3 Commits

Author	SHA1	Message	Date
Roman Lebedev	0727e2b90c	[DAGCombiner][X86][AArch64] Generalize `A-(A&B)`->`A&(~B)` fold (PR44448) ... The fold 'A - (A & (B - 1))' -> 'A & (0 - B)' added in 8dab0a4a7d is too specific. It should/can just be 'A - (A & B)' -> 'A & (~B)' Even if we don't manage to fold `~` into B, we have likely formed `ANDN` node. Also, this way there's less similar-but-duplicate folds. Name: X - (X & Y) -> X & (~Y) %o = and i32 %X, %Y %r = sub i32 %X, %o => %n = xor i32 %Y, -1 %r = and i32 %X, %n https://rise4fun.com/Alive/kOUl See https://bugs.llvm.org/show_bug.cgi?id=44448 https://reviews.llvm.org/D71499	2020-01-03 17:55:47 +03:00
Roman Lebedev	8dab0a4a7d	[DAGCombine][X86][AArch64] 'A - (A & (B - 1))' -> 'A & (0 - B)' fold (PR44448) ... While we do manage to fold integer-typed IR in middle-end, we can't do that for the main motivational case of pointers. There is @llvm.ptrmask() intrinsic which may or may not be helpful, but i'm not sure it is fully considered canonical yet, not everything is fully aware of it likely. https://rise4fun.com/Alive/ZVdp Name: ptr - (ptr & (alignment-1)) -> ptr & (0 - alignment) %mask = add i64 %alignment, -1 %bias = and i64 %ptr, %mask %r = sub i64 %ptr, %bias => %highbitmask = sub i64 0, %alignment %r = and i64 %ptr, %highbitmask See https://bugs.llvm.org/show_bug.cgi?id=44448 https://reviews.llvm.org/D71499	2020-01-03 13:58:36 +03:00
Roman Lebedev	c0cbe3fbb7	[NFC][DAGCombine][X86][AArch64] Tests for 'A - (A & (B - 1))' pattern (PR44448) ... https://rise4fun.com/Alive/ZVdp Name: ptr - (ptr & (alignment-1)) -> ptr & (0 - alignment) %mask = add i64 %alignment, -1 %bias = and i64 %ptr, %mask %r = sub i64 %ptr, %bias => %highbitmask = sub i64 0, %alignment %r = and i64 %ptr, %highbitmask The main motivational pattern involes pointer-typed values, so this transform can't really be done in middle-end. See https://bugs.llvm.org/show_bug.cgi?id=44448 https://reviews.llvm.org/D71499	2020-01-03 13:58:36 +03:00