On 5/28/2026 3:46 AM, Robin Dapp wrote:
I was thinking of "scaling" dependent filters right away, i.e. allowing a list
of dependent opnos and passing the current operands[] to the backend. I
wouldn't dare though :)
Initially, before starting the implementation, my assumption was that adjusting
or adding early-clobber semantics would likely not be well received and that's
why I went the fully-generic route. If new earlyclobbers kinds are acceptable,
maybe that would be a less invasive way?
It really never even occurred to me to consider a new earlyclobber like
constraint. I think I kept zeroing in on our inability to look at more
than just one operand to determine validity.
If we can write the rules in a way that's easy to understand and doing
so allows us to determine validity of an operand without additional
state (like the registers used in a different operand), then that's a
win in my book and better than the filtering approach I suspect.
jeff
