On 7/4/25 1:43 PM, Vineet Gupta wrote:
Hmm. I'm a little surprised that's needed. But there's new rules in
the era of LRA and define_constraint.
Yeah I was wondering why the fallback reg alternative is not already built-in.
Is that meant to catch bugs in the backends ?
I doubt it's meant to catch bugs in the backend. More likely a need to
clearly guide LRA that when all else fails the value can be reloaded
into a REG and everything will just work. Otherwise LRA has to look at
the constraint and somehow magically conclude that a REG is always safe.
Not 100% sure how that would work in practice.
I briefly thought a match_code clause might help with that problem, but
it didn't make any difference.
jeff
