Commit Graph

8 Commits

Author SHA1 Message Date
Roman Lebedev ccdad6ef48 [InstCombine] foldShiftIntoShiftInAnotherHandOfAndInICmp(): avoid constantexpr pitfail (PR42962)
Instead of matching value and then blindly casting to BinaryOperator
just to get the opcode, just match instruction and do no cast.

Fixes https://bugs.llvm.org/show_bug.cgi?id=42962

llvm-svn: 368554
2019-08-12 11:28:02 +00:00
Roman Lebedev a8d20b4467 [InstCombine] Shift amount reassociation in bittest: relax one-use check when shifting constant
If one of the values being shifted is a constant, since the new shift
amount is known-constant, the new shift will end up being constant-folded
so, we don't need that one-use restriction then.

llvm-svn: 368519
2019-08-10 19:28:54 +00:00
Roman Lebedev 64fe806c4e [InstCombine] Shift amount reassociation in bittest: drop pointless one-use restriction
That one-use restriction is not needed for correctness - we have already
ensured that one of the shifts will go away, so we know we won't increase
the instruction count. So there is no need for that restriction.

llvm-svn: 368518
2019-08-10 19:28:44 +00:00
Roman Lebedev 45e9990c02 [NFC][InstCombine] Tests for shift amount reassociation in bittest with shift of const
llvm-svn: 368517
2019-08-10 19:28:12 +00:00
Roman Lebedev 72b8d41ce8 [InstCombine] Shift amount reassociation in bittest (PR42399)
Summary:
Given pattern:
`icmp eq/ne (and ((x shift Q), (y oppositeshift K))), 0`
we should move shifts to the same hand of 'and', i.e. rewrite as
`icmp eq/ne (and (x shift (Q+K)), y), 0`  iff `(Q+K) u< bitwidth(x)`

It might be tempting to not restrict this to situations where we know
we'd fold two shifts together, but i'm not sure what rules should there be
to avoid endless combine loops.

We pick the same shift that was originally used to shift the variable we picked to shift:
https://rise4fun.com/Alive/6x1v

Should fix [[ https://bugs.llvm.org/show_bug.cgi?id=42399 | PR42399]].

Reviewers: spatel, nikic, RKSimon

Reviewed By: spatel

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63829

llvm-svn: 364791
2019-07-01 15:55:15 +00:00
Roman Lebedev 34a0b16e29 [NFC][InstCombine] Better commutative tests for "shift amount reassociation in bittest" pattern.
As discussed in https://reviews.llvm.org/D63829
*if* *both* shifts are one-use, we'd most likely want to produce `lshr`,
and not rely on ordering.

Also, there should likely be a *separate* fold to do this reordering.

llvm-svn: 364772
2019-07-01 14:28:24 +00:00
Roman Lebedev 3f3eacfec1 [NFC][InstCombine] Revisit one-use tests in shift-amount-reassociation-in-bittest.ll
llvm-svn: 364433
2019-06-26 14:42:39 +00:00
Roman Lebedev 78edfc4bf0 [NFC][InstCombine] Add shift amount reassociation in bittest tests (PR42399)
https://bugs.llvm.org/show_bug.cgi?id=42399
https://rise4fun.com/Alive/kBb
https://rise4fun.com/Alive/1SB

llvm-svn: 364430
2019-06-26 14:24:41 +00:00