On Sun, Oct 20, 2019 at 12:21:21PM +0100, Richard Earnshaw (lists) wrote: > On 19/10/2019 17:31, Segher Boessenkool wrote: > > I have a bunch of testcases from when I did something similar for PowerPC > > that I wanted to test... But I cannot get your series to apply. Do you > > have a git repo I can pull from? > > Perhaps because it's already committed to trunk?
Oh probably. Duh. Thanks :-) > > u64 addH(u64 a) { return a + 0x123400005678ULL; } > > u64 addH0(u64 a) { return a + 0x123400000000ULL; } (If you change those to 0x3400000078ULL etc. it'll test as meant on arm as well: to see if it uses immediates in the insn where it can. It looks like it'll work fine fwiw). > We do pretty well on this. Only addSHm1 needs three insns (except where > the constant isn't valid for arm), and I think that's the minimum for > this case anyway. Several of the tests only need one insn. Yeah, very nice :-) Segher