It seems TypeSize is currently broken in the sense that:
TypeSize::Fixed(4) + TypeSize::Scalable(4) => TypeSize::Fixed(8)
without failing its assert that explicitly tests for this case:
assert(LHS.Scalable == RHS.Scalable && ...);
The reason this fails is that `Scalable` is a static method of class
TypeSize,
and LHS and RHS are both objects of class TypeSize. So this is
evaluating
if the pointer to the function Scalable == the pointer to the function
Scalable,
which is always true because LHS and RHS have the same class.
This patch fixes the issue by renaming `TypeSize::Scalable` ->
`TypeSize::getScalable`, as well as `TypeSize::Fixed` to
`TypeSize::getFixed`,
so that it no longer clashes with the variable in
FixedOrScalableQuantity.
The new methods now also better match the coding standard, which
specifies that:
* Variable names should be nouns (as they represent state)
* Function names should be verb phrases (as they represent actions)
When x is not known to be nonzero, ctpop(x) == 1 is expanded to
x != 0 && (x & (x - 1)) == 0
resulting in codegen like
leal -1(%rdi), %eax
testl %eax, %edi
sete %cl
testl %edi, %edi
setne %al
andb %cl, %al
But another expression that works is
(x ^ (x - 1)) > x - 1
which has nicer codegen:
leal -1(%rdi), %eax
xorl %eax, %edi
cmpl %eax, %edi
seta %al
This patch lowers `sdiv x, +/-2**k` to `add + select + shift` when the
short forward branch optimization is enabled. The latter inst seq
performs faster than the seq generated by target-independent
DAGCombiner. This algorithm is described in ***Hacker's Delight***.
This patch also removes duplicate logic in the X86 and AArch64 backend.
But we cannot do this for the PowerPC backend since it generates a
special instruction `addze`.
This adds the nneg flag to SDNodeFlags and the node printing code.
SelectionDAGBuilder will add this flag to the node if the target doesn't
prefer sign extend.
A future RISC-V patch can remove the sign extend preference from
SelectionDAGBuilder.
I've also added the flag to the DAG combine that converts
ISD::SIGN_EXTEND to ISD::ZERO_EXTEND.
D146121 needs to set the NSW flag, but given the result is NUW then we know that the result has leading zeros, so we don't need to call ComputeNumSignBits - just reuse the existing KnownBits value instead.
If a shl node leaves the upper half bits zero / undemanded, then see if we can profitably perform this with a half-width shl and a free trunc/zext.
Followup to D146121
Reapplied - moved after the ShrinkDemandedOp call; reuse the existing KnownBits result; ensure that we only attempt this if all the upper bits are demanded; 547dc461225ba should address the remaining regressions that were noticed in the previous commit.
Differential Revision: https://reviews.llvm.org/D155472
If a shl node leaves the upper half bits zero / undemanded, then see if we can profitably perform this with a half-width shl and a free trunc/zext.
Followup to D146121
Differential Revision: https://reviews.llvm.org/D155472
Splitting up patches for #20571. I found these comments generally useful
to add and not predicated on those changes. Hopefully they help future
travelers.
I accidentally introduced this in
commit 330fa7d2a4e0 ("[TargetLowering] Deduplicate choosing InlineAsm
constraint between ISels (#67057)")
Fix forward.
After 330fa7d2a4e0cfbb4b078 we were seeing nondeterministic failures of
llvm/test/CodeGen/ARM/thumb-big-stack.ll, with different code being
generated in different runs.
Switching sort -> stable_sort fixes this.
It looks like the old algorithm picked the first best option, and using
stable_sort restores that behavior.
Given a list of constraints for InlineAsm (ex. "imr") I'm looking to
modify the order in which they are chosen. Before doing so, I noticed a
fair
amount of logic is duplicated between SelectionDAGISel and GlobalISel
for this.
That is because SelectionDAGISel is also trying to lower immediates
during selection. If we detangle these concerns into:
1. choose the preferred constraint
2. attempt to lower that constraint
Then we can slide down the list of constraints until we find one that
can be lowered. That allows the implementation to be shared between
instruction selection frameworks.
This makes it so that later I might only need to adjust the priority of
constraints in one place, and have both selectors behave the same.
While simplifying some vector operators in DAG combine, we may need to
create new instructions for simplified vectors. At that time, we need to
make sure that all the flags of the new instruction are copied/modified
from the old instruction.
If "contract" is dropped from an instruction like FMUL, it may not
generate FMA instruction which would impact performance.
Here's an example where "contract" flag is dropped when FMUL is created.
Replacing.2 t42: v2f32 = fmul contract t41, t38
With: t48: v2f32 = fmul t38, t38
Co-authored-by: Sirish Pande <sirish.pande@amd.com>
If the SRL for Hi constant folds, but we don't remoe those bits from
the Lo, we can end up with strange constant folding through DAGCombine later.
I've only seen this with constants being lowered to constant pools
during lowering on RISC-V.
This patch adjusts the legality check for riscv to use `cpop/cpopw` since `isOperationLegal(ISD::CTPOP, MVT::i32)` returns false on rv64gc_zbb.
Clang vs gcc: https://godbolt.org/z/rc3s4hjPh
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D156390
Extension to the signbit case, if the signbits extend down through all the demanded bits then SMIN/SMAX/UMIN/UMAX nodes can be simplified to a OR/AND/AND/OR.
Alive2: https://alive2.llvm.org/ce/z/mFVFAn (general case)
Differential Revision: https://reviews.llvm.org/D158364
The CodeView `S_ARMSWITCHTABLE` debug symbol is used to describe the layout of a jump table, it contains the following information:
* The address of the branch instruction that uses the jump table.
* The address of the jump table.
* The "base" address that the values in the jump table are relative to.
* The type of each entry (absolute pointer, a relative integer, a relative integer that is shifted).
Together this information can be used by debuggers and binary analysis tools to understand what an jump table indirect branch is doing and where it might jump to.
Documentation for the symbol can be found in the Microsoft PDB library dumper: 0fe89a942f/cvdump/dumpsym7.cpp (L5518)
This change adds support to LLVM to emit the `S_ARMSWITCHTABLE` debug symbol as well as to dump it out (for testing purposes).
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D149367
This improves some cases where a splat_vector uses a build_pair that can be
simplified, e.g:
(rotl x:i64, splat_vector (build_pair x1:i32, x2:i32))
rotl only demands the bottom 6 bits, so this patch allows it to simplify it to:
(rotl x:i64, splat_vector (build_pair x1:i32, undef:i32))
Which in turn improves some cases where a splat_vector_parts is lowered on
RV32.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D158839
This reverts commit 8d0c3db388143f4e058b5f513a70fd5d089d51c3.
Causes crashes, see comments in https://reviews.llvm.org/D149367.
Some follow-up fixes are also reverted:
This reverts commit 636269f4fca44693bfd787b0a37bb0328ffcc085.
This reverts commit 5966079cf4d4de0285004eef051784d0d9f7a3a6.
This reverts commit e7294dbc85d24a08c716d9babbe7f68390cf219b.
The CodeView `S_ARMSWITCHTABLE` debug symbol is used to describe the layout of a jump table, it contains the following information:
* The address of the branch instruction that uses the jump table.
* The address of the jump table.
* The "base" address that the values in the jump table are relative to.
* The type of each entry (absolute pointer, a relative integer, a relative integer that is shifted).
Together this information can be used by debuggers and binary analysis tools to understand what an jump table indirect branch is doing and where it might jump to.
Documentation for the symbol can be found in the Microsoft PDB library dumper: 0fe89a942f/cvdump/dumpsym7.cpp (L5518)
This change adds support to LLVM to emit the `S_ARMSWITCHTABLE` debug symbol as well as to dump it out (for testing purposes).
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D149367
DemandedBits is forced to all ones if there are multiple users.
The changes X86 test cases looks like they were miscompiles before.
The value of eax/rax from the cmov is returned from the function in
addition to being used by the sar. That usage needs all bits even
though the sar doesn't.
This reverts commit 54d663d5896008c09c938f80357e2a056454bc65, which breaks the test CodeGen/SystemZ/ctpop-01.ll for stage2-ubsan check (see https://lab.llvm.org/buildbot/#/builders/85/builds/18410)
I manually confirmed that the test had been passing immediately prior to that commit
(BUILDBOT_REVISION=4772c66cfb00d60f8f687930e9dd3aa1b6872228 llvm-zorg/zorg/buildbot/builders/sanitizers/buildbot_bootstrap_ubsan.sh)
On X86 for vec types `(X & Y) == Y` is generally preferable to
`(X & Y) != 0`. Creating zero requires an extra instruction and on
pre-avx512 targets there is no vector `pcmpne` so it requires two
additional instructions to invert the `pcmpeq`.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D157014
Move the sra(shl(x,c1),c1) -> sign_extend_inreg(x) fold inside SimplifyDemandedBits so we can recognize hidden splats with DemandedElts masks.
Because the c1 shift amount has multiple uses, hidden splats won't get simplified to a splat constant buildvector - meaning the existing fold in DAGCombiner::visitSRA can't fire as it won't see a uniform shift amount.
I also needed to add TLI preferSextInRegOfTruncate hook to help keep truncate(sign_extend_inreg(x)) vector patterns on X86 so we can use PACKSS more efficiently.
Differential Revision: https://reviews.llvm.org/D157972
Extending to f32 first (as is done for f16) results in better generated
code for RISC-V (and affects no other in-tree tests). Additionally,
performing the FP_EXTEND first seems equally justified for bf16 as for
f16.
Differential Revision: https://reviews.llvm.org/D156944
We only attempted to determine KnownBits for uniform constant shift amounts, but ComputeKnownBits is able to handle some non-uniform cases as well that we can use as a fallback.
Inspired by some of the cases from D145468
Let SimplifyDemandedBits handle the narrowing of lshr to half-width if we don't require the upper bits, the narrowed shift is profitable and the zext/trunc are free.
A future patch will propose the equivalent shl narrowing combine.
Differential Revision: https://reviews.llvm.org/D146121
TargetLowering::SimplifySetCC wants to swap the operands of a SETCC to
canonicalize the constant to the RHS. The bug here was that it did so whether
or not the RHS was already a constant, leading to an infinite loop.
rdar://111847838
Divverential revision: https://reviews.llvm.org/D155095
This reverts commit cdc633e4bc93d4bf241ecd4c29691ae065749313.
TargetLowering::SimplifySetCC wants to swap the operands of a SETCC to
canonicalize the constant to the RHS. The bug here was that it did so whether
or not the RHS was already a constant, leading to an infinite loop.
rdar://111847838
Differential revision: https://reviews.llvm.org/D155095
InstCombine should have taken care of this, but I think
this is more useful in the future when the expansion
tries to handle multiple cases at a time with fcmp.
x87 looks worse to me but the only thing I know about it is that
I aggressively do not care about it.
https://reviews.llvm.org/D143198
