Previous code build the model that tile config register is the user of
each AMX instruction. There is a problem for the tile config register
spill. When across function, the ldtilecfg instruction may be inserted
on each AMX instruction which use tile config register. This cause all
tile data register clobber.
To fix this issue, we remove the model of tile config register. We
analyze the regmask of call instruction and insert ldtilecfg if there is
any tile data register live across the call. Inserting the sttilecfg
before the call is unneccessary, because the tile config doesn't change
and we can just reload the config.
Besides we also need check tile config register interference. Since we
don't model the config register we should check interference from the
ldtilecfg to each tile data register def.
ldtilecfg
/ \
BB1 BB2
/ \
call BB3
/ \
%1=tileload %2=tilezero
We can start from the instruction of each tile def, and backward to
ldtilecfg. If there is any call instruction, and tile data register is
not preserved, we should insert ldtilecfg after the call instruction.
Differential Revision: https://reviews.llvm.org/D94155
In case of indirect calls or address taken functions,
skip propagating any attributes to them. We just
propagate features to such functions.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D94585
We already have an experimental option to tune loop alignment. Its impact
is very wide (and there is a suspicion that it's not always profitable). We want
to have something more narrow to play with. This patch adds similar option that
overrides preferred alignment for innermost loops. This is for experimental
purposes, default values do not change the existing behavior.
Differential Revision: https://reviews.llvm.org/D94895
Reviewed By: pengfei
Previously we only matched (and (shl X, C1), 0xffffffff << C1)
which matches the InstCombine canonicalization order. But its
possible to see (shl (and X, 0xffffffff), C1) if the pattern
is introduced in SelectionDAG. For example, through expansion of
a GEP.
It caused "Vector shift amounts must be in the same as their first arg"
asserts in Chromium builds. See the code review for repro instructions.
> Add DemandedElts support inside the TRUNCATE analysis.
>
> Differential Revision: https://reviews.llvm.org/D56387
This reverts commit cad4275d69.
We already handle "vperm2x128 (ins ?, X, C1), (ins ?, X, C1), 0x31" for shuffling of the upper subvectors, but we weren't dealing with the case when we were splatting the upper subvector from a single source.
Exploits the instruction xxsplti32dx.
It can be used to materialize any 64 bit scalar/vector splat by using two instances, one for the upper 32 bits and the other for the lower 32 bits. It should not materialize the cases which can be materialized by using the instruction xxspltidp.
Differential Revision: https://https://reviews.llvm.org/D90173
There can be muliple patterns that map to the same compressed
instruction. Reversing those leads to multiple ways to uncompress
an instruction, but its not easily controllable which one will
be chosen by the tablegen backend.
This patch adds a flag to mark patterns that should only be used
for compressing. This allows us to leave one canonical pattern
for uncompressing.
The obvious benefit of this is getting c.mv to uncompress to
the addi patern that is aliased to the mv pseudoinstruction. For
the add/and/or/xor/li patterns it just removes some unreachable
code from the generated code.
Reviewed By: frasercrmck
Differential Revision: https://reviews.llvm.org/D94894
As discussed on D56387, if we're shifting to extract the upper/lower half of a vXi64 vector then we're actually better off performing this at the subvector level as its very likely to fold into something.
combineConcatVectorOps can perform this in reverse if necessary.
Add the aarch64[_be]-*-gnu_ilp32 targets to support the GNU ILP32 ABI for AArch64.
The needed codegen changes were mostly already implemented in D61259, which added support for the watchOS ILP32 ABI. The main changes are:
- Wiring up the new target to enable ILP32 codegen and MC.
- ILP32 va_list support.
- ILP32 TLSDESC relocation support.
There was existing MC support for ELF ILP32 relocations from D25159 which could be enabled by passing "-target-abi ilp32" to llvm-mc. This was changed to check for "gnu_ilp32" in the target triple instead. This shouldn't cause any issues since the existing support was slightly broken: it was generating ELF64 objects instead of the ELF32 object files expected by the GNU ILP32 toolchain.
This target has been tested by running the full rustc testsuite on a big-endian ILP32 system based on the GCC ILP32 toolchain.
Reviewed By: kristof.beyls
Differential Revision: https://reviews.llvm.org/D94143
If constants are hidden behind G_ANYEXT we can treat them same way as G_SEXT.
For that purpose we extend getConstantVRegValWithLookThrough with option
to handle G_ANYEXT same way as G_SEXT.
Differential Revision: https://reviews.llvm.org/D92219
With tfe on there can be a vgpr write to vdata+1.
Add tablegen support for 5 register vdata store.
This is required for 4 register vdata store with tfe.
Differential Revision: https://reviews.llvm.org/D94960
When constraining an operand register using constrainOperandRegClass(),
the function may emit a COPY in case the provided register class does
not match the current operand register class. However, the operand
itself is not updated to make use of the COPY, thereby resulting in
incorrect code. This patch fixes that bug by updating the machine
operand accordingly.
Reviewed By: dsanders
Differential Revision: https://reviews.llvm.org/D91244
X86 allows for the "addr32" and "addr16" address size override prefixes.
Also, these and the segment override prefixes should be recognized as
valid prefixes.
Differential Revision: https://reviews.llvm.org/D94726
For Zvlsseg, we need continuous vector registers for the values. We need
to define new register classes for the different combinations of (number
of fields and LMUL). For example,
when the number of fields(NF) = 3, LMUL = 2, the values will be assigned
to (V0M2, V2M2, V4M2), (V2M2, V4M2, V6M2), (V4M2, V6M2, V8M2), ...
We define the vlseg intrinsics with multiple outputs. There is no way to
describe the codegen patterns with multiple outputs in the tablegen
files. We do the codegen in RISCVISelDAGToDAG and use EXTRACT_SUBREG to
extract the values of output.
The multiple scalable vector values will be put into a struct. This
patch is depended on the support for scalable vector struct.
Differential Revision: https://reviews.llvm.org/D94229
Make it easier to reuse for intrinsic vrgatherei16
which needs to encode both LMUL & EMUL in the instruction name,
like PseudoVRGATHEREI16_VV_M1_M1 and PseudoVRGATHEREI16_VV_M1_M2.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D94951
This reverts commit 418df4a6ab.
This change broke emscripten tests, I believe because it started
generating 5-byte a wide table index in the call_indirect instruction.
Neither v8 nor wabt seem to be able to handle that. The spec
currently says that this is single 0x0 byte and:
"In future versions of WebAssembly, the zero byte occurring in the
encoding of the call_indirectcall_indirect instruction may be used to
index additional tables."
So we need to revisit this change. For backwards compat I guess
we need to guarantee that __indirect_function_table is always at
address zero. We could also consider making this a single-byte
relocation with and assert if have more than 127 tables (for now).
Differential Revision: https://reviews.llvm.org/D95005
NotHasStdExtZbb doesn't have an AssemblerPredicate associated with it
so it didn't do anything. We don't need it either because the sorting
rules in tablegen prioritize by number of predicates. So the
dedicated instructions in the B extension that have predicates
will be prioritized automatically.
If we are able to compare with 0 instead of 1, we might be able
to fold the setcc into a beqz/bnez.
Often these setccs start life as an xor that gets converted to
a setcc by DAG combiner's rebuildSetcc. I looked into a detecting
(xor X, 1) and converting to (seteq X, 0) based on boolean contents
being 0/1 in rebuildSetcc instead of using computeKnownBits. It was
very perturbing to AMDGPU tests which I didn't look closely at.
It had a few changes on a couple other targets, but didn't seem
to be much if any improvement.
Reviewed By: lenary
Differential Revision: https://reviews.llvm.org/D94730
The Hexagon Vector Combine pass genertes stores for a complete
aligned vector. The start of each section is a multiple of the
vector size, so that value is passed to normalize to compute
the offset of the stores in the section. The first store may
not occur at offset 0 when there is a gap between sections.
Rename the *_gfx9_gfx10 ttmp registers to *_gfx9plus for simplicity,
and use the corresponding isGFX9Plus predicate to decide when to use
them instead of the old *_vi versions.
Differential Revision: https://reviews.llvm.org/D94975
Original patch by @rogfer01.
This patch adds support for sign-, zero-, and any-extension from
scalable mask vector types to integer vector types, as well as
truncation in the opposite direction.
Authored-by: Roger Ferrer Ibanez <rofirrim@gmail.com>
Co-Authored-by: Fraser Cormack <fraser@codeplay.com>
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D94590
We have no lowering for VSELECT vXi1, vXi1, vXi1, so mark them as
expanded to turn them into a series of logical operations.
Differential Revision: https://reviews.llvm.org/D94946
This patch factors out the "VLMax" operand passed to most
scalable-vector ISel patterns into a property of each VType.
This is seen as a preparatory change to allow RVV in the future to
more easily support fixed-length vector types with constrained vector
lengths, with the AVL operand set to the length of the fixed-length
vector. It has no effect on the scalable code generation path.
Reviewed By: HsiangKai
Differential Revision: https://reviews.llvm.org/D94594
This adds some basic MVE sadd_sat/ssub_sat/uadd_sat/usub_sat costs,
based on when the instruction is legal. With smaller than legal types
that are promoted we generate shr(qadd(shl, shl)), so the cost is 4
appropriately.
Differential Revision: https://reviews.llvm.org/D94958
When performing peephole optimization to simplify the code, after removing
passed FPSP/XSRSP instruction we will set any uses of that FRSP/XSRSP to the
source of the FRSP/XSRSP.
We are finding the machine instruction using virtual register holding FRSP/XSRSP
results by searching all following instructions and encountering an issue
that the first use of the virtual register is a debug MI causing:
1. virtual register in the debug MI removed unexpectedly.
2. virtual register used in non-debug MI not replaced with the source of
FRSP/XSRSP. which stays in a undef status.
This patch fix the issue by only searching non-debug machine instruction using
virtual register holding FRSP/XSRSP results when the vr only has one non debug
usage.
Differential Revisien: https://reviews.llvm.org/D94711
Reviewed by: nemanjai
This CPU supports all v8.5a features except BTI, and so identifies as v8.5a to
Clang. A bit weird, but the best way for things like xnu to detect the new
features it cares about.
This patch computes the cost for vector.reduce<operand> for scalable vectors.
The cost is split into two parts: the legalization cost and the horizontal
reduction.
Differential Revision: https://reviews.llvm.org/D93639
If a srl doesn't introduce any sign bits into the truncated result, then replace with a sra to let us use a PACKSS truncation - fixes a regression noticed in D56387 on pre-SSE41 targets that don't have PACKUSDW.
This patch handles cases where we have to save/restore the link register
into the stack and and load/store instruction which use the stack are
part of the outlined region. It checks that there will be no overflow
introduced by the new offset and fixup these instructions accordingly.
Differential Revision: https://reviews.llvm.org/D92934
Previously uniqueCallSite could have race conditions between different
threads. Now it is accessed with an atomic RMW and will be unique
between different threads.
Differential Revision: https://reviews.llvm.org/D94784
This patch changes to make call_indirect explicitly refer to the
corresponding function table, residualizing TABLE_NUMBER relocs against
it.
With this change, wasm-ld now sees all references to tables, and can
link multiple tables.
Differential Revision: https://reviews.llvm.org/D90948
As of 8dacca943a, we sign extend the atomic loaded
operand for signed subword comparisons. However, the assumption that the other
operand is correctly sign extended doesn't always hold. This patch sign extends
the other operand if it needs to be sign extended.
This is a second fix for https://bugs.llvm.org/show_bug.cgi?id=30451
Differential revision: https://reviews.llvm.org/D94058
Luo, Yuanke