
> this should be relatively easy since all you need to do is add the proper
> instructions for the generic register move ops generated by the register
> allocator:
> move bitvector from int to cr
> move bitvector from cr to int

if the CR masks are striped down all the eq bits, all the lt bits etc. this
meshes well.  the register allocator becomes a weird bitlevel allocator but

i would very much prefer that interactions between the cr mask intrinsics *not*
require tight integration and understanding in gcc other than in the regfile
allocator (keep gcc off the critical path)

that implies keeping *away* from CR0-7 as much as possible, so that use of a
crweird intrinsic doesn't completely destabilise scalar code.  once we have
bootstrapped up, got VC funding, this can be revisited.

You are receiving this mail because:
You are on the CC list for the bug.

More information about the libre-soc-bugs mailing list