[mips] Explicitly select `mips32r2` CPU for test cases require 64-bit FPU. NFC

Support for 64-bit coprocessors on a 32-bit architecture
was added in `MIPS32 R2`.

llvm-svn: 365507

[opaque pointer type] Add textual IR support for explicit type parameter to load instruction

Essentially the same as the GEP change in r230786.

A similar migration script can be used to update test

[opaque pointer type] Add textual IR support for explicit type parameter to load instruction

Essentially the same as the GEP change in r230786.

A similar migration script can be used to update test cases, though a few more
test case improvements/changes were required this time around: (r229269-r229278)

import fileinput
import sys
import re

pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)")

for line in sys.stdin:
sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line))

Reviewers: rafael, dexonsmith, grosser

Differential Revision: http://reviews.llvm.org/D7649

llvm-svn: 230794

show more ...

[mips] BSEL's and BINS[RL] operands are reversed compared to the vselect node used in the pattern.

Summary:
Correct the match patterns and the lowerings that made the CodeGen tests pass despite the

[mips] BSEL's and BINS[RL] operands are reversed compared to the vselect node used in the pattern.

Summary:
Correct the match patterns and the lowerings that made the CodeGen tests pass despite the mistakes.

The original testcase that discovered the problem was SingleSource/UnitTests/SignlessType/factor.c in test-suite.
During review, we also found that some of the existing CodeGen tests were incorrect and fixed them:
* bitwise.ll: In bsel_v16i8 the IfSet/IfClear were reversed because bsel and bmnz have different operand orders and the test didn't correctly account for this. bmnz goes 'IfClear, IfSet, CondMask', while bsel goes 'CondMask, IfClear, IfSet'.
* vec.ll: In the cases where a bsel is emitted as a bmnz (they are the same operation with a different input tied to the result) the operands were in the wrong order.
* compare.ll and compare_float.ll: The bsel operand order was correct for a greater-than comparison, but a greater-than comparison instruction doesn't exist. Lowering this operation inverts the condition so the IfSet/IfClear need to be swapped to match.

The differences between BSEL, BMNZ, and BMZ and how they map to/from vselect are rather confusing. I've therefore added a note to MSA.txt to explain this in a single place in addition to the comments that explain each case.

Reviewers: matheusalmeida, jacksprat

Reviewed By: matheusalmeida

Differential Revision: http://llvm-reviews.chandlerc.com/D3028

llvm-svn: 203657

show more ...

[DAGCombiner] Teach how to fold sext/aext/zext of constant build vectors.

This patch teaches the DAGCombiner how to fold a sext/aext/zext dag node when
the operand in input is a build vector of cons

[DAGCombiner] Teach how to fold sext/aext/zext of constant build vectors.

This patch teaches the DAGCombiner how to fold a sext/aext/zext dag node when
the operand in input is a build vector of constants (or UNDEFs).

The inability to fold a sext/zext of a constant build_vector was the root
cause of some pcg bugs affecting vselect expansion on x86-64 with AVX support.

Before this change, the DAGCombiner only knew how to fold a sext/zext/aext of a
ConstantSDNode.

llvm-svn: 200234

show more ...

[mips][msa] Build all the tests in little and big endian modes and correct an incorrect test.

Summary:
This patch (correctly) breaks some MSA tests by exposing the cases when
SelectionDAG::getConsta

[mips][msa] Build all the tests in little and big endian modes and correct an incorrect test.

Summary:
This patch (correctly) breaks some MSA tests by exposing the cases when
SelectionDAG::getConstant() produces illegal types. These have been temporarily
marked XFAIL and the XFAIL flag will be removed when
SelectionDAG::getConstant() is fixed.

There are three categories of failure:
* Immediate instructions are not selected in one endian mode.
* Immediates used in ldi.[bhwd] must be different according to endianness.
(this only affects cases where the 'wrong' ldi is used to load the correct
bitpattern. E.g. (bitcast:v2i64 (build_vector:v4i32 ...)))
* Non-immediate instructions that rely on immediates affected by the
previous two categories as part of their match pattern.
For example, the bset match pattern is the vector equivalent of
'ws | (1 << wt)'.

One test needed correcting to expect different output depending on whether big
or little endian was in use. This test was
test/CodeGen/Mips/msa/basic_operations.ll and experiences the second category
of failure shown above. The little endian version of this test is named
basic_operations_little.ll and will be merged back into basic_operations.ll in
a follow up commit now that FileCheck supports multiple check prefixes.

Reviewers: bkramer, jacksprat, dsanders

Reviewed By: dsanders

CC: llvm-commits

Differential Revision: http://llvm-reviews.chandlerc.com/D1972

llvm-svn: 194806

show more ...

[mips][msa] MSA requires FR=1 mode (64-bit FPU register file). Report fatal error when using it in FR=0 mode.

llvm-svn: 191498

[mips][msa] Remove the VSPLAT and VSPLATD nodes in favour of matching BUILD_VECTOR.

Most constant BUILD_VECTOR's are matched using ComplexPatterns which cover
bitcasted as well as normal vectors. Ho

[mips][msa] Remove the VSPLAT and VSPLATD nodes in favour of matching BUILD_VECTOR.

Most constant BUILD_VECTOR's are matched using ComplexPatterns which cover
bitcasted as well as normal vectors. However, it doesn't seem to be possible to
match ldi.[bhwd] in a type-agnostic manner (e.g. to support the widest range of
immediates, it should be possible to use ldi.b to load v2i64) using TableGen so
ldi.[bhwd] is matched using custom code in MipsSEISelDAGToDAG.cpp

This made the majority of the constant splat BUILD_VECTOR lowering redundant.
The only transformation remaining for constant splats is when an (up-to) 32-bit
constant splat is possible but the value does not fit into a 10-bit signed
integer. In this case, the BUILD_VECTOR is transformed into a bitcasted
BUILD_VECTOR so that fill.[bhw] can be used to splat the vector from a GPR32
register (which is initialized using the usual lui/addui sequence).

There are no additional tests since this is a re-implementation of previous
functionality. The change is intended to make it easier to implement some of
the upcoming instruction selection patches since they can rely on existing
support for BUILD_VECTOR's in the DAGCombiner.

compare_float.ll changed slightly because a BITCAST is no longer
introduced during legalization.

llvm-svn: 191299

show more ...

[mips][msa] Added support for matching max, maxi, min, mini from normal IR (i.e. not intrinsics)

llvm-svn: 191291

[mips][msa] Added support for matching bsel and bseli from normal IR (i.e. not intrinsics)

This required correcting the definition of the bsel and bseli intrinsics.

llvm-svn: 191290

[mips][msa] Added support for matching comparisons from normal IR (i.e. not intrinsics)

MIPS SelectionDAG changes:
* Added VCEQ, VCL[ET]_[SU] nodes to represent vector comparisons that produce a bit

[mips][msa] Added support for matching comparisons from normal IR (i.e. not intrinsics)

MIPS SelectionDAG changes:
* Added VCEQ, VCL[ET]_[SU] nodes to represent vector comparisons that produce a bitmask.

llvm-svn: 191286

show more ...