Summary:
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: arsenm, dschuff, jyknight, sdardis, nemanjai, jvesely, nhaehnle, sbc100, jgravelle-google, hiraditya, aheejin, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, kerbowa, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73885
These can be lowered to code sequences using CMPFP and CMPFPE which then get
selected to VCMP and VCMPE. The implementation isn't fully correct, as the chain
operand isn't handled correctly, but resolving that looks like it would involve
changes around FPSCR-handling instructions and how the FPSCR is modelled.
The fp-intrinsics test was already testing some of this but as the entire test
was being XFAILed it wasn't noticed. Un-XFAIL the test and instead leave the
cases where we aren't generating the right instruction sequences as FIXME.
Differential Revision: https://reviews.llvm.org/D73194
Summary:
In big-endian MVE, the simple vector load/store instructions (i.e.
both contiguous and non-widening) don't all store the bytes of a
register to memory in the same order: it matters whether you did a
VSTRB.8, VSTRH.16 or VSTRW.32. Put another way, the in-register
formats of different vector types relate to each other in a different
way from the in-memory formats.
So, if you want to 'bitcast' or 'reinterpret' one vector type as
another, you have to carefully specify which you mean: did you want to
reinterpret the //register// format of one type as that of the other,
or the //memory// format?
The ACLE `vreinterpretq` intrinsics are specified to reinterpret the
register format. But I had implemented them as LLVM IR bitcast, which
is specified for all types as a reinterpretation of the memory format.
So a `vreinterpretq` intrinsic, applied to values already in registers,
would code-generate incorrectly if compiled big-endian: instead of
emitting no code, it would emit a `vrev`.
To fix this, I've introduced a new IR intrinsic to perform a
register-format reinterpretation: `@llvm.arm.mve.vreinterpretq`. It's
implemented by a trivial isel pattern that expects the input in an
MQPR register, and just returns it unchanged.
In the clang codegen, I only emit this new intrinsic where it's
actually needed: I prefer a bitcast wherever it will have the right
effect, because LLVM understands bitcasts better. So we still generate
bitcasts in little-endian mode, and even in big-endian when you're
casting between two vector types with the same lane size.
For testing, I've moved all the codegen tests of vreinterpretq out
into their own file, so that they can have a different set of RUN
lines to check both big- and little-endian.
Reviewers: dmgreen, MarkMurrayARM, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D73786
Summary:
These instructions generate a vector of consecutive elements starting
from a given base value and incrementing by 1, 2, 4 or 8. The `wdup`
versions also wrap the values back to zero when they reach a given
limit value. The instruction updates the scalar base register so that
another use of the same instruction will continue the sequence from
where the previous one left off.
At the IR level, I've represented these instructions as a family of
target-specific intrinsics with two return values (the constructed
vector and the updated base). The user-facing ACLE API provides a set
of intrinsics that throw away the written-back base and another set
that receive it as a pointer so they can update it, plus the usual
predicated versions.
Because the intrinsics return two values (as do the underlying
instructions), the isel has to be done in C++.
This is the first family of MVE intrinsics that use the `imm_1248`
immediate type in the clang Tablegen framework, so naturally, I found
I'd given it the wrong C integer type. Also added some tests of the
check that the immediate has a legal value, because this is the first
time those particular checks have been exercised.
Finally, I also had to fix a bug in MveEmitter which failed an
assertion when I nested two `seq` nodes (the inner one used to extract
the two values from the pair returned by the IR intrinsic, and the
outer one put on by the predication multiclass).
Reviewers: dmgreen, MarkMurrayARM, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D73357
Summary:
The unpredicated case of this is trivial: the clang codegen just makes
a vector splat of the input, and LLVM isel is already prepared to
handle that. For the predicated version, I've generated a `select`
between the same vector splat and the `inactive` input parameter, and
added new Tablegen isel rules to match that pattern into a predicated
`MVE_VDUP` instruction.
Reviewers: dmgreen, MarkMurrayARM, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D73356
This reverts commit e34801c8e6 and the followup due to multiple
problems.
I've tried to keep the tests and RDA parts where possible, as those
still seem useful.
Summary: This is a first step before changing the types to llvm::Align and introduce functions to ease client code.
Reviewers: courbet
Subscribers: arsenm, sdardis, nemanjai, jvesely, nhaehnle, hiraditya, kbarton, jrtc27, atanasyan, jsji, kerbowa, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73785
fadd/fmul reductions without reassoc are lowered to
VECREDUCE_STRICT_FADD/FMUL nodes, which don't have legalization
support. Until that is in place, expand these intrinsics on
ARM and AArch64. Other targets always expand the vector reduction
intrinsics.
Additionally expand fmax/fmin reductions without nonan flag on
AArch64, as the backend asserts that the flag is present when
lowering VECREDUCE_FMIN/FMAX.
This fixes https://bugs.llvm.org/show_bug.cgi?id=44600.
Differential Revision: https://reviews.llvm.org/D73135
Summary:
The initialization of RegisterBank needs to be done only once. The
logic of AlreadyInit has data race, use llvm::call_once instead.
This is continuing work of D73587.
Reviewers: arsenm, rovka, dsanders, t.p.northover, efriedma, apazos
Reviewed By: arsenm
Subscribers: wdng, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73605
This adds some missing MVE opcodes to evaluateBranch, which results in
llvm-objdump being able to print the PC relative branch target as an
annotation.
Differential Revision: https://reviews.llvm.org/D73553
Add several new helpers to RDA:
- hasLocalDefBefore
- isRegDefinedAfter
- isSafeToDefRegAt
And move two bits of logic from ARMLowOverheadLoops into RDA:
- isSafeToMove
- isSafeToRemove
Both of these have some wrappers too to make them more convienent to
use.
Differential Revision: https://reviews.llvm.org/D73460
This is how it should've been and brings it more in line with
std::string_view. There should be no functional change here.
This is mostly mechanical from a custom clang-tidy check, with a lot of
manual fixups. It uncovers a lot of minor inefficiencies.
This doesn't actually modify StringRef yet, I'll do that in a follow-up.
The MVE_VLDRWU32_qi_pre gather loads, like the other _pre/_post mve
loads returns the writeback as result 0, the value as result 1. The llvm
ir intrinsic seems to have this the other way around though, and so when
lowering from one to the other we need to switch the first two outputs.
I've also fixed up the types of _pre/_post on normal MVE loads. There we
were already getting the values the right way around, just not for the
types. I don't believe this was causing anything to go wrong, but it was
very confusing to read in the debug output.
Differential Revision: https://reviews.llvm.org/D73370
We had support for runtime trip count values, but not constants, and this adds
supports for that.
And added a minor optimisation while I was add it: don't invoke Cleanup when
there's nothing to clean up.
Differential Revision: https://reviews.llvm.org/D73198
When expanding the LoopStart, we try to remove the iteration count
calculation. However, if part of the calculation was also used to
calculate the number of elements we could end up deleting
instructions that were required to feed DLSTP/WLSTP.
Differential Revision: https://reviews.llvm.org/D73275
The codegen for splitting a llvm.vector.reduction intrinsic into parts
will be better than the codegen for the generic reductions. This will
only directly effect when vectorization factors are specified by the
user.
Also added tests to make sure the codegen for larger reductions is OK.
Differential Revision: https://reviews.llvm.org/D72257
Summary:
This is a follow up on https://reviews.llvm.org/D71473#inline-647262.
There's a caveat here that `Align(1)` relies on the compiler understanding of `Log2_64` implementation to produce good code. One could use `Align()` as a replacement but I believe it is less clear that the alignment is one in that case.
Reviewers: xbolva00, courbet, bollu
Subscribers: arsenm, dylanmckay, sdardis, nemanjai, jvesely, nhaehnle, hiraditya, kbarton, jrtc27, atanasyan, jsji, Jim, kerbowa, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D73099
Similar to the function attribute `prefix` (prefix data),
"patchable-function-prefix" inserts data (M NOPs) before the function
entry label.
-fpatchable-function-entry=2,1 (1 NOP before entry, 1 NOP after entry)
will look like:
```
.type foo,@function
.Ltmp0: # @foo
nop
foo:
.Lfunc_begin0:
# optional `bti c` (AArch64 Branch Target Identification) or
# `endbr64` (Intel Indirect Branch Tracking)
nop
.section __patchable_function_entries,"awo",@progbits,get,unique,0
.p2align 3
.quad .Ltmp0
```
-fpatchable-function-entry=N,0 + -mbranch-protection=bti/-fcf-protection=branch has two reasonable
placements (https://gcc.gnu.org/ml/gcc-patches/2020-01/msg01185.html):
```
(a) (b)
func: func:
.Ltmp0: bti c
bti c .Ltmp0:
nop nop
```
(a) needs no additional code. If the consensus is to go for (b), we will
need more code in AArch64BranchTargets.cpp / X86IndirectBranchTracking.cpp .
Differential Revision: https://reviews.llvm.org/D73070
Summary:
Immediate vmvnq is code-generated as a simple vector constant in IR,
and left to the backend to recognize that it can be created with an
MVE VMVN instruction. The predicated version is represented as a
select between the input and the same constant, and I've added a
Tablegen isel rule to turn that into a predicated VMVN. (That should
be better than the previous VMVN + VPSEL: it's the same number of
instructions but now it can fold into an adjacent VPT block.)
The unpredicated forms of VBIC and VORR are done by enabling the same
isel lowering as for NEON, recognizing appropriate immediates and
rewriting them as ARMISD::VBICIMM / ARMISD::VORRIMM SDNodes, which I
then instruction-select into the right MVE instructions (now that I've
also reworked those instructions to use the same MC operand encoding).
In order to do that, I had to promote the Tablegen SDNode instance
`NEONvorrImm` to a general `ARMvorrImm` available in MVE as well, and
similarly for `NEONvbicImm`.
The predicated forms of VBIC and VORR are represented as a vector
select between the original input vector and the output of the
unpredicated operation. The main convenience of this is that it still
lets me use the existing isel lowering for VBICIMM/VORRIMM, and not
have to write another copy of the operand encoding translation code.
This intrinsic family is the first to use the `imm_simd` system I put
into the MveEmitter tablegen backend. So, naturally, it showed up a
bug or two (emitting bogus range checks and the like). Fixed those,
and added a full set of tests for the permissible immediates in the
existing Sema test.
Also adjusted the isel pattern for `vmovlb.u8`, which stopped matching
because lowering started turning its input into a VBICIMM. Now it
recognizes the VBICIMM instead.
Reviewers: dmgreen, MarkMurrayARM, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D72934
Summary:
In NEON, the immediate forms of VBIC and VORR are each represented as
a single MC instruction, which takes its immediate operand already
encoded in a NEON-friendly format: 8 data bits, plus some control bits
indicating how to expand them into a full vector.
In MVE, we represented immediate VBIC and VORR as four separate MC
instructions each, for an 8-bit immediate shifted left by 0, 8, 16 or
24 bits. For each one, the value of the immediate operand is in the
'natural' form, i.e. the numerical value that would actually be BICed
or ORRed into each vector lane (and also the same value shown in
assembly). For example, MVE_VBICIZ16v4i32 takes an operand such as
0xab0000, which NEON would represent as 0xab | (control bits << 8).
The MVE approach is superficially nice (it makes assembly input and
output easy, and it's also nice if you're manually constructing
immediate VBICs). But it turns out that it's better for isel if we
make the NEON and MVE instructions work the same, because the
ARMISD::VBICIMM and VORRIMM node types already encode their immediate
into the NEON format, so it's easier if we can just use it.
Also, this commit reduces the total amount of code rather than
increasing it, which is surely an indication that it really is simpler
to do it this way!
Reviewers: dmgreen, ostannard, miyuki, MarkMurrayARM
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73205
The hasSideEffect parameter is usually automatically inferred from
instruction patterns. For some of our MVE instructions, we do not have
patterns though, such as for the pre/post inc loads and stores. This
instead specifies the flag manually on the base MVE_VLDRSTR_base
tablegen class, making sure we get this correct.
This can help with scheduling multiple loads more optimally. Here I've
added a unittest as a more direct form of testing.
Differential Revision: https://reviews.llvm.org/D73117
This is a very basic MVE gather/scatter cost model, based roughly on the
code that we will currently produce. It does not handle truncating
scatters or extending gathers correctly yet, as it is difficult to tell
that they are going to be correctly extended/truncated from the limited
information in the cost function.
This can be improved as we extend support for these in the future.
Based on code originally written by David Sherwood.
Differential Revision: https://reviews.llvm.org/D73021
Extends the gather/scatter pass in MVEGatherScatterLowering.cpp to
enable the transformation of masked scatters into calls to MVE's masked
scatter intrinsic.
Differential Revision: https://reviews.llvm.org/D72856
Summary:
A recent commit accidentally defined names like `MVE_VMAXAs8` as
instances of the multiclass `MVE_VMINA`, and vice versa. This has no
effect on the test suite, because nothing directly refers to those
instruction names (the isel patterns are generated in Tablegen using
`!cast<Instruction>(NAME)` inside a lower-level multiclass). But it
means that `llvm-mc -show-inst` was listing VMAXA as VMINA, and it
would also affect any further draft code gen patches that use those
instruction ids.
Reviewers: MarkMurrayARM, dmgreen, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73034
This patch uses helper function rewriteLoopExitValues that is refactored in
D72602 to rematerialise the iteration count in exit blocks, so that we can
clean-up loop update expressions inside the hardware-loops later in
ARMLowOverheadLoops, which is necessary to get actual performance gains for
tail-predicated loops.
Differential Revision: https://reviews.llvm.org/D72714
This adds Post inc variants of the VLD2/4 and VST2/4 instructions in
MVE. It uses the same mechanism/nodes as Neon, transforming the
intrinsic+add pair into a ARMISD::VLD2_UPD, which gets selected to a
post-inc instruction. The code to do that is mostly taken from the
existing Neon code, but simplified as less variants are needed.
It also fills in some getTgtMemIntrinsic for the arm.mve.vld2/4
instrinsics, which allow the nodes to have MMO's, calculated as the full
length to the memory being loaded/stored.
Differential Revision: https://reviews.llvm.org/D71194
We were previously not necessarily favouring postinc for the MVE loads
and stores, leading to extra code prior to the loop to set up the
preinc. MVE in general can benefit from postinc (as we don't have
unrolled loops), and certain instructions like the VLD2's only post-inc
versions are available.
Differential Revision: https://reviews.llvm.org/D70790
Except AMDGPU/R600RegisterInfo (a bunch of MIR tests seem to have
problems), every target overrides it with true. PostMachineScheduler
requires livein information. Not providing it can cause assertion
failures in ScheduleDAGInstrs::addSchedBarrierDeps().
Introduce a method to walk through use-def chains to decide whether
it's possible to remove a given instruction and its users. These
instructions are then stored in a set until the end of the transform
when they're erased. This is now used to perform checks on the
iteration count (LoopDec chain), element count (VCTP chain) and the
possibly redundant iteration count.
As well as being able to remove chains of instructions, we know also
check that the sub feeding the vctp is producing the expected value.
Differential Revision: https://reviews.llvm.org/D71837
Recommitting e93e0d413f after reverting due to test failures, which
will hopefully now be fixed. Original commit message:
After expanding the pseudo instructions, update the liveness info.
We do this in a post-order traversal of the loop, including its
exit blocks and preheader(s).
Differential Revision: https://reviews.llvm.org/D72131
Summary:
For builds with LLVM_BUILD_LLVM_DYLIB=ON and BUILD_SHARED_LIBS=OFF
this change makes all symbols in the target specific libraries hidden
by default.
A new macro called LLVM_EXTERNAL_VISIBILITY has been added to mark symbols in these
libraries public, which is mainly needed for the definitions of the
LLVMInitialize* functions.
This patch reduces the number of public symbols in libLLVM.so by about
25%. This should improve load times for the dynamic library and also
make abi checker tools, like abidiff require less memory when analyzing
libLLVM.so
One side-effect of this change is that for builds with
LLVM_BUILD_LLVM_DYLIB=ON and LLVM_LINK_LLVM_DYLIB=ON some unittests that
access symbols that are no longer public will need to be statically linked.
Before and after public symbol counts (using gcc 8.2.1, ld.bfd 2.31.1):
nm before/libLLVM-9svn.so | grep ' [A-Zuvw] ' | wc -l
36221
nm after/libLLVM-9svn.so | grep ' [A-Zuvw] ' | wc -l
26278
Reviewers: chandlerc, beanz, mgorny, rnk, hans
Reviewed By: rnk, hans
Subscribers: merge_guards_bot, luismarques, smeenai, ldionne, lenary, s.egerton, pzheng, sameer.abuasal, MaskRay, wuzish, echristo, Jim, hiraditya, michaelplatings, chapuni, jholewinski, arsenm, dschuff, jyknight, dylanmckay, sdardis, nemanjai, jvesely, javed.absar, sbc100, jgravelle-google, aheejin, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, zzheng, edward-jones, mgrang, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, kristina, jsji, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D54439
Fix a missing and broken test: 2 VPT blocks predicated on the same VCMP
instruction that can be folded. The problem was that for each VPT block, we
record the predicate statements with a list, but the same instruction was added
twice. Thus, we were running in an assert trying to remove the same instruction
twice. To avoid this the instructions are now recorded with a set.
Differential Revision: https://reviews.llvm.org/D72699
Summary:
This cleans up a lot of ugly `foreach` bodges that I've been using to
work around the lack of those two language features. Now they both
exist, I can make then all into something more legible!
In particular, in the common pattern in `ARMInstrMVE.td` where a
multiclass defines an `Instruction` instance plus one or more `Pat` that
select it, I've used a `defvar` to wrap `!cast<Instruction>(NAME)` so
that the patterns themselves become a little more legible.
Replacing a `foreach` with a `defvar` removes a level of block
structure, so several pieces of code have their indentation changed by
this patch. Best viewed with whitespace ignored.
NFC: the output of `llvm-tblgen -print-records` on the two affected
Tablegen sources is exactly identical before and after this change, so
there should be no effect at all on any of the other generated files.
Reviewers: MarkMurrayARM, miyuki
Reviewed By: MarkMurrayARM
Subscribers: kristof.beyls, hiraditya, dmgreen, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D72690
We have a whitelist of instructions that we allow when tail
predicating, since these are trivial ones that we've deemed need no
special handling. Now change ARMLowOverheadLoops to allow the
non-trivial instructions if they're contained within a valid VPT
block. Since a valid block is one that is predicated upon the VCTP so
we know that these non-trivial instructions will still behave as
expected once the implicit predication is used instead.
This also fixes a previous test failure.
Differential Revision: https://reviews.llvm.org/D72509
Use the already provided helper function to get the operand type so
that we can detect whether the vpr is being used as a predicate or
not. Also use existing helpers to get the predicate indices when we
converting the vpt blocks. This enables us to support both types of
vpr predicate operand.
Differential Revision: https://reviews.llvm.org/D72504
Summary:
This patch fixes pr23772 [ARM] r226200 can emit illegal thumb2 instruction: "sub sp, r12, #80".
The violation was that SUB and ADD (reg, immediate) instructions can only write to SP if the source register is also SP. So the above instructions was unpredictable.
To enforce that the instruction t2(ADD|SUB)ri does not write to SP we now enforce the destination register to be rGPR (That exclude PC and SP).
Different than the ARM specification, that defines one instruction that can read from SP, and one that can't, here we inserted one that can't write to SP, and other that can only write to SP as to reuse most of the hard-coded size optimizations.
When performing this change, it uncovered that emitting Thumb2 Reg plus Immediate could not emit all variants of ADD SP, SP #imm instructions before so it was refactored to be able to. (see test/CodeGen/Thumb2/mve-stacksplot.mir where we use a subw sp, sp, Imm12 variant )
It also uncovered a disassembly issue of adr.w instructions, that were only written as SUBW instructions (see llvm/test/MC/Disassembler/ARM/thumb2.txt).
Reviewers: eli.friedman, dmgreen, carwil, olista01, efriedma, andreadb
Reviewed By: efriedma
Subscribers: gbedwell, john.brawn, efriedma, ostannard, kristof.beyls, hiraditya, dmgreen, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70680
Due to the current way that we collect predicated instructions, we
can't easily handle vpsel in tail predicated loops. There are a
couple of issues:
1) It will use the VPR as a predicate operand, but doesn't have to be
instead a VPT block, which means we can assert while building up
the VPT block because we don't find another VPST to being a new
one.
2) VPSEL still requires a VPR operand even after tail predicating,
which means we can't remove it unless there is another
instruction, such as vcmp, that can provide the VPR def.
The first issue should be a relatively simple fix in the logic of the
LowOverheadLoops pass, whereas the second will require us to
represent the 'implicit' tail predication with an explicit value.
Differential Revision: https://reviews.llvm.org/D72629
Enables the masked gather pass to create a masked
gather loading from a base and vector of offsets.
This also enables v8i16 and v16i8 gather loads.
Differential Revision: https://reviews.llvm.org/D72330
The argument is llvm::null() everywhere except llvm::errs() in
llvm-objdump in -DLLVM_ENABLE_ASSERTIONS=On builds. It is used by no
target but X86 in -DLLVM_ENABLE_ASSERTIONS=On builds.
If we ever have the needs to add verbose log to disassemblers, we can
record log with a member function, instead of passing it around as an
argument.
Summary:
This always just used the same libcall as unordered, but the comparison predicate was different. This change appears to have been made when targets were given the ability to override the predicates. Before that they were hardcoded into the type legalizer. At that time we never inverted predicates and we handled ugt/ult/uge/ule compares by emitting an unordered check ORed with a ogt/olt/oge/ole checks. So only ordered needed an inverted predicate. Later ugt/ult/uge/ule were optimized to only call a single libcall and invert the compare.
This patch removes the ordered entries and just uses the inverting logic that is now present. This removes some odd things in both the Mips and WebAssembly code.
Reviewers: efriedma, ABataev, uweigand, cameron.mcinally, kpn
Reviewed By: efriedma
Subscribers: dschuff, sdardis, sbc100, arichardson, jgravelle-google, kristof.beyls, hiraditya, aheejin, sunfish, atanasyan, Petar.Avramovic, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D72536
Summary:
This patch fixes pr23772 [ARM] r226200 can emit illegal thumb2 instruction: "sub sp, r12, #80".
The violation was that SUB and ADD (reg, immediate) instructions can only write to SP if the source register is also SP. So the above instructions was unpredictable.
To enforce that the instruction t2(ADD|SUB)ri does not write to SP we now enforce the destination register to be rGPR (That exclude PC and SP).
Different than the ARM specification, that defines one instruction that can read from SP, and one that can't, here we inserted one that can't write to SP, and other that can only write to SP as to reuse most of the hard-coded size optimizations.
When performing this change, it uncovered that emitting Thumb2 Reg plus Immediate could not emit all variants of ADD SP, SP #imm instructions before so it was refactored to be able to. (see test/CodeGen/Thumb2/mve-stacksplot.mir where we use a subw sp, sp, Imm12 variant )
It also uncovered a disassembly issue of adr.w instructions, that were only written as SUBW instructions (see llvm/test/MC/Disassembler/ARM/thumb2.txt).
Reviewers: eli.friedman, dmgreen, carwil, olista01, efriedma
Reviewed By: efriedma
Subscribers: john.brawn, efriedma, ostannard, kristof.beyls, hiraditya, dmgreen, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70680
The current implementation assumes there is an instruction associated
with the transform, but this is not the case for
timm/TargetConstant/immarg values. These transforms should directly
operate on a specific MachineOperand in the source
instruction. TableGen would assert if you attempted to define an
equivalent GISDNodeXFormEquiv using timm when it failed to find the
instruction matcher.
Specially recognize SDNodeXForms on timm, and pass the operand index
to the render function.
Ideally this would be a separate render function type that looks like
void renderFoo(MachineInstrBuilder, const MachineOperand&), but this
proved to be somewhat mechanically painful. Add an optional operand
index which will only be passed if the transform should only look at
the one source operand.
Theoretically it would also be possible to only ever pass the
MachineOperand, and the existing renderers would check the parent. I
think that would be somewhat ugly for the standard usage which may
want to inspect other operands, and I also think MachineOperand should
eventually not carry a pointer to the parent instruction.
Use it in one sample pattern. This isn't a great example, since the
transform exists to satisfy DAG type constraints. This could also be
avoided by just changing the MachineInstr's arbitrary choice of
operand type from i16 to i32. Other patterns have nontrivial uses, but
this serves as the simplest example.
One flaw this still has is if you try to use an SDNodeXForm defined
for imm, but the source pattern uses timm, you still see the "Failed
to lookup instruction" assert. However, there is now a way to avoid
it.
Only PPC seems to be using it, and only checks some simple cases and
doesn't distinguish between FP. Just switch to using LLT to simplify
use from GlobalISel.
We don't unroll vector loops for MVE targets, but we miss the case
when loops only contain intrinsic calls. So just move the logic a
bit to catch this case.
Differential Revision: https://reviews.llvm.org/D72440
After expanding the pseudo instructions, update the liveness info.
We do this in a post-order traversal of the loop, including its
exit blocks and preheader(s).
Differential Revision: https://reviews.llvm.org/D72131
This batch of intrinsics fills in all the shift instructions that take
a variable shift distance in a register, instead of an immediate. Some
of these instructions take a single shift distance in a scalar
register and apply it to all lanes; others take a vector of per-lane
distances.
These instructions are all basically one family, varying in whether
they saturate out-of-range values, and whether they round when bits
are shifted off the bottom. I've implemented them at the IR level by a
much smaller family of IR intrinsics, which take flag parameters to
indicate saturating and/or rounding (along with the usual one to
specify signed/unsigned integers).
An oddity is that all of them are //left// shift instructions – but if
you pass a negative shift count, they'll shift right. So the vector
shift distances are always vectors of //signed// integers, regardless
of whether you're considering the other input vector to be of signed
or unsigned. Also, even the simplest `vshlq` instruction in this
family (neither saturating nor rounding) has to be implemented as an
IR intrinsic, because the ordinary LLVM IR `shl` operation would
consider an out-of-range shift count to be undefined behavior.
Reviewers: dmgreen, MarkMurrayARM, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D72329
This batch of intrinsics covers two sets of immediate shift
instructions, which have in common that they only overwrite part of
their output register and so they need an extra input giving its
previous value.
The VSLI and VSRI instructions shift each lane of the input vector
left or right just as if they were normal immediate VSHL/VSHR, but
then they only overwrite the output bits that correspond to actual
shifted bits of the input. So VSLI will leave the low n bits of each
output lane unchanged, and VSRI the same with the top n bits.
The V[Q][R]SHR[U]N family are all narrowing shifts: they take an input
vector of 2n-bit integers, shift each lane right by a constant, and
then narrowing the shifted result to only n bits. So they only
overwrite half of the n-bit lanes in the output register, and the B/T
suffix indicates whether it's the bottom or top half of each 2n-bit
lane.
I've implemented the whole of the latter family using a single IR
intrinsic `vshrn`, which takes a lot of i32 parameters indicating
which instruction it expands to (by specifying signedness of the input
and output types, whether it saturates and/or rounds, etc).
Reviewers: dmgreen, MarkMurrayARM, miyuki, ostannard
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D72328
Adds a pass to the ARM backend that takes a v4i32
gather and transforms it into a call to MVE's
masked gather intrinsics.
Differential Revision: https://reviews.llvm.org/D71743
This is a recommit of D71330, but with a few things fixed and changed:
1) ReachingDefAnalysis: this was not running with optnone as it was checking
skipFunction(), which other analysis passes don't do. I guess this is a
copy-paste from a codegen pass.
2) VPTBlockPass: here I've added skipFunction(), because like most/all
optimisations, we don't want to run this with optnone.
This fixes the issues with the initial/previous commit: the VPTBlockPass was
running with optnone, but ReachingDefAnalysis wasn't, and so VPTBlockPass was
crashing querying ReachingDefAnalysis.
I've added test case mve-vpt-block-optnone.mir to check that we don't run
VPTBlock with optnone.
Differential Revision: https://reviews.llvm.org/D71470
Summary:
Instead of generating two i32 instructions for each load or store of a volatile
i64 value (two LDRs or STRs), now emit LDRD/STRD.
These improvements cover architectures implementing ARMv5TE or Thumb-2.
Reviewers: dmgreen, efriedma, john.brawn, nickdesaulniers
Reviewed By: efriedma, nickdesaulniers
Subscribers: nickdesaulniers, vvereschaka, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70072
printInst prints a branch/call instruction as `b offset` (there are many
variants on various targets) instead of `b address`.
It is a convention to use address instead of offset in most external
symbolizers/disassemblers. This difference makes `llvm-objdump -d`
output unsatisfactory.
Add `uint64_t Address` to printInst(), so that it can pass the argument to
printInstruction(). `raw_ostream &OS` is moved to the last to be
consistent with other print* methods.
The next step is to pass `Address` to printInstruction() (generated by
tablegen from the instruction set description). We can gradually migrate
targets to print addresses instead of offsets.
In any case, downstream projects which don't know `Address` can pass 0 as
the argument.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D72172
The segmented stack lowering code appears to be using ARM opcodes under
Thumb2. The MRC opcode will be the same for Thumb and ARM, but t2LDR
seems wrong. Either way, using the correct thumb vs arm opcodes is more
correct.
Differential Revision: https://reviews.llvm.org/D72074
We were previously unconditionally using the ARM::TRAP opcode, even
under Thumb. My understanding is that these are essentially the same
thing (they both result in a trap under Thumb), but the ARM::TRAP opcode
is marked as requiring IsARM, so it is more correct to use ARM::tTRAP.
Differential Revision: https://reviews.llvm.org/D72075
Summary:
Running an end-to-end test last week I noticed that a lot of the ACLE
intrinsics that operate differently on vectors of signed and unsigned
integers were ending up generating the signed version of the
instruction unconditionally. This is because the IR intrinsics had no
way to distinguish signed from unsigned: the LLVM type system just
calls them both `v8i16` (or whatever), so you need either separate
intrinsics for signed and unsigned, or a flag parameter that tells
ISel which one to choose.
This patch fixes all the problems of that kind that I've noticed, by
adding an i32 flag parameter to many of the IR intrinsics which is set
to 1 for unsigned (matching the existing practice in cases where we
got it right), and conditioning all the isel patterns on that flag. So
the fundamental change is in `IntrinsicsARM.td`, changing the
low-level IR intrinsics API; there are knock-on changes in
`arm_mve.td` (adjusting code gen for the ACLE intrinsics to use the
modified API) and in `ARMInstrMVE.td` (adjusting isel to expect the
new unsigned flags). The rest of this patch is boringly updating tests.
Reviewers: dmgreen, miyuki, MarkMurrayARM
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D72270
Summary:
Due to a copy-paste error in the isel patterns, the predicated version
of this intrinsic was expanding to the `VMAXNMT.F32` instruction
instead of `VMAXNMT.F16`. Similarly for vminnm.
Reviewers: dmgreen, miyuki, MarkMurrayARM
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D72269
I've added a few more debug messages to MVETailPredication because I wanted to
trace better which instructions are added/removed. And while I was at it, I
factored out one function which I thought was clearer, and have added some
comments to describe better the flow between MVETailPredication and
ARMLowOverheadLoops.
Differential Revision: https://reviews.llvm.org/D71549
This simplifies the generic interface and also makes SHF_ARM_PURECODE
more robust (fixes a TODO). Inspecting MCDataFragment contents covers
more cases than MCObjectStreamer::EmitBytes.
This adds extra scalar handling to isFMAFasterThanFMulAndFAdd, allowing
the target independent code to handle more folds in more situations (for
example if the fast math flags are present, but the global
AllowFPOpFusion option isnt). It also splits apart the HasSlowFPVMLx
into HasSlowFPVFMx, to allow VFMA and VMLA to be controlled separately
if needed.
Differential Revision: https://reviews.llvm.org/D72139
SCEVExpander modifies the underlying function so it is more suitable in
Transforms/Utils, rather than Analysis. This allows using other
transform utils in SCEVExpander.
Reviewers: sanjoy.google, efriedma, reames
Reviewed By: sanjoy.google
Differential Revision: https://reviews.llvm.org/D71537
AMDGPU can't unambiguously go back from the selected instruction
register class to the register bank without knowing if this was used
in a boolean context.
When the "disable-tail-calls" attribute was added, checks were added for
it in various backends. Now this code has proliferated, and it is
something the target is responsible for checking. Move that
responsibility back to the ISels (fast, global, and SD).
There's no major functionality change, except for targets that never
implemented this check.
This LLVM attribute was originally added in
d9699bc7bd (2015).
Reviewers: echristo, MaskRay
Differential Revision: https://reviews.llvm.org/D72118
For now, we didn't set the default operation action for SIGN_EXTEND_INREG for
vector type, which is 0 by default, that is legal. However, most target didn't
have native instructions to support this opcode. It should be set as expand by
default, as what we did for ANY_EXTEND_VECTOR_INREG.
Differential Revision: https://reviews.llvm.org/D70000
Summary:
Add missing part of patch D71361. Now that the stack-frame
can be operated using a addw/subw instruction, they should
appear in the unwinding list.
Reviewers: dmgreen, efriedma
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D72000
This adds ICmp to the list of instructions that we sink a splat to in a
loop, allowing the register forms of instructions to be selected more
often. It does not add FCmp yet as the results look a little odd, trying
to keep the register in an float reg and having to move it back to a GPR.
Differential Revision: https://reviews.llvm.org/D70997
Summary:
This patch allows to emit thumb2 add and sub
instructions with 12 bit immediates in the
emitT2RegPlusImmediate function.
- Splitting parts of the D70680
Reviewers: eli.friedman, olista01, efriedma
Reviewed By: efriedma
Subscribers: efriedma, kristof.beyls, hiraditya, dmgreen, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71361
This allows us to delete InlineAsm::Constraint_i workarounds in
SelectionDAGISel::SelectInlineAsmMemoryOperand overrides and
TargetLowering::getInlineAsmMemConstraint overrides.
They were introduced to X86 in r237517 to prevent crashes for
constraints like "=*imr". They were later copied to other targets.
As the extern_weak target might be missing, resolving to the absolute
address zero, we can't use the normal direct PC-relative branch
instructions (as that would result in relocations out of range).
Instead check the shouldAssumeDSOLocal method and load the address
from a COFF stub.
This matches what was done for X86 in 6bf108d77a.
Differential Revision: https://reviews.llvm.org/D71720
1) Fix an issue with the incorrect value being used for the number of
elements being passed to [d|w]lstp. We were trying to check that
the value was available at LoopStart, but this doesn't consider
that the last instruction in the block could also define the
register. Two helpers have been added to RDA for this.
2) Insert some code to now try to move the element count def or the
insertion point so that we can perform more tail predication.
3) Related to (1), the same off-by-one could prevent us from
generating a low-overhead loop when a mov lr could have been
the last instruction in the block.
4) Fix up some instruction attributes so that not all the
low-overhead loop instructions are labelled as branches and
terminators - as this is not true for dls/dlstp.
Differential Revision: https://reviews.llvm.org/D71609
Record the discovered VPT blocks while checking for validity and, for
now, only handle blocks that begin with VPST and not VPT. We're now
allowing more than one instruction to define vpr, but each block must
somehow be predicated using the vctp. This leaves us with several
scenarios which need fixing up:
1) A VPT block with is only predicated by the vctp and has no
internal vpr defs.
2) A VPT block which is only predicated by the vctp but has an
internal vpr def.
3) A VPT block which is predicated upon the vctp as well as another
vpr def.
4) A VPT block which is not predicated upon a vctp, but contains it
and all instructions within the block are predicated upon in.
The changes needed are, for:
1) The easy one, just remove the vpst and unpredicate the
instructions in the block.
2) Remove the vpst and unpredicate the instructions up to the
internal vpr def. Need insert a new vpst to predicate the
remaining instructions.
3) No nothing.
4) The vctp will be inside a vpt and the instruction will be removed,
so adjust the size of the mask on the vpst.
Differential Revision: https://reviews.llvm.org/D71107
Summary:
Instead of generating two i32 instructions for each load or store of a volatile
i64 value (two LDRs or STRs), now emit LDRD/STRD.
These improvements cover architectures implementing ARMv5TE or Thumb-2.
Reviewers: dmgreen, efriedma, john.brawn
Reviewed By: efriedma
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70072
Add an extra parameter so alignment can be taken under
consideration in gather/scatter legalization.
Differential Revision: https://reviews.llvm.org/D71610
In ARMLowOverheadLoops.cpp, MVETailPredication.cpp, and MVEVPTBlock.cpp we have
quite a few helper functions all looking at the opcodes of MVE instructions.
This moves all these utility functions to ARMBaseInstrInfo.
Diferential Revision: https://reviews.llvm.org/D71426
The big switch in `ARMBaseInstrInfo::getNumMicroOps` is missing cases for
`VLLDM` and `VLSTM`, which are currently defined with itineraries having a
dynamic count of micro-ops.
Assuming an optimistic case in which these instruction do not actually perform
loads or stores, and with the idea that Armv8-m cores are supposed to use the
new style scheduling models, this patch just sets the itinerary for those two
instructions to `NoItinerary`.
Differential Revision: https://reviews.llvm.org/D71266
We've been marking VPT incompatible instructions as invalid for tail
predication too, though this may not strictly be true. VPT are
incompatible and, unless its the first predicate def in a loop,
they shouldn't be compatible for tail predication either.
Differential Revision: https://reviews.llvm.org/D71410
Summary:
This patch adds intrinsics for the following MVE instructions:
* VABAV
* VMLADAV, VMLSDAV
* VMLALDAV, VMLSLDAV
* VRMLALDAVH, VRMLSLDAVH
Each of the above 4 groups has a corresponding new LLVM IR intrinsic,
since the instructions cannot be easily represented using
general-purpose IR operations.
Reviewers: simon_tatham, ostannard, dmgreen, MarkMurrayARM
Reviewed By: MarkMurrayARM
Subscribers: merge_guards_bot, kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71062
Summary:
This fills in the remaining shift operations that take a single vector
input and an immediate shift count: the `vqshl`, `vqshlu`, `vrshr` and
`vshll[bt]` families.
`vshll[bt]` (which shifts each input lane left into a double-width
output lane) is the most interesting one. There are separate MC
instruction ids for shifting by exactly the input lane width and
shifting by less than that, because the instruction encoding is so
completely different for the lane-width special case. So I had to
write two sets of patterns to match based on the immediate shift
count, which involved adding a ComplexPattern matcher to avoid the
general-case pattern accidentally matching the special case too. For
that family I've made sure to add an llc codegen test for both
versions of each instruction.
I'm experimenting with a new strategy for parametrising the isel
patterns for all these instructions: adding extra fields to the
relevant `Instruction` subclass itself, which are ignored by the
Tablegen backends that generate the MC data, but can be retrieved from
each instance of that instruction subclass when it's passed as a
template parameter to the multiclass that generates its isel patterns.
A nice effect of that is that I can fill in those informational fields
using `let` blocks, rather than having to type them out once per
instruction at `defm` time.
(As a result, quite a lot of existing instruction `def`s are
reindented by this patch, so it's clearer to read with whitespace
changes ignored.)
Reviewers: dmgreen, MarkMurrayARM, miyuki, ostannard
Reviewed By: MarkMurrayARM
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71458
We have custom lowering for operations converting to/from floating-point types
when we don't have hardware support for those types, and this doesn't interact
well with the target-independent legalization of the strict versions of these
operations. Fix this by adding similar custom lowering of the strict versions.
This fixes the last of the assertion failures in the CodeGen/ARM/fp-intrinsics
test, with the remaining failures due to poor instruction selection.
Differential Revision: https://reviews.llvm.org/D71127
Summary:
The use of a boolean isInteger flag (generally initialized using
VT.isInteger()) caused errors in our out-of-tree CHERI backend
(https://github.com/CTSRD-CHERI/llvm-project).
In our backend, pointers use a separate ValueType (iFATPTR) and therefore
.isInteger() returns false. This meant that getSetCCInverse() was using the
floating-point variant and generated incorrect code for us:
`(void *)0x12033091e < (void *)0xffffffffffffffff` would return false.
Committing this change will significantly reduce our merge conflicts
for each upstream merge.
Reviewers: spatel, bogner
Reviewed By: bogner
Subscribers: wuzish, arsenm, sdardis, nemanjai, jvesely, nhaehnle, hiraditya, kbarton, jrtc27, atanasyan, jsji, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70917
This reverts commit 9468e3334b.
There's a test that doesn't like this change. The RDA analysis
gets invalided by changes in the block, which is not taken into
account. Revert while I work on a fix for this.
Summary:
Better use of multiclass is used, and this helped find some existing
bugs in the predicated VMULL* intrinsics, which are now fixed.
The refactored VMULL[TB]Q_(INT|POLY)_M() intrinsics were discovered
to have an argument ("inactive") with incorrect type, and this required
a fix that is included in this whole patch. The argument "inactive"
should have been the same width (per vector element) as the return
type of the intrinsic, but was not in the case where the return type
was double the element width of the input types.
To assist in testing the multiclassing , and to thwart further gremlins,
the unit tests are improved in scope.
The *.ll tests are all generated by a small bit of throw-away scripting
from the corresponding *.c tests, and as such the diffs are large and
nasty. Look at the file rather than the diff.
Reviewers: dmgreen, miyuki, ostannard, simon_tatham
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71421
This adds ReachingDefAnalysis (RDA) to the VPTBlock pass, so that we can
reimplement findVCMPToFoldIntoVPS with just a few calls to RDA.
Differential Revision: https://reviews.llvm.org/D71330
Recommit e0b966643f. sub instructions were being generated for the
negated value, and for some reason they were the register only ones.
I think the problem was because I was grabbing the 'zero' from
vmovimm, which is a target constant. Now I'm just generating a new
Constant zero and so rsb instructions are now generated.
Original commit message:
The shift amount operand can be provided in a general purpose
register so sink it. Flip the vdup and negate so the existing
patterns can be used for matching.
Differential Revision: https://reviews.llvm.org/D70841
The shift amount operand can be provided in a general purpose
register so sink it. Flip the vdup and negate so the existing
patterns can be used for matching.
Differential Revision: https://reviews.llvm.org/D70841
This has two main effects:
- Optimizes debug info size by saving 221.86 MB of obj file size in a
Windows optimized+debug build of 'all'. This is 3.03% of 7,332.7MB of
object file size.
- Incremental step towards decoupling target intrinsics.
The enums are still compact, so adding and removing a single
target-specific intrinsic will trigger a rebuild of all of LLVM.
Assigning distinct target id spaces is potential future work.
Part of PR34259
Reviewers: efriedma, echristo, MaskRay
Reviewed By: echristo, MaskRay
Differential Revision: https://reviews.llvm.org/D71320
Soon Intrinsic::ID will be a plain integer, so this overload will not be
possible.
Rename both overloads to ensure that downstream targets observe this as
a build failure instead of a runtime failure.
Split off from D71320
Reviewers: efriedma
Differential Revision: https://reviews.llvm.org/D71381
After creating a low-overhead loop, the loop update instruction was still
lingering around hurting performance. This removes dead loop update
instructions, which in our case are mostly SUBS instructions.
To support this, some helper functions were added to MachineLoopUtils and
ReachingDefAnalysis to analyse live-ins of loop exit blocks and find uses
before a particular loop instruction, respectively.
This is a first version that removes a SUBS instruction when there are no other
uses inside and outside the loop block, but there are some more interesting
cases in test/CodeGen/Thumb2/LowOverheadLoops/mve-tail-data-types.ll which
shows that there is room for improvement. For example, we can't handle this
case yet:
..
dlstp.32 lr, r2
.LBB0_1:
mov r3, r2
subs r2, #4
vldrh.u32 q2, [r1], #8
vmov q1, q0
vmla.u32 q0, q2, r0
letp lr, .LBB0_1
@ %bb.2:
vctp.32 r3
..
which is a lot more tricky because r2 is not only used by the subs, but also by
the mov to r3, which is used outside the low-overhead loop by the vctp
instruction, and that requires a bit of a different approach, and I will follow
up on this.
Differential Revision: https://reviews.llvm.org/D71007
This adds the family of `vshlq_n` and `vshrq_n` ACLE intrinsics, which
shift every lane of a vector left or right by a compile-time
immediate. They mostly work by expanding to the IR `shl`, `lshr` and
`ashr` operations, with their second operand being a vector splat of
the immediate.
There's a fiddly special case, though. ACLE specifies that the
immediate in `vshrq_n` can take values up to //and including// the bit
size of the vector lane. But LLVM IR thinks that shifting right by the
full size of the lane is UB, and feels free to replace the `lshr` with
an `undef` half way through the optimization pipeline. Hence, to keep
this legal in source code, I have to detect it at codegen time.
Logical (unsigned) right shifts by the element size are handled by
simply emitting the zero vector; arithmetic ones are converted into a
shift of one bit less, which will always give the same output.
In order to do that check, I also had to enhance the tablegen
MveEmitter so that it can cope with converting a builtin function's
operand into a bare integer to pass to a code-generating subfunction.
Previously the only bare integers it knew how to handle were flags
generated from within `arm_mve.td`.
Reviewers: dmgreen, miyuki, MarkMurrayARM, ostannard
Reviewed By: dmgreen, MarkMurrayARM
Subscribers: echristo, hokein, rdhindsa, kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71065
Summary:
This patch refactors instruction selection of the complex vector
addition, multiplication and multiply-add intrinsics, so that it is
now based on TableGen patterns rather than C++ code.
It also changes the first parameter (halving vs non-halving) of the
arm_mve_vcaddq IR intrinsic to match the corresponding instruction
encoding, hence it requires some changes in the tests.
The patch addresses David's comment in https://reviews.llvm.org/D71190
Reviewers: dmgreen, ostannard, simon_tatham, MarkMurrayARM
Reviewed By: dmgreen
Subscribers: merge_guards_bot, kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71245
I rewrote the isel tablegen for MVE immediate shifts, and accidentally
removed the `let Predicates=[HasMVEInt]` that was wrapping the old
version, which seems to have allowed those rules to cause trouble on
non-MVE targets. That's what I get for only re-running the MVE tests.
Summary:
This adds the family of `vshlq_n` and `vshrq_n` ACLE intrinsics, which
shift every lane of a vector left or right by a compile-time
immediate. They mostly work by expanding to the IR `shl`, `lshr` and
`ashr` operations, with their second operand being a vector splat of
the immediate.
There's a fiddly special case, though. ACLE specifies that the
immediate in `vshrq_n` can take values up to //and including// the bit
size of the vector lane. But LLVM IR thinks that shifting right by the
full size of the lane is UB, and feels free to replace the `lshr` with
an `undef` half way through the optimization pipeline. Hence, to keep
this legal in source code, I have to detect it at codegen time.
Logical (unsigned) right shifts by the element size are handled by
simply emitting the zero vector; arithmetic ones are converted into a
shift of one bit less, which will always give the same output.
In order to do that check, I also had to enhance the tablegen
MveEmitter so that it can cope with converting a builtin function's
operand into a bare integer to pass to a code-generating subfunction.
Previously the only bare integers it knew how to handle were flags
generated from within `arm_mve.td`.
Reviewers: dmgreen, miyuki, MarkMurrayARM, ostannard
Reviewed By: MarkMurrayARM
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71065
Summary:
This patch adds intrinsics for the following MVE instructions:
* VCADD, VHCADD
* VCMUL
* VCMLA
Each of the above 3 groups has a corresponding new LLVM IR intrinsic.
Reviewers: simon_tatham, MarkMurrayARM, ostannard, dmgreen
Reviewed By: MarkMurrayARM
Subscribers: merge_guards_bot, kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71190
With the extra optimisations we have done, these should now be fine to
enable by default. Which is what this patch does.
Differential Revision: https://reviews.llvm.org/D70968
This attempts to teach the cost model in Arm that code such as:
%s = shl i32 %a, 3
%a = and i32 %s, %b
Can under Arm or Thumb2 become:
and r0, r1, r2, lsl #3
So the cost of the shift can essentially be free. To do this without
trying to artificially adjust the cost of the "and" instruction, it
needs to get the users of the shl and check if they are a type of
instruction that the shift can be folded into. And so it needs to have
access to the actual instruction in getArithmeticInstrCost, which if
available is added as an extra parameter much like getCastInstrCost.
We otherwise limit it to shifts with a single user, which should
hopefully handle most of the cases. The list of instruction that the
shift can be folded into include ADC, ADD, AND, BIC, CMP, EOR, MVN, ORR,
ORN, RSB, SBC and SUB. This translates to Add, Sub, And, Or, Xor and
ICmp.
Differential Revision: https://reviews.llvm.org/D70966
This adds some extra cost model tests for shifts, and does some minor
adjustments to some Neon code to make it clear as to what it applies to.
Both NFC.
Summary:
Currently the describeLoadedValue() hook is assumed to describe the
value of the instruction's first explicit define. The hook will not be
called for instructions with more than one explicit define.
This commit adds a register parameter to the describeLoadedValue() hook,
and invokes the hook for all registers in the worklist.
This will allow us to for example describe instructions which produce
more than two parameters' values; e.g. Hexagon's various combine
instructions.
This also fixes situations in our downstream target where we may pass
smaller parameters in the high part of a register. If such a parameter's
value is produced by a larger copy instruction, we can't describe the
call site value using the super-register, and we instead need to know
which sub-register that should be used.
This also allows us to handle cases like this:
$ebx = [...]
$rdi = MOVSX64rr32 $ebx
$esi = MOV32rr $edi
CALL64pcrel32 @call
The hook will first be invoked for the MOV32rr instruction, which will
say that @call's second parameter (passed in $esi) is described by $edi.
As $edi is not preserved it will be added to the worklist. When we get
to the MOVSX64rr32 instruction, we need to describe two values; the
sign-extended value of $ebx -> $rdi for the first parameter, and $ebx ->
$edi for the second parameter, which is now possible.
This commit modifies the dbgcall-site-lea-interpretation.mir test case.
In the test case, the values of some 32-bit parameters were produced
with LEA64r. Perhaps we can in general cases handle such by emitting
expressions that AND out the lower 32-bits, but I have not been able to
land in a case where a LEA64r is used for a 32-bit parameter instead of
LEA64_32 from C code.
I have not found a case where it would be useful to describe parameters
using implicit defines, so in this patch the hook is still only invoked
for explicit defines of forwarding registers.
Reviewers: djtodoro, NikolaPrica, aprantl, vsk
Reviewed By: djtodoro, vsk
Subscribers: ormris, hiraditya, llvm-commits
Tags: #debug-info, #llvm
Differential Revision: https://reviews.llvm.org/D70431
Currently the describeLoadedValue() hook is assumed to describe the
value of the instruction's first explicit define. The hook will not be
called for instructions with more than one explicit define.
This commit adds a register parameter to the describeLoadedValue() hook,
and invokes the hook for all registers in the worklist.
This will allow us to for example describe instructions which produce
more than two parameters' values; e.g. Hexagon's various combine
instructions.
This also fixes a case in our downstream target where we may pass
smaller parameters in the high part of a register. If such a parameter's
value is produced by a larger copy instruction, we can't describe the
call site value using the super-register, and we instead need to know
which sub-register that should be used.
This also allows us to handle cases like this:
$ebx = [...]
$rdi = MOVSX64rr32 $ebx
$esi = MOV32rr $edi
CALL64pcrel32 @call
The hook will first be invoked for the MOV32rr instruction, which will
say that @call's second parameter (passed in $esi) is described by $edi.
As $edi is not preserved it will be added to the worklist. When we get
to the MOVSX64rr32 instruction, we need to describe two values; the
sign-extended value of $ebx -> $rdi for the first parameter, and $ebx ->
$edi for the second parameter, which is now possible.
This commit modifies the dbgcall-site-lea-interpretation.mir test case.
In the test case, the values of some 32-bit parameters were produced
with LEA64r. Perhaps we can in general cases handle such by emitting
expressions that AND out the lower 32-bits, but I have not been able to
land in a case where a LEA64r is used for a 32-bit parameter instead of
LEA64_32 from C code.
I have not found a case where it would be useful to describe parameters
using implicit defines, so in this patch the hook is still only invoked
for explicit defines of forwarding registers.
MVE doesn't have the range of shuffle instructions available in Neon. We
also cannot use the trick of cutting a difficult vector shuffle in half
to simplify things. Instead we need to be more careful about how we
lower shuffles.
This patch adds an extra combine that attempts to find "whole lane"
vmovs when lowering shuffles of smaller types. This helps us make some
shuffles a lot simpler, generating single lane movs for the parts that
can make use of it, falling back to the original shuffle for the rest.
Differential Revision: https://reviews.llvm.org/D69509
Alas, using half the available vector registers in a single instruction
is just too much for the register allocator to handle. The mve-vldst4.ll
test here fails when these instructions are enabled at present. This
patch disables the generation of VLD4 and VST4 by adding a
mve-max-interleave-factor option, which we currently default to 2.
Differential Revision: https://reviews.llvm.org/D71109
Summary:
The immediate forms of the MVE VQSHL instruction have MC names like
`MVE_VSLIimms8` and `MVE_VSLIimmu32`. Those names are confusing,
because VSLI is a completely different shift instruction with no
semantic relation to VQSHL. But it just happens to be defined
immediately before VQSHL in `ARMInstrMVE.td`, so this looks like a
copy-paste error. Renamed the ids to match the instruction name.
Reviewers: ostannard, dmgreen, MarkMurrayARM, miyuki
Reviewed By: miyuki
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71114
Convert ARMCodeGenPrepare into a generic type promotion pass by:
- Removing the insertion of arm specific intrinsics to handle narrow
types as we weren't using this.
- Removing ARMSubtarget references.
- Now query a generic TLI object to know which types should be
promoted and what they should be promoted to.
- Move all codegen tests into Transforms folder and testing using opt
and not llc, which is how they should have been written in the
first place...
The pass searches up from icmp operands in an attempt to safely
promote types so we can avoid generating unnecessary unsigned extends
during DAG ISel.
Differential Revision: https://reviews.llvm.org/D69556
Similar to the parent, this adds some constants to tablegen to replace
the existing magic values.
Differential Revision: https://reviews.llvm.org/D70825
I got tired of looking at magic constants in tablegen files. This adds
condition codes like ARMCCeq and makes use of them.
I also removed the extra patterns for reverse condition codes from
D70296, they should now be covered by the parent commit.
Differential Revision: https://reviews.llvm.org/D70824
The VCMP instructions in MVE can accept a register or ZR, but only as
the right hand operator. Most of the time this will already be correct
because the icmp will have been canonicalised that way already. There
are some cases in the lowering of float conditions that this will not
apply to though. This code should fix up those cases.
Differential Revision: https://reviews.llvm.org/D70822
Summary:
This commit adds the `vpselq` intrinsics which take an MVE predicate
word and select lanes from two vectors; the `vctp` intrinsics which
create a tail predicate word suitable for processing the first m
elements of a vector (e.g. in the last iteration of a loop); and
`vpnot`, which simply complements a predicate word and is just
syntactic sugar for the `~` operator.
The `vctp` ACLE intrinsics are lowered to the IR intrinsics we've
already added (and which D70592 just reorganized). I've filled in the
missing isel rule for VCTP64, and added another set of rules to
generate the predicated forms.
I needed one small tweak in MveEmitter to allow the `unpromoted` type
modifier to apply to predicates as well as integers, so that `vpnot`
doesn't pointlessly convert its input integer to an `<n x i1>` before
complementing it.
Reviewers: ostannard, MarkMurrayARM, dmgreen
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D70485
Summary:
D65884 added a set of Arm IR intrinsics for the MVE VCTP instruction,
to use in tail predication. But the 64-bit one doesn't work properly:
its predicate type is `<2 x i1>` / `v2i1`, which isn't a legal MVE
type (due to not having a full set of instructions that manipulate it
usefully). The test of `vctp64` in `basic-tail-pred.ll` goes through
`opt` fine, as the test expects, but if you then feed it to `llc` it
causes a type legality failure at isel time.
The usual workaround we've been using in the rest of the MVE
intrinsics family is to bodge `v2i1` into `v4i1`. So I've adjusted the
`vctp64` IR intrinsic to do that, and completely removed the code (and
test) that uses that intrinsic for 64-bit tail predication. That will
allow me to add isel rules (upcoming in D70485) that actually generate
the VCTP64 instruction.
Also renamed all four of these IR intrinsics so that they have `mve`
in the name, since its absence was confusing.
Reviewers: ostannard, MarkMurrayARM, dmgreen
Reviewed By: MarkMurrayARM
Subscribers: samparker, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70592
Summary:
Add support for vcadd_* family of intrinsics. This set of intrinsics is
available in Armv8.3-A.
The fp16 versions require the FP16 extension, which has been available
(opt-in) since Armv8.2-A.
Reviewers: t.p.northover
Reviewed By: t.p.northover
Subscribers: t.p.northover, kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D70862
These instructions do not work quite like I expected them to. They
perform the addition and then shift in a higher precision integer, so do
not match up with the patterns that we added.
For example with s8s, adding 100 and 100 should wrap leaving the shift
to work on a negative number. VHADD will instead do the arithmetic in
higher precision, giving 100 overall. The vhadd gives a "better" result,
but not one that matches up with the input.
I am just removing the patterns here. We might be able to re-add them in
the future by checking for wrap flags or changing bitwidths. But for the
moment just remove them to remove the problem cases.
Summary:
In the cases where the CMOV (f16) SDNode is used with condition codes
LT, LE, VC or NE, it is successfully selected into a VSEL instruction.
In the remaining cases, however, instruction selection fails since VSEL
does not support other condition codes.
This patch handles such cases by using the single-precision version of
the VMOV instruction.
Reviewers: ostannard, dmgreen
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70667
This replaces the A32 NEON vqadds, vqaddu, vqsubs and vqsubu intrinsics
with the target independent sadd_sat, uadd_sat, ssub_sat and usub_sat.
This helps generate vqadds from standard IR nodes, which might be
produced from the vectoriser. The old variants are removed in the
process.
Differential Revision: https://reviews.llvm.org/D69350
Some of these patterns have grown quite organically. I've tried to
organise them a little here, moving all the PatFlags together and giving
them a more consistent naming scheme, to allow some of the later
patterns to be merged into a single multiclass.
Differential Revision: https://reviews.llvm.org/D70178
MVE has a basic symmetry between it's normal loads/store operations and
the masked variants. This means that masked loads and stores can use
pre-inc and post-inc addressing modes, just like the standard loads and
stores already do.
To enable that, this patch adds all the relevant infrastructure for
treating masked loads/stores addressing modes in the same way as normal
loads/stores.
This involves:
- Adding an AddressingMode to MaskedLoadStoreSDNode, along with an extra
Offset operand that is added after the PtrBase.
- Extending the IndexedModeActions from 8bits to 16bits to store the
legality of masked operations as well as normal ones. This array is
fairly small, so doubling the size still won't make it very large.
Offset masked loads can then be controlled with
setIndexedMaskedLoadAction, similar to standard loads.
- The same methods that combine to indexed loads, such as
CombineToPostIndexedLoadStore, are adjusted to handle masked loads in
the same way.
- The ARM backend is then adjusted to make use of these indexed masked
loads/stores.
- The X86 backend is adjusted to hopefully be no functional changes.
Differential Revision: https://reviews.llvm.org/D70176
Add some more helper functions to ReachingDefs to query the uses of
a given MachineInstr and also to query whether two MachineInstrs use
the same def of a register.
For Arm, while tail-predicating, these helpers are used in the
low-overhead loops to remove the dead code that calculates the number
of loop iterations.
Differential Revision: https://reviews.llvm.org/D70240
Add several new methods to ReachingDefAnalysis:
- getReachingMIDef, instead of returning an integer, return the
MachineInstr that produces the def.
- getInstFromId, return a MachineInstr for which the given integer
corresponds to.
- hasSameReachingDef, return whether two MachineInstr use the same
def of a register.
- isRegUsedAfter, return whether a register is used after a given
MachineInstr.
These methods have been used in ARMLowOverhead to replace searching
for uses/defs.
Differential Revision: https://reviews.llvm.org/D70009
When inserting a non-decrementing LE, the basic block was being
resized to take into consideration that a tCMP and tBcc had been
combined into one T1 instruction. This is not true in the LE case
where we generate a CBN?Z and an LE.
Differential Revision: https://reviews.llvm.org/D70536
This patch adds instruction selection patterns for the TT, TTT, TTA, and TTAT
instructions and tests for llvm.arm.cmse.tt, llvm.arm.cmse.ttt,
llvm.arm.cmse.tta, and llvm.arm.cmse.ttat intrinsics (added in a previous
patch).
Patch by Javed Absar.
Differential Revision: https://reviews.llvm.org/D70407
Adds a pattern to ARMInstrMVE.td to use a VQNEG
instruction if an equivalent multi-instruction
construct is found.
Differential Revision: https://reviews.llvm.org/D70491
Summary:
Most libraries are defined in the lib/ directory but there are also a
few libraries defined in tools/ e.g. libLLVM, libLTO. I'm defining
"Component Libraries" as libraries defined in lib/ that may be included in
libLLVM.so. Explicitly marking the libraries in lib/ as component
libraries allows us to remove some fragile checks that attempt to
differentiate between lib/ libraries and tools/ libraires:
1. In tools/llvm-shlib, because
llvm_map_components_to_libnames(LIB_NAMES "all") returned a list of
all libraries defined in the whole project, there was custom code
needed to filter out libraries defined in tools/, none of which should
be included in libLLVM.so. This code assumed that any library
defined as static was from lib/ and everything else should be
excluded.
With this change, llvm_map_components_to_libnames(LIB_NAMES, "all")
only returns libraries that have been added to the LLVM_COMPONENT_LIBS
global cmake property, so this custom filtering logic can be removed.
Doing this also fixes the build with BUILD_SHARED_LIBS=ON
and LLVM_BUILD_LLVM_DYLIB=ON.
2. There was some code in llvm_add_library that assumed that
libraries defined in lib/ would not have LLVM_LINK_COMPONENTS or
ARG_LINK_COMPONENTS set. This is only true because libraries
defined lib lib/ use LLVMBuild.txt and don't set these values.
This code has been fixed now to check if the library has been
explicitly marked as a component library, which should now make it
easier to remove LLVMBuild at some point in the future.
I have tested this patch on Windows, MacOS and Linux with release builds
and the following combinations of CMake options:
- "" (No options)
- -DLLVM_BUILD_LLVM_DYLIB=ON
- -DLLVM_LINK_LLVM_DYLIB=ON
- -DBUILD_SHARED_LIBS=ON
- -DBUILD_SHARED_LIBS=ON -DLLVM_BUILD_LLVM_DYLIB=ON
- -DBUILD_SHARED_LIBS=ON -DLLVM_LINK_LLVM_DYLIB=ON
Reviewers: beanz, smeenai, compnerd, phosek
Reviewed By: beanz
Subscribers: wuzish, jholewinski, arsenm, dschuff, jyknight, dylanmckay, sdardis, nemanjai, jvesely, nhaehnle, mgorny, mehdi_amini, sbc100, jgravelle-google, hiraditya, aheejin, fedor.sergeev, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, steven_wu, rogfer01, MartinMosbeck, brucehoult, the_o, dexonsmith, PkmX, jocewei, jsji, dang, Jim, lenary, s.egerton, pzheng, sameer.abuasal, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70179
Adds a pattern to ARMInstrMVE.td to use a VQABS
instruction if an equivalent multi-instruction
construct is found.
Differential revision: https://reviews.llvm.org/D70181
Now that we have the intrinsics, we can add VLD2/4 and VST2/4 lowering
for MVE. This works the same way as Neon, recognising the load/shuffles
combination and converting them into intrinsics in a pre-isel pass,
which just calls getMaxSupportedInterleaveFactor, lowerInterleavedLoad
and lowerInterleavedStore.
The main difference to Neon is that we do not have a VLD3 instruction.
Otherwise most of the code works very similarly, with just some minor
differences in the form of the intrinsics to work around. VLD3 is
disabled by making isLegalInterleavedAccessType return false for those
cases.
We may need some other future adjustments, such as VLD4 take up half the
available registers so should maybe cost more. This patch should get the
basics in though.
Differential Revision: https://reviews.llvm.org/D69392
This fills in the small family of MVE intrinsics that have nothing to
do with vectors: they implement bit-shift operations on 32- or 64-bit
values held in one or two general-purpose registers. Most of these
shift operations saturate if shifting left, and round to nearest if
shifting right, although LSLL and ASRL behave like ordinary shifts.
When these instructions take a variable shift count in a register,
they pay attention to its sign, so that (for example) LSLL or UQRSHLL
will shift left if given a positive number but right if given a
negative one. That makes even LSLL and ASRL different enough from
standard LLVM IR shift semantics that I couldn't see any better
alternative than to simply model the whole family as a set of
MVE-specific IR intrinsics.
(The //immediate// forms of LSLL and ASRL, on the other hand, do
behave exactly like a standard IR shift of a 64-bit value. In fact,
those forms don't have ACLE intrinsics defined at all, because you can
just write an ordinary C shift operation if you want one of those.)
The 64-bit shifts have to be instruction-selected in C++, because they
deliver two output values. But the 32-bit ones are simple enough that
I could write a DAG isel pattern directly into each Instruction
record.
Reviewers: ostannard, MarkMurrayARM, dmgreen
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D70319
AMDGPU needs to know the FP mode for the function to answer this
correctly when this is removed from the subtarget.
AArch64 had to make this more complicated by using this from an IR
hook, so add an IR typed overload.
Remove the restriction, from the mve tail predication pass, that the
all masked vectors instructions need to be 128-bits. This allows us
to supported extending loads and truncating stores.
Differential Revision: https://reviews.llvm.org/D69946
This patch modifies ARMLowOverheadLoops to convert a predicated
vector low-overhead loop into a tail-predicatd one. This is currently
a very basic conversion, with the following restrictions:
- Operates only on single block loops.
- The loop can only contain a single vctp instruction.
- No other instructions can write to the vpr.
- We only allow a subset of the mve instructions in the loop.
TODO: Pass the number of elements, not the number of iterations to
dlstp/wlstp.
Differential Revision: https://reviews.llvm.org/D69945
* Implements scalable size queries for MVTs, split out from D53137.
* Contains a fix for FindMemType to avoid using scalable vector type
to contain non-scalable types.
* Explicit casts for several places where implicit integer sign
changes or promotion from 32 to 64 bits caused problems.
* CodeGenDAGPatterns will treat scalable and non-scalable vector types
as different.
Reviewers: greened, cameron.mcinally, sdesmalen, rovka
Reviewed By: rovka
Differential Revision: https://reviews.llvm.org/D66871
Provides support for using r6-r11 as globally scoped
register variables. This requires a -ffixed-rN flag
in order to reserve rN against general allocation.
If for a given GRV declaration the corresponding flag
is not found, or the the register in question is the
target's FP, we fail with a diagnostic.
Differential Revision: https://reviews.llvm.org/D68862
In MCObjectStreamer, when there is no current fragment, initially
symbols are created in a "pending" state and assigned to a dummy
empty fragment.
Previously, they were not being assigned an offset, and thus
evaluateAbsolute would fail if trying to evaluate an expression 'a -
b', where both 'a' and 'b' were in this pending state.
Also slightly refactored the EmitLabel overload which takes an
MCFragment for clarity.
Fixes: https://llvm.org/PR41825
Differential Revision: https://reviews.llvm.org/D70062
Summary:
As well as vector/vector compare instructions, MVE also has a family
of comparisons taking a vector and a scalar, which compare every lane
of the vector against the same value. We generate those at isel time
using isel patterns that match `(ARMvcmp vector, (ARMvdup scalar))`.
This commit adds corresponding patterns for the operand-reversed form
`(ARMvcmp (ARMvdup scalar), vector)`, with condition codes swapped as
necessary. That way, we can still generate the vector/scalar compare
instruction if the IR happens to have been rearranged to put the
operands the other way round, which can happen in some optimization
phases. Previously, a vcmp the other way round was handled by emitting
a `vdup` instruction to //explicitly// replicate the scalar input into
a vector, and then doing a vector/vector comparison.
I haven't added a new test, because it turned out that several
existing tests were already exhibiting that failure mode. So just
updating the expected output in the existing MVE codegen tests
demonstrates what's been improved.
Reviewers: ostannard, MarkMurrayARM, dmgreen
Reviewed By: dmgreen
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70296
This is a follow up of d90804d, to also flag fmcp instructions as instructions
that we do not support in tail-predicated vector loops.
Differential Revision: https://reviews.llvm.org/D70295
During register coalescing, we update the live-intervals on-the-fly.
To do that we are in this strange mode where the live-intervals can
be slightly out-of-sync (more precisely they are forward looking)
compared to what the IR actually represents.
This happens because the register coalescer only updates the IR when
it is done with updating the live-intervals and it has to do it this
way because updating the IR on-the-fly would actually clobber some
information on how the live-ranges that are being updated look like.
This is problematic for updates that rely on the IR to accurately
represents the state of the live-ranges. Right now, we have only
one of those: stripValuesNotDefiningMask.
To reconcile this need of out-of-sync IR, this patch introduces a
new argument to LiveInterval::refineSubRanges that allows the code
doing the live range updates to reason about how the code should
look like after the coalescer will have rewritten the registers.
Essentially this captures how a subregister index with be offseted
to match its position in a new register class.
E.g., let say we want to merge:
V1.sub1:<2 x s32> = COPY V2.sub3:<4 x s32>
We do that by choosing a class where sub1:<2 x s32> and sub3:<4 x s32>
overlap, i.e., by choosing a class where we can find "offset + 1 == 3".
Put differently we align V2's sub3 with V1's sub1:
V2: sub0 sub1 sub2 sub3
V1: <offset> sub0 sub1
This offset will look like a composed subregidx in the the class:
V1.(composed sub2 with sub1):<4 x s32> = COPY V2.sub3:<4 x s32>
=> V1.(composed sub2 with sub1):<4 x s32> = COPY V2.sub3:<4 x s32>
Now if we didn't rewrite the uses and def of V1, all the checks for V1
need to account for this offset to match what the live intervals intend
to capture.
Prior to this patch, we would fail to recognize the uses and def of V1
and would end up with machine verifier errors: No live segment at def.
This could lead to miscompile as we would drop some live-ranges and
thus, miss some interferences.
For this problem to trigger, we need to reach stripValuesNotDefiningMask
while having a mismatch between the IR and the live-ranges (i.e.,
we have to apply a subreg offset to the IR.)
This requires the following three conditions:
1. An update of overlapping subreg lanes: e.g., dsub0 == <ssub0, ssub1>
2. An update with Tuple registers with a possibility to coalesce the
subreg index: e.g., v1.dsub_1 == v2.dsub_3
3. Subreg liveness enabled.
looking at the IR to decide what is alive and what is not, i.e., calling
stripValuesNotDefiningMask.
coalescer maintains for the live-ranges information.
None of the targets that currently use subreg liveness (i.e., the targets
that fulfill #3, Hexagon, AMDGPU, PowerPC, and SystemZ IIRC) expose #1 and
and #2, so this patch also artificial enables subreg liveness for ARM,
so that a nice test case can be attached.
This implements TTI hook 'preferPredicateOverEpilogue' for MVE. This is a
first version and it operates on single block loops only. With this change, the
vectoriser will now determine if tail-folding scalar remainder loops is
possible/desired, which is the first step to generate MVE tail-predicated
vector loops.
This is disabled by default for now. I.e,, this is depends on option
-disable-mve-tail-predication, which is off by default.
I will follow up on this soon with a patch for the vectoriser to respect loop
hint 'vectorize.predicate.enable'. I.e., with this loop hint set to Disabled,
we don't want to tail-fold and we shouldn't query this TTI hook, which is
done in D70125.
Differential Revision: https://reviews.llvm.org/D69845
MVE includes instructions that extract an 8- or 16-bit lane from a
vector and sign-extend it into the output 32-bit GPR. `ARMInstrMVE.td`
already included isel patterns to select those instructions in
response to the `ARMISD::VGETLANEs` selection-DAG node type. But
`ARMISD::VGETLANEs` was never actually generated, because the code
that creates it was conditioned on NEON only.
It's an easy fix to enable the same code for integer MVE, and now IR
that sign-extends the result of an extractelement (whether explicitly
or as part of the function call ABI) will use `vmov.s8` instead of
`vmov.u8` followed by `sxtb`.
Reviewers: SjoerdMeijer, dmgreen, ostannard
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70132
These relocations are specified by the ARM EHABI (section 6.3). As I understand
it, their purpose is to accommodate unwinder implementations that wish to
reduce code size by placing the implementations of the compact unwinding
decoders in a separate translation unit, and using extern weak symbols to
refer to them from the main unwinder implementation, so that they are only
linked when something in the binary needs them in order to unwind.
However, neither of the unwinders used on Android (libgcc, LLVM libunwind)
use this technique, and in fact emitting these relocations ends up being
counterproductive to code size because they cause a copy of the unwinder
to be statically linked into most binaries, regardless of whether it is
actually needed. Furthermore, these relocations create circular dependencies
(between libc and the unwinder) in cases where the unwinder is dynamically
linked and libc contains compact unwind info.
Therefore, deviate from the EHABI here and stop emitting these relocations
on Android.
Differential Revision: https://reviews.llvm.org/D70027
We had some code for this for 32-bit ARM, but this doesn't really need
to be in target-specific code; generalize it.
(I think this started showing up recently because we added an
optimization that converts pow to powi.)
Differential Revision: https://reviews.llvm.org/D69013
Refactor usage of isCopyInstrImpl, isCopyInstr and isAddImmediate methods
to return optional machine operand pair of destination and source
registers.
Patch by Nikola Prica
Differential Revision: https://reviews.llvm.org/D69622
We have two ways to steer creating a predicated vector body over creating a
scalar epilogue. To force this, we have 1) a command line option and 2) a
pragma available. This adds a third: a target hook to TargetTransformInfo that
can be queried whether predication is preferred or not, which allows the
vectoriser to make the decision without forcing it.
While this change behaves as a non-functional change for now, it shows the
required TTI plumbing, usage of this new hook in the vectoriser, and the
beginning of an ARM MVE implementation. I will follow up on this with:
- a complete MVE implementation, see D69845.
- a patch to disable this, i.e. we should respect "vector_predicate(disable)"
and its corresponding loophint.
Differential Revision: https://reviews.llvm.org/D69040
This patch adds two new families of intrinsics, both of which are
memory accesses taking a vector of locations to load from / store to.
The vldrq_gather_base / vstrq_scatter_base intrinsics take a vector of
base addresses, and an immediate offset to be added consistently to
each one. vldrq_gather_offset / vstrq_scatter_offset take a scalar
base address, and a vector of offsets to add to it. The
'shifted_offset' variants also multiply each offset by the element
size type, so that the vector is effectively of array indices.
At the IR level, these operations are represented by a single set of
four IR intrinsics: {gather,scatter} × {base,offset}. The other
details (signed/unsigned, shift, and memory element size as opposed to
vector element size) are all specified by IR intrinsic polymorphism
and immediate operands, because that made the selection job easier
than making a huge family of similarly named intrinsics.
I considered using the standard IR representations such as
llvm.masked.gather, but they're not a good fit. In order to use
llvm.masked.gather to represent a gather_offset load with element size
smaller than a pointer, you'd have to expand the <8 x i16> vector of
offsets into an <8 x i16*> vector of pointers, which would be split up
during legalization, so you'd spend most of your time undoing the mess
it had made. Also, ISel support for llvm.masked.gather would be easy
enough in a trivial way (you can expand it into a gather-base load
with a zero immediate offset), but instruction-selecting lots of
fiddly idioms back into all the _other_ MVE load instructions would be
much more work. So I think dedicated IR intrinsics are the more
sensible approach, at least for the moment.
On the clang tablegen side, I've added two new features to the
Tablegen source accepted by MveEmitter: a 'CopyKind' type node for
defining a type that varies with the parameter type (it lets you ask
for an unsigned integer type of the same width as the parameter), and
an 'unsignedflag' value node for passing an immediate IR operand which
is 0 for a signed integer type or 1 for an unsigned one. That lets me
write each kind of intrinsic just once and get all its subtypes and
immediate arguments generated automatically.
Also I've tweaked the handling of pointer-typed values in the code
generation part of MveEmitter: they're generated as Address rather
than Value (i.e. including an alignment) so that they can be given to
the ordinary IR load and store operations, but I'd omitted the code to
convert them back to Value when they're going to be used as an
argument to an IR intrinsic.
On the MC side, I've enhanced MVEVectorVTInfo so that it can tell you
not only the full assembly-language suffix for a given vector type
(like 's32' or 'u16') but also the numeric-only one used by store
instructions (just '32' or '16').
Reviewers: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D69791
Summary:
G_GEP is rather poorly named. It's a simple pointer+scalar addition and
doesn't support any of the complexities of getelementptr. I therefore
propose that we rename it. There's a G_PTR_MASK so let's follow that
convention and go with G_PTR_ADD
Reviewers: volkan, aditya_nandakumar, bogner, rovka, arsenm
Subscribers: sdardis, jvesely, wdng, nhaehnle, hiraditya, jrtc27, atanasyan, arphaman, Petar.Avramovic, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69734
This feature controls whether AA is used into the backend, and was
previously turned on for certain subtargets to help create less
constrained scheduling graphs. This patch turns it on for all
subtargets, so that they can all make use of the extra information to
produce better code.
Differential Revision: https://reviews.llvm.org/D69796
In the ARM backend, for historical reasons we have only some targets
using Machine Scheduling. The rest use the old list scheduler as they
are using itinaries and the list scheduler seems to produce better code
(and not crash running out of register on v6m codes). So whether to use
the MIScheduler or not is checked at runtime from the subtarget
features.
This is fine, except for post-ra scheduling. Whether to use the old
post-ra list scheduler or the post-ra machine schedule is decided as the
pass manager is set up, in arms case from a newly constructed subtarget.
Under some situations, like LTO, this won't include the correct cpu so
can pick the wrong option. This can have a surprising effect on
performance.
To fix that, this patch overrides targetSchedulesPostRAScheduling and
addPreSched2 in the ARM backend, adding _both_ post-ra schedulers and
picking at runtime which to execute. To pick between the two I've had to
add a enablePostRAMachineScheduler() method that normally returns
enableMachineScheduler() && enablePostRAScheduler(), which can be
overridden to enable just one of PostRAMachineScheduler vs
PostRAScheduler.
Thanks to David Penry for the identifying this problem.
Differential Revision: https://reviews.llvm.org/D69775
The Arm backend will usually return false for isFMAFasterThanFMulAndFAdd,
where both the fused VFMA.f32 and a non-fused VMLA.f32 are usually
available for scalar code. For MVE we don't have the non-fused version
though. It makes more sense for isFMAFasterThanFMulAndFAdd to return
true, allowing us to simplify some of the existing ISel patterns.
The tests here are that non of the existing tests failed, and so we are
still selecting VFMA and VFMS. The one test that changed shows we can
now select from fast math flags, as opposed to just relying on the
isFMADLegalForFAddFSub option.
Differential Revision: https://reviews.llvm.org/D69115
Refactor usage of isCopyInstrImpl, isCopyInstr and isAddImmediate methods
to return optional machine operand pair of destination and source
registers.
Patch by Nikola Prica
Differential Revision: https://reviews.llvm.org/D69622
Extend the describeLoadedValue() with support for target specific ARM and
AArch64 instructions interpretation. The patch provides specialization for
ADD and SUB operations that include a register and an immediate/offset
operand. Some of the instructions can operate with global string addresses
or constant pool indexes but such cases are omitted since we currently lack
flexible support for processing such operands at DWARF production stage.
Patch by Nikola Prica
Differential Revision: https://reviews.llvm.org/D67556
llvm/test/DebugInfo/MIR/X86/live-debug-values-reg-copy.mir failed with
EXPENSIVE_CHECKS enabled, causing the patch to be reverted in
rG2c496bb5309c972d59b11f05aee4782ddc087e71.
This patch relands the patch with a proper fix to the
live-debug-values-reg-copy.mir tests, by ensuring the MIR encodes the
callee-saves correctly so that the CalleeSaved info is taken from MIR
directly, rather than letting it be recalculated by the PEI pass. I've
done this by running `llc -stop-before=prologepilog` on the LLVM
IR as captured in the test files, adding the extra MOV instructions
that were manually added in the original test file, then running `llc
-run-pass=prologepilog` and finally re-added the comments for the MOV
instructions.
The static functions `positiveOffsetOpcode`, `negativeOffsetOpcode` and
`immediateOffsetOpcode` (lib/Target/ARM/Thumb2InstrInfo.cpp) currently can
return `0` as default opcode which is meaningless in this situation.
This patch replaces this default value by llvm_unreachable.
Reviewers: t.p.northover, tellenbach
Reviewed By: tellenbach
Subscribers: tellenbach, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69432
Patch By: Lorenzo Casalino <lorenzo.casalino93@gmail.com>
Summary:
A new function pass (Transforms/CFGuard/CFGuard.cpp) inserts CFGuard checks on
indirect function calls, using either the check mechanism (X86, ARM, AArch64) or
or the dispatch mechanism (X86-64). The check mechanism requires a new calling
convention for the supported targets. The dispatch mechanism adds the target as
an operand bundle, which is processed by SelectionDAG. Another pass
(CodeGen/CFGuardLongjmp.cpp) identifies and emits valid longjmp targets, as
required by /guard:cf. This feature is enabled using the `cfguard` CC1 option.
Reviewers: thakis, rnk, theraven, pcc
Subscribers: ychen, hans, metalcanine, dmajor, tomrittervg, alex, mehdi_amini, mgorny, javed.absar, kristof.beyls, hiraditya, steven_wu, dexonsmith, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D65761
Summary:
Writing support for three ACLE functions:
unsigned int __cls(uint32_t x)
unsigned int __clsl(unsigned long x)
unsigned int __clsll(uint64_t x)
CLS stands for "Count number of leading sign bits".
In AArch64, these two intrinsics can be translated into the 'cls'
instruction directly. In AArch32, on the other hand, this functionality
is achieved by implementing it in terms of clz (count number of leading
zeros).
Reviewers: compnerd
Reviewed By: compnerd
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D69250
The VST2 and VST4 instructions take two or four vector registers as
input, and store part of each register to memory in an interleaved
pattern. They come in variants indicating which part of each register
they store (VST20 and VST21; VST40 to VST43 inclusive); the intention
is that issuing each of those variants in turn has the combined effect
of loading or storing the whole set of registers to a memory block of
equal size. The corresponding VLD2 and VLD4 instructions load from
memory in the same interleaved format: each one overwrites only part
of its output register set, and again, the idea is that if you use
VLD4{0,1,2,3} or VLD2{0,1} together, you end up having written to the
whole of each register.
I've implemented the stores and loads quite differently. The loads
were easiest to implement as a single intrinsic that expands to all
four VLD4x instructions or both VLD2x, delivering four complete output
registers. (Implementing each individual load as a separate
instruction taking four input registers to partially overwrite is
possible in theory, but pointless, and when I tried it, I found it
would need extra work to get the register allocation not to be
horrible.) Since that intrinsic delivers multiple outputs, it has to
be instruction-selected in custom C++.
But the store instructions are easier to model individually, because
they don't overwrite any register at all and you can write a DAG Isel
pattern in Tablegen for each one.
Hence, my new intrinsic `int_arm_mve_vld4q` expands to four load
instructions, delivers four full output vectors, and is handled by C++
code, whereas `int_arm_mve_vst4q` expands to just one store
instruction, takes four input vectors and a constant indicating which
lanes to store, and is handled entirely in Tablegen. (And similarly
for vld2q/vst2q.) This is asymmetric, but it was the easiest way to do
each one.
Reviewers: dmgreen, miyuki, ostannard
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68700
This adds some initial example IR intrinsics for MVE instructions that
deliver multiple output values, and hence, have to be instruction-
selected by custom C++ code instead of Tablegen patterns.
I've added the writeback gather load instructions (taking a vector of
base addresses and a single common offset, returning a vector of
loaded values and an updated vector of base addresses); one example
from the long shift family (taking and returning a 64-bit value in two
GPRs); and the VADC instruction (which propagates a carry bit from
each vector-lane addition to the next, taking an input carry flag in
FPSCR and outputting the final one in FPSCR as well).
To support the VPT-predicated forms of these instructions, I've
written some helper functions to add the cluster of MVE predicate
operands to the end of a MachineInstr. `AddMVEPredicateToOps` is used
when the instruction actually is predicated (so it takes a predicate
mask argument), and `AddEmptyMVEPredicateToOps` is for when the
instruction is unpredicated (so it fills in $noreg for the mask). Each
one comes in a form suitable for `vpred_n`, and one for `vpred_r`
which takes the extra 'inactive' parameter.
For VADC, the representation of the carry flag in the IR intrinsic is
a word intended to be moved directly to and from `FPSCR_nzcvqc`, i.e.
with the carry flag in bit 29 of the word. (The user-facing ACLE
intrinsic will want it to be in bit 0, but I'll do that on the clang
side.)
Reviewers: dmgreen, miyuki, ostannard
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68699
This commit, together with the next few, will add a representative
sample of the kind of IR intrinsics that we'll need in order to
implement the user-facing ACLE intrinsics for MVE. Supporting all of
them will take more work; the intention of this initial series of
commits is to implement an intrinsic or two from lots of different
categories, as examples and proofs of concept.
This initial commit introduces a small number of IR intrinsics for
instructions simple enough that they can use Tablegen ISel patterns:
the predicated versions of the VADD and VSUB instructions (both
integer and FP), VMIN and VMAX, and the float->half VCVT instruction
(predicated and unpredicated).
When using VPT-predicated instructions in automatic code generation,
it will be convenient to specify the predicate value as a vector of
the appropriate number of i1. To make it easy to specify all sizes of
an instruction in one go and give each one the matching predicate
vector type, I've added a system of Tablegen informational records
describing MVE's vector types: each one gives the underlying LLVM IR
ValueType (which may not be the same if the MVE vector is of
explicitly signed or unsigned integers) and an appropriate vNi1 to use
as the predicate vector.
(Also, those info records include the usual encoding for the types, so
that as we add associations between each instruction encoding and one
of the new `MVEVectorVTInfo` records, we can remove some of the
existing template parameters and replace them with references to the
vector type info's fields.)
The user-facing ACLE intrinsics will receive a predicate mask as a
16-bit integer, so I've also provided a pair of intrinsics i2v and
v2i, to convert between an integer and a vector of i1 by just changing
the register class.
Reviewers: dmgreen, miyuki, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67158
MipsMCAsmInfo was using '$' prefix for Mips32 and '.L' for Mips64
regardless of -target-abi option. By passing MCTargetOptions to MCAsmInfo
we can find out Mips ABI and pick appropriate prefix.
Tags: #llvm, #clang, #lldb
Differential Revision: https://reviews.llvm.org/D66795
Commit message from D66935:
This patch fixes a bug exposed by D65653 where a subsequent invocation
of `determineCalleeSaves` ends up with a different size for the callee
save area, leading to different frame-offsets in debug information.
In the invocation by PEI, `determineCalleeSaves` tries to determine
whether it needs to spill an extra callee-saved register to get an
emergency spill slot. To do this, it calls 'estimateStackSize' and
manually adds the size of the callee-saves to this. PEI then allocates
the spill objects for the callee saves and the remaining frame layout
is calculated accordingly.
A second invocation in LiveDebugValues causes estimateStackSize to return
the size of the stack frame including the callee-saves. Given that the
size of the callee-saves is added to this, these callee-saves are counted
twice, which leads `determineCalleeSaves` to believe the stack has
become big enough to require spilling an extra callee-save as emergency
spillslot. It then updates CalleeSavedStackSize with a larger value.
Since CalleeSavedStackSize is used in the calculation of the frame
offset in getFrameIndexReference, this leads to incorrect offsets for
variables/locals when this information is recalculated after PEI.
This patch fixes the lldb unit tests in `functionalities/thread/concurrent_events/*`
Changes after D66935:
Ensures AArch64FunctionInfo::getCalleeSavedStackSize does not return
the uninitialized CalleeSavedStackSize when running `llc` on a specific
pass where the MIR code has already been expected to have gone through PEI.
Instead, getCalleeSavedStackSize (when passed the MachineFrameInfo) will try
to recalculate the CalleeSavedStackSize from the CalleeSavedInfo. In debug
mode, the compiler will assert the recalculated size equals the cached
size as calculated through a call to determineCalleeSaves.
This fixes two tests:
test/DebugInfo/AArch64/asan-stack-vars.mir
test/DebugInfo/AArch64/compiler-gen-bbs-livedebugvalues.mir
that otherwise fail when compiled using msan.
Reviewed By: omjavaid, efriedma
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68783
llvm-svn: 375425
This adds some new qdadd patterns to go along with the other recently added
qadd's.
Differential Revision: https://reviews.llvm.org/D68999
llvm-svn: 375414
This lowers a sadd_sat to a qadd by treating it as legal. Also adds qsub at the
same time.
The qadd instruction sets the q flag, but we already have many cases where we
do not model this in llvm.
Differential Revision: https://reviews.llvm.org/D68976
llvm-svn: 375411
Lower the target independent signed saturating intrinsics to qadd8 and qadd16.
This custom lowers them from a sadd_sat, catching the node early before it is
promoted. It also adds a QADD8b and QADD16b node to mean the bottom "lane" of a
qadd8/qadd16, so that we can call demand bits on it to show that it does not
use the upper bits.
Also handles QSUB8 and QSUB16.
Differential Revision: https://reviews.llvm.org/D68974
llvm-svn: 375402
Summary:
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: arsenm, dschuff, jyknight, sdardis, jvesely, nhaehnle, sbc100, jgravelle-google, hiraditya, aheejin, fedor.sergeev, jrtc27, atanasyan, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69216
llvm-svn: 375398
MachineInstr.h included AliasAnalysis.h, which includes a world of IR
constructs mostly unneeded in CodeGen. Prune it. Same for
DebugInfoMetadata.h.
Noticed with -ftime-trace.
llvm-svn: 375311
The default implementation of isIncomingArgumentHandler could lead
to generating incorrect code.
Make it a pure virtual method, so that targets know they have to
override it to produce correct code.
NFC
Differential Revision: https://reviews.llvm.org/D69187
llvm-svn: 375277
Allow us to generate truncating masked store which take v4i32 and
v8i16 vectors and can store to v4i8, v4i16 and v8i8 and memory.
Removed support for unaligned masked stores.
Differential Revision: https://reviews.llvm.org/D68461
llvm-svn: 375108
Add generic DAG combine for extending masked loads.
Allow us to generate sext/zext masked loads which can access v4i8,
v8i8 and v4i16 memory to produce v4i32, v8i16 and v4i32 respectively.
Differential Revision: https://reviews.llvm.org/D68337
llvm-svn: 375085
Summary:
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: jholewinski, arsenm, dschuff, jyknight, dylanmckay, sdardis, nemanjai, jvesely, nhaehnle, sbc100, jgravelle-google, hiraditya, aheejin, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, jsji, Jim, lenary, s.egerton, pzheng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68993
llvm-svn: 375084
Summary:
Currently Thumb2InstrInfo.cpp uses a register class which is
auto-generated by tablegen. Such approach is fragile because
auto-generated classes might change when other register classes are
added. For example, before https://reviews.llvm.org/D62667
we were using GPRPair_with_gsub_1_in_rGPRRegClass, but had to
change it to GPRPair_with_gsub_1_in_GPRwithAPSRnospRegClass
because the former class stopped being generated (this did not change
the functionality though).
This patch adds a register class consisting of even-odd GPR register
pairs from (R0, R1) to (R10, R11), which excludes (R12, SP) and uses
it in Thumb2InstrInfo.cpp instead of
GPRPair_with_gsub_1_in_GPRwithAPSRnospRegClass.
Reviewers: ostannard, simon_tatham, dmgreen, efriedma
Reviewed By: simon_tatham
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69026
llvm-svn: 374990
Instead of inserting everything after the 'root' of the reduction,
insert all instructions as close to their operands as possible. This
can help reduce register pressure.
Differential Revision: https://reviews.llvm.org/D67392
llvm-svn: 374981
Reverse the logic for valid tail predication instructions and create
a whitelist instead. Added other instruction groups that aren't
obviously safe:
- instructions that 'narrow' their result.
- lane moves.
- byte swapping instructions.
- interleaving loads and stores.
- cross-beat carries.
- top/bottom instructions.
- complex operations.
Hopefully we should be able to add more of these instructions to the
whitelist, once we have a more concrete idea of the transform.
Differential Revision: https://reviews.llvm.org/D67904
llvm-svn: 374887
Summary:
Integrated assembler does not accept offset expressions surrounded by
parenthesis. Handle this case for GAS compability.
https://bugs.llvm.org/show_bug.cgi?id=43631
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68764
llvm-svn: 374832
The adds both VMOVNt and VMOVNb instruction selection from the appropriate
shuffles. We detect shuffle masks of the form:
0, N, 2, N+2, 4, N+4, ...
or
0, N+1, 2, N+3, 4, N+5, ...
ISel will also try the opposite patterns, with inputs reversed. These are
selected to VMOVNt and VMOVNb respectively.
Differential Revision: https://reviews.llvm.org/D68283
llvm-svn: 374781
Add an extra parameter so the backend can take the alignment into
consideration.
Differential Revision: https://reviews.llvm.org/D68400
llvm-svn: 374763
In loop-vectorize, interleave count and vector factor depend on target register number. Currently, it does not
estimate different register pressure for different register class separately(especially for scalar type,
float type should not be on the same position with int type), so it's not accurate. Specifically,
it causes too many times interleaving/unrolling, result in too many register spills in loop body and hurting performance.
So we need classify the register classes in IR level, and importantly these are abstract register classes,
and are not the target register class of backend provided in td file. It's used to establish the mapping between
the types of IR values and the number of simultaneous live ranges to which we'd like to limit for some set of those types.
For example, POWER target, register num is special when VSX is enabled. When VSX is enabled, the number of int scalar register is 32(GPR),
float is 64(VSR), but for int and float vector register both are 64(VSR). So there should be 2 kinds of register class when vsx is enabled,
and 3 kinds of register class when VSX is NOT enabled.
It runs on POWER target, it makes big(+~30%) performance improvement in one specific bmk(503.bwaves_r) of spec2017 and no other obvious degressions.
Differential revision: https://reviews.llvm.org/D67148
llvm-svn: 374634
This selects MVE VQADD from the vector llvm.sadd.sat or llvm.uadd.sat
intrinsics.
Differential Revision: https://reviews.llvm.org/D68566
llvm-svn: 374336
Currently, the heuristics the if-conversion pass uses for diamond if-conversion
are based on execution time, with no consideration for code size. This adds a
new set of heuristics to be used when optimising for code size.
This is mostly target-independent, because the if-conversion pass can
see the code size of the instructions which it is removing. For thumb,
there are a few passes (insertion of IT instructions, selection of
narrow branches, and selection of CBZ instructions) which are run after
if conversion and affect these heuristics, so I've added target hooks to
better predict the code-size effect of a proposed if-conversion.
Differential revision: https://reviews.llvm.org/D67350
llvm-svn: 374301
Also Revert "[LoopVectorize] Fix non-debug builds after rL374017"
This reverts commit 9f41deccc0.
This reverts commit 18b6fe07bc.
The patch is breaking PowerPC internal build, checked with author, reverting
on behalf of him for now due to timezone.
llvm-svn: 374091
During the If-Converter optimization pay attention when copying or
deleting call instructions in order to keep call site information in
valid state.
Reviewers: aprantl, vsk, efriedma
Reviewed By: vsk, efriedma
Differential Revision: https://reviews.llvm.org/D66955
llvm-svn: 374068
Support for tracking registers that forward function parameters into the
following function frame. For now we only support cases when parameter
is forwarded through single register.
Reviewers: aprantl, vsk, t.p.northover
Reviewed By: vsk
Differential Revision: https://reviews.llvm.org/D66953
llvm-svn: 374033
Based on the discussion in
http://lists.llvm.org/pipermail/llvm-dev/2019-October/135574.html, the
conclusion was reached that the ARM backend should produce vcmp instead
of vcmpe instructions by default, i.e. not be producing an Invalid
Operation exception when either arguments in a floating point compare
are quiet NaNs.
In the future, after constrained floating point intrinsics for floating
point compare have been introduced, vcmpe instructions probably should
be produced for those intrinsics - depending on the exact semantics
they'll be defined to have.
This patch logically consists of the following parts:
- Revert http://llvm.org/viewvc/llvm-project?rev=294945&view=rev and
http://llvm.org/viewvc/llvm-project?rev=294968&view=rev, which
implemented fine-tuning for when to produce vcmpe (i.e. not do it for
equality comparisons). The complexity introduced by those patches
isn't needed anymore if we just always produce vcmp instead. Maybe
these patches need to be reintroduced again once support is needed to
map potential LLVM-IR constrained floating point compare intrinsics to
the ARM instruction set.
- Simply select vcmp, instead of vcmpe, see simple changes in
lib/Target/ARM/ARMInstrVFP.td
- Adapt lots of tests that tested for vcmpe (instead of vcmp). For all
of these test, the intent of what is tested for isn't related to
whether the vcmp should produce an Invalid Operation exception or not.
Fixes PR43374.
Differential Revision: https://reviews.llvm.org/D68463
llvm-svn: 374025
In loop-vectorize, interleave count and vector factor depend on target register number. Currently, it does not
estimate different register pressure for different register class separately(especially for scalar type,
float type should not be on the same position with int type), so it's not accurate. Specifically,
it causes too many times interleaving/unrolling, result in too many register spills in loop body and hurting performance.
So we need classify the register classes in IR level, and importantly these are abstract register classes,
and are not the target register class of backend provided in td file. It's used to establish the mapping between
the types of IR values and the number of simultaneous live ranges to which we'd like to limit for some set of those types.
For example, POWER target, register num is special when VSX is enabled. When VSX is enabled, the number of int scalar register is 32(GPR),
float is 64(VSR), but for int and float vector register both are 64(VSR). So there should be 2 kinds of register class when vsx is enabled,
and 3 kinds of register class when VSX is NOT enabled.
It runs on POWER target, it makes big(+~30%) performance improvement in one specific bmk(503.bwaves_r) of spec2017 and no other obvious degressions.
Differential revision: https://reviews.llvm.org/D67148
llvm-svn: 374017
Darwin platforms need the frame register to always point at a valid record even
if it's not updated in a leaf function. Backtraces are more important than one
extra GPR.
llvm-svn: 373738
Identity shuffles, of the form (0, 1, 2, 3, ...) are perfectly OK under MVE
(they essentially just become bitcasts). We were not catching that in the
existing set of what we considered legal though. On NEON, they would be covered
by vext's, but that is not generally available in MVE.
This uses ShuffleVectorInst::isIdentityMask which is a little odd to use here
but does what we want and prevents us from just rewriting what is the same
function.
Differential Revision: https://reviews.llvm.org/D68241
llvm-svn: 373446
Replace with the MachineFunction. X86 is the only user, and only uses
it for the function. This removes one obstacle from using this in
GlobalISel. The other is the more tolerable EVT argument.
The X86 use of the function seems questionable to me. It checks hasFP,
before frame lowering.
llvm-svn: 373292
The VCTP instruction will calculate the predicate masked based upon
the number of elements that need to be processed. I had inserted the
sub before the vctp intrinsic and supplied it as the operand, but
this is incorrect as the phi should directly feed the vctp. The sub
is calculating the value for the next iteration.
Differential Revision: https://reviews.llvm.org/D67921
llvm-svn: 373188
As we perform a zext on any arguments used in the promoted tree, it
doesn't matter if they're marked as signext. The only permitted
user(s) in the tree which would interpret the sign bits are signed
icmps. For these instructions, their promoted operands are truncated
before the icmp uses them.
Differential Revision: https://reviews.llvm.org/D68019
llvm-svn: 373186
This is an attempt to fill in some of the missing instructions from the
Cortex-M4 schedule, and make it easier to do the same for other ARM cpus.
- Some instructions are marked as hasNoSchedulingInfo as they are pseudos or
otherwise do not require scheduling info
- A lot of features have been marked not supported
- Some WriteRes's have been added for cvt instructions.
- Some extra instruction latencies have been added, notably by relaxing the
regex for dsp instruction to catch more cases, and some fp instructions.
This goes a long way to get the CompleteModel working for this CPU. It does not
go far enough as to get all scheduling info for all output operands correct.
Differential Revision: https://reviews.llvm.org/D67957
llvm-svn: 373163
The static analyzer is warning about potential null dereferences, but we should be able to use cast<> directly and if not assert will fire for us.
llvm-svn: 372992
During legalisation we can end up with some pretty strange nodes, like shifts
of 0. We need to make sure we don't try to make long shifts of these, ending up
with invalid assembly instructions. A long shift with a zero immediate actually
encodes a shift by 32.
Differential Revision: https://reviews.llvm.org/D67664
llvm-svn: 372839
Similar to rL372717, we can force the splitting of extends of vector loads in
MVE, in order to use the better widening loads as opposed to going through
expensive extends. This adds a combine to early-on detect extends of loads and
split the load in two, from where normal legalisation will kick in and we get a
series of widening loads.
Differential Revision: https://reviews.llvm.org/D67909
llvm-svn: 372721
MVE does not have a simple sign extend instruction that can move elements
across lanes. We currently often end up moving each lane into and out of a GPR,
in order to get elements into the correct places. When we have a store of a
trunc (or a extend of a load), we can instead just split the store/load in two,
using the narrowing/widening load/store instructions from each half of the
vector.
This does that for stores. It happens very early in a store combine, so as to
easily detect the truncates. (It would be possible to do this later, but that
would involve looking through a buildvector of extract elements. Not impossible
but this way seemed simpler).
By enabling store combines we also get a vmovdrr combine for free, helping some
other tests.
Differential Revision: https://reviews.llvm.org/D67828
llvm-svn: 372717
Summary:
The functions different in two ways:
- getLLVMRegNum could return both "eh" and "other" dwarf register
numbers, while getLLVMRegNumFromEH only returned the "eh" number.
- getLLVMRegNum asserted if the register was not found, while the second
function returned -1.
The second distinction was pretty important, but it was very hard to
infer that from the function name. Aditionally, for the use case of
dumping dwarf expressions, we needed a function which can work with both
kinds of number, but does not assert.
This patch solves both of these issues by merging the two functions into
one, returning an Optional<unsigned> value. While the same thing could
be achieved by adding an "IsEH" argument to the (renamed)
getLLVMRegNumFromEH function, it seemed better to avoid the confusion of
two functions and put the choice of asserting into the hands of the
caller -- if he checks the Optional value, he can safely process
"untrusted" input, and if he blindly dereferences the Optional, he gets
the assertion.
I've updated all call sites to the new API, choosing between the two
options according to the function they were calling originally, except
that I've updated the usage in DWARFExpression.cpp to use the "safe"
method instead, and added a test case which would have previously
triggered an assertion failure when processing (incorrect?) dwarf
expressions.
Reviewers: dsanders, arsenm, JDevlieghere
Subscribers: wdng, aprantl, javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67154
llvm-svn: 372710
Remove any predicate that we replace with a vctp intrinsic, and try
to remove their operands too. Also look into the exit block to see if
there's any duplicates of the predicates that we've replaced and
clone the vctp to be used there instead.
Differential Revision: https://reviews.llvm.org/D67709
llvm-svn: 372567
Check whether there are any uses or defs between the LoopDec and
LoopEnd. If there's not, then we can use a subs to set the cpsr and
skip generating a cmp.
Differential Revision: https://reviews.llvm.org/D67801
llvm-svn: 372560
This reverts r372314, reapplying r372285 and the commits which depend
on it (r372286-r372293, and r372296-r372297)
This was missing one switch to getTargetConstant in an untested case.
llvm-svn: 372338
This broke the Chromium build, causing it to fail with e.g.
fatal error: error in backend: Cannot select: t362: v4i32 = X86ISD::VSHLI t392, Constant:i8<15>
See llvm-commits thread of r372285 for details.
This also reverts r372286, r372287, r372288, r372289, r372290, r372291,
r372292, r372293, r372296, and r372297, which seemed to depend on the
main commit.
> Encode them directly as an imm argument to G_INTRINSIC*.
>
> Since now intrinsics can now define what parameters are required to be
> immediates, avoid using registers for them. Intrinsics could
> potentially want a constant that isn't a legal register type. Also,
> since G_CONSTANT is subject to CSE and legalization, transforms could
> potentially obscure the value (and create extra work for the
> selector). The register bank of a G_CONSTANT is also meaningful, so
> this could throw off future folding and legalization logic for AMDGPU.
>
> This will be much more convenient to work with than needing to call
> getConstantVRegVal and checking if it may have failed for every
> constant intrinsic parameter. AMDGPU has quite a lot of intrinsics wth
> immarg operands, many of which need inspection during lowering. Having
> to find the value in a register is going to add a lot of boilerplate
> and waste compile time.
>
> SelectionDAG has always provided TargetConstant for constants which
> should not be legalized or materialized in a register. The distinction
> between Constant and TargetConstant was somewhat fuzzy, and there was
> no automatic way to force usage of TargetConstant for certain
> intrinsic parameters. They were both ultimately ConstantSDNode, and it
> was inconsistently used. It was quite easy to mis-select an
> instruction requiring an immediate. For SelectionDAG, start emitting
> TargetConstant for these arguments, and using timm to match them.
>
> Most of the work here is to cleanup target handling of constants. Some
> targets process intrinsics through intermediate custom nodes, which
> need to preserve TargetConstant usage to match the intrinsic
> expectation. Pattern inputs now need to distinguish whether a constant
> is merely compatible with an operand or whether it is mandatory.
>
> The GlobalISelEmitter needs to treat timm as a special case of a leaf
> node, simlar to MachineBasicBlock operands. This should also enable
> handling of patterns for some G_* instructions with immediates, like
> G_FENCE or G_EXTRACT.
>
> This does include a workaround for a crash in GlobalISelEmitter when
> ARM tries to uses "imm" in an output with a "timm" pattern source.
llvm-svn: 372314
We needn't BFI each lane individually into a predicate register when each lane
in the same. A simple sign extend and a vmsr will do.
Differential Revision: https://reviews.llvm.org/D67653
llvm-svn: 372313
Encode them directly as an imm argument to G_INTRINSIC*.
Since now intrinsics can now define what parameters are required to be
immediates, avoid using registers for them. Intrinsics could
potentially want a constant that isn't a legal register type. Also,
since G_CONSTANT is subject to CSE and legalization, transforms could
potentially obscure the value (and create extra work for the
selector). The register bank of a G_CONSTANT is also meaningful, so
this could throw off future folding and legalization logic for AMDGPU.
This will be much more convenient to work with than needing to call
getConstantVRegVal and checking if it may have failed for every
constant intrinsic parameter. AMDGPU has quite a lot of intrinsics wth
immarg operands, many of which need inspection during lowering. Having
to find the value in a register is going to add a lot of boilerplate
and waste compile time.
SelectionDAG has always provided TargetConstant for constants which
should not be legalized or materialized in a register. The distinction
between Constant and TargetConstant was somewhat fuzzy, and there was
no automatic way to force usage of TargetConstant for certain
intrinsic parameters. They were both ultimately ConstantSDNode, and it
was inconsistently used. It was quite easy to mis-select an
instruction requiring an immediate. For SelectionDAG, start emitting
TargetConstant for these arguments, and using timm to match them.
Most of the work here is to cleanup target handling of constants. Some
targets process intrinsics through intermediate custom nodes, which
need to preserve TargetConstant usage to match the intrinsic
expectation. Pattern inputs now need to distinguish whether a constant
is merely compatible with an operand or whether it is mandatory.
The GlobalISelEmitter needs to treat timm as a special case of a leaf
node, simlar to MachineBasicBlock operands. This should also enable
handling of patterns for some G_* instructions with immediates, like
G_FENCE or G_EXTRACT.
This does include a workaround for a crash in GlobalISelEmitter when
ARM tries to uses "imm" in an output with a "timm" pattern source.
llvm-svn: 372285
This patch fixes a bug exposed by D65653 where a subsequent invocation
of `determineCalleeSaves` ends up with a different size for the callee
save area, leading to different frame-offsets in debug information.
In the invocation by PEI, `determineCalleeSaves` tries to determine
whether it needs to spill an extra callee-saved register to get an
emergency spill slot. To do this, it calls 'estimateStackSize' and
manually adds the size of the callee-saves to this. PEI then allocates
the spill objects for the callee saves and the remaining frame layout
is calculated accordingly.
A second invocation in LiveDebugValues causes estimateStackSize to return
the size of the stack frame including the callee-saves. Given that the
size of the callee-saves is added to this, these callee-saves are counted
twice, which leads `determineCalleeSaves` to believe the stack has
become big enough to require spilling an extra callee-save as emergency
spillslot. It then updates CalleeSavedStackSize with a larger value.
Since CalleeSavedStackSize is used in the calculation of the frame
offset in getFrameIndexReference, this leads to incorrect offsets for
variables/locals when this information is recalculated after PEI.
Reviewers: omjavaid, eli.friedman, thegameg, efriedma
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D66935
llvm-svn: 372204
r361845 changed the way we handle "D16" vs. "D32" targets; there used to
be a negative "d16" which removed instructions from the instruction set,
and now there's a "d32" feature which adds instructions to the
instruction set. This is good, but there was an oversight in the
implementation: the behavior of VFPv2 was changed. In particular, the
"vfp2" feature was changed to imply "d32". This is wrong: VFPv2 only
supports 16 D registers.
In practice, this means if you specify -mfpu=vfpv2, the compiler will
generate illegal instructions.
This patch gets rid of "vfp2d16" and "vfp2d16sp", and fixes "vfp2" and
"vfp2sp" so they don't imply "d32".
Differential Revision: https://reviews.llvm.org/D67375
llvm-svn: 372186
The static analyzer is warning about potential null dereferences of dyn_cast<> results - in these cases we can safely use cast<> directly as we know that these cases should all be the correct type, which is why its working atm and anyway cast<> will assert if they aren't.
llvm-svn: 372145
We were previously using the SelectT2AddrModeImm7 for both normal and narrowing
MVE loads/stores. As the narrowing instructions do not accept sp as a register,
it makes little sense to optimise a FrameIndex into the load, only to have to
recover that later on. This adds a SelectTAddrModeImm7 which does not do that
folding, and uses it for narrowing load/store patterns.
Differential Revision: https://reviews.llvm.org/D67489
llvm-svn: 372134
Similar to D67327, but this time for the FP16 VLDR and VSTR instructions that
use the AddrMode5FP16 addressing mode. We need to reserve an emergency spill
slot for instructions that will be out of range to use sp directly.
AddrMode5FP16 is 8 bits with a scale of 2.
Differential Revision: https://reviews.llvm.org/D67483
llvm-svn: 372132
Remove setPreservesCFG from ARMConstantIslandPass and add a couple
of -verify-machine-dom-info instances into the existing codegen
tests.
llvm-svn: 372126
MVE loads and stores have a 7 bit immediate range, scaled by the length of the type. This needs to be taught to the stack estimation code to ensure that an emergency spill slot is reserved in case we run out of registers when materialising stack indices.
Also the narrowing loads/stores can be created with frame indices even though they do not accept SP as a register. We need in those cases to make sure we have an emergency register to use as the frame base, as SP can never be used.
Differential Revision: https://reviews.llvm.org/D67327
llvm-svn: 372114
Converting the *LoopStart pseudo instructions into DLS/WLS results in
LR being defined. These instructions were inserted on the assumption
that LR would already contain the loop counter because a mov is
introduced during ISel as the the consumers in the loop can only use
LR. That assumption proved wrong!
So perform a safety check, finding an appropriate place to insert the
DLS/WLS instructions or revert if this isn't possible.
Differential Revision: https://reviews.llvm.org/D67539
llvm-svn: 372111
* Reordered MVT simple types to group scalable vector types
together.
* New range functions in MachineValueType.h to only iterate over
the fixed-length int/fp vector types.
* Stopped backends which don't support scalable vector types from
iterating over scalable types.
Reviewers: sdesmalen, greened
Reviewed By: greened
Differential Revision: https://reviews.llvm.org/D66339
llvm-svn: 372099
The low-overhead branch extension provides a loop-end 'LE' instruction
that performs no decrement nor compare, it just jumps backwards. This
patch modifies the constant islands pass to try to insert LE
instructions in place of a Thumb2 conditional branch, instead of
shrinking it. This only happens if a cmp can be converted to a cbn/z
and used to exit the loop.
Differential Revision: https://reviews.llvm.org/D67404
llvm-svn: 372085
Set this bit for the MVE reduction instructions to prevent a loop from
becoming tail predicated in their presence.
Differential Revision: https://reviews.llvm.org/D67444
llvm-svn: 372076
The adds some very basic folding of PREDICATE_CASTS, removing cases when they
are chained together. These would already be removed eventually, as these are
lowered to copies. This just allows it to happen earlier, which can help other
simplifications.
Differential Revision: https://reviews.llvm.org/D67591
llvm-svn: 372012
Lower CTTZ on MVE using VBRSR and VCLS which will reverse the bits and
count the leading zeros, equivalent to a count trailing zeros (CTTZ).
llvm-svn: 372000
MVE has VPT instructions, which perform the duties of both a VCMP and a VPST in
a single instruction, performing the compare and starting the VPT block in one.
This teaches the MVEVPTBlockPass to fold them, searching back through the
basicblock for a valid VCMP and creating the VPT from its operands.
There are some changes to the VPT instructions to accommodate this, altering
the order of the operands to match the VCMP better, and changing P0 register
defs to be VPR defs, as is used in other places.
Differential Revision: https://reviews.llvm.org/D66577
llvm-svn: 371982
Masked loads and store fit naturally with MVE, the instructions being easily
predicated. This adds lowering for the simple cases of masked loads and stores.
It does not yet deal with widening/narrowing or pre/post inc, and so is
currently behind an option.
The llvm masked load intrinsic will accept a "passthru" value, dictating the
values used for the zero masked lanes. In MVE the instructions write 0 to the
zero predicated lanes, so we need to match a passthru that isn't 0 (or undef)
with a select instruction to pull in the correct data after the load.
Differential Revision: https://reviews.llvm.org/D67186
llvm-svn: 371932
rL367544 added @earlyclobbers for the MVE VREV64 instruction. This adds the
same for a number of other 32bit instructions that are similarly unpredictable
if the destination equals the source (due to the cross beat nature of the
instructions).
This includes:
VCADD.f32
VCADD.i32
VCMUL.f32
VHCADD.s32
VMULLT/B.s/u32
VQDMLADH{X}.s32
VQRDMLADH{X}.s32
VQDMLSDH{X}.s32
VQRDMLSDH{X}.s32
VQDMULLT/B.s32 with Qm and Rm
No tests here as this would require intrinsics (or very interesting codegen) to
manifest. The tests will follow naturally as the intrinsics are added.
Differential Revision: https://reviews.llvm.org/D67462
llvm-svn: 371838
This patch adds vecreduce_smax, vecredude_umax, vecreduce_smin, vecreduce_umin and selection for vmaxv and minv.
Differential Revision: https://reviews.llvm.org/D66413
llvm-svn: 371827
Summary:
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet, JDevlieghere, alexshap, rupprecht, jhenderson
Subscribers: sdardis, nemanjai, hiraditya, kbarton, jakehehrlich, jrtc27, MaskRay, atanasyan, jsji, seiya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D67499
llvm-svn: 371742
Summary:
This catches malformed mir files which specify alignment as log2 instead of pow2.
See https://reviews.llvm.org/D65945 for reference,
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: MatzeB, qcolombet, dschuff, arsenm, sdardis, nemanjai, jvesely, nhaehnle, hiraditya, kbarton, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, jsji, Petar.Avramovic, asbirlea, s.egerton, pzheng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67433
llvm-svn: 371608
These predicate vectors can usually be loaded and stored with a single
instruction, a VSTR_P0. However this instruction will store the entire P0
predicate, 16 bits, zeroextended to 32bits. Each lane of the the
v4i1/v8i1/v16i1 representing 4/2/1 bits.
As far as I understand, when llvm says "store this v4i1", it really does need
to store 4 bits (or 8, that being the size of a byte, with this bottom 4 as the
interesting bits). For example a bitcast from a v8i1 to a i8 is defined as a
store followed by a load, which is how the code is expanded.
So this instead lowers the v4i1/v8i1 load/store through some shuffles to get
the bits into the correct positions. This, as you might imagine, is not as
efficient as a single instruction. But I believe it is needed for correctness.
v16i1 equally should not load/store 32bits, only storing the 16bits of data.
Stack loads/stores are still using the VSTR_P0 (as can be seen by the test not
changing). This is fine as they are self-consistent, it is only "externally
observable loads/stores" (from our point of view) that need to be corrected.
Differential revision: https://reviews.llvm.org/D67085
llvm-svn: 371419
The family of 'dual-accumulating' vector multiply-add instructions
(VMLADAV, VMLALDAV and VRMLALDAVH) can all operate on both signed and
unsigned integer types, and they all have an 'exchange' variant (with
an X in the name) that modifies which pairs of vector lanes in the two
inputs are multiplied together. But there's a clause in the spec that
says that the X variants //don't// operate on unsigned integer types,
only signed. You can have X, or unsigned, or neither, but not both.
We didn't notice that clause when we implemented the MC support for
these instructions, so LLVM believes that things like VMLADAVX.U8 do
exist, contradicting the spec. Here I fix that by conditioning them
out in Tablegen.
In order to do that, I've reversed the nesting order of the Tablegen
multiclasses for those instructions. Previously, the innermost
multiclass generated the X and not-X variants, and the one outside
that generated the A and not-A variants. Now X is done by the outer
multiclass, which allows me to bypass that one when I only want the
two not-X variants.
Changing the multiclass nesting order also changes the names of the
instruction ids unless I make a special effort not to. I decided that
while I was changing them anyway I'd make them look nicer; so now the
instructions have names like MVE_VMLADAVs32 or MVE_VMLADAVaxs32,
instead of cumbersome _noacc_noexch suffixes.
The corresponding multiply-subtract instructions are unaffected. Those
don't accept unsigned types at all, either in the spec or in LLVM.
Reviewers: ostannard, dmgreen
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67214
llvm-svn: 371405
We should not be generating Neon stack loads/stores even for these large
registers.
No test here because my understanding is we will only generate these QQPR regs
for intrinsics and VLDn's. The tests will follow once those are available.
Differential revision: https://reviews.llvm.org/D67169
llvm-svn: 371386
Specify the Unpredictable bits, and return softfails when appropriate.
Patch by Mark Murray!
Differential revision: https://reviews.llvm.org/D66939
llvm-svn: 371374
The incoming accumulator value can be discovered through a sext, in
which case there will be a mismatch between the input and the result.
So sign extend the accumulator input if we're performing a 64-bit mac.
Differential Revision: https://reviews.llvm.org/D67220
llvm-svn: 371370
Summary:
This is the first change to enable the TLI to be built per-function so
that -fno-builtin* handling can be migrated to use function attributes.
See discussion on D61634 for background. This is an enabler for fixing
handling of these options for LTO, for example.
This change should not affect behavior, as the provided function is not
yet used to build a specifically per-function TLI, but rather enables
that migration.
Most of the changes were very mechanical, e.g. passing a Function to the
legacy analysis pass's getTLI interface, or in Module level cases,
adding a callback. This is similar to the way the per-function TTI
analysis works.
There was one place where we were looking for builtins but not in the
context of a specific function. See FindCXAAtExit in
lib/Transforms/IPO/GlobalOpt.cpp. I'm somewhat concerned my workaround
could provide the wrong behavior in some corner cases. Suggestions
welcome.
Reviewers: chandlerc, hfinkel
Subscribers: arsenm, dschuff, jvesely, nhaehnle, mehdi_amini, javed.absar, sbc100, jgravelle-google, eraman, aheejin, steven_wu, george.burgess.iv, dexonsmith, jfb, asbirlea, gchatelet, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66428
llvm-svn: 371284
This patch sinks add/mul(shufflevector(insertelement())) into the basic block in which they are used so that they can then be selected together.
This is useful for various MVE instructions, such as vmla and others that take R registers.
Loop tests have been added to the vmla test file to make sure vmlas are generated in loops.
Differential revision: https://reviews.llvm.org/D66295
llvm-svn: 371218
Summary:
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: jyknight, sdardis, nemanjai, javed.absar, hiraditya, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, jsji, s.egerton, pzheng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67229
llvm-svn: 371200
The MVE and LOB extensions of Armv8.1m can be combined to enable
'tail predication' which removes the need for a scalar remainder
loop after vectorization. Lane predication is performed implicitly
via a system register. The effects of predication is described in
Section B5.6.3 of the Armv8.1-m Arch Reference Manual, the key points
being:
- For vector operations that perform reduction across the vector and
produce a scalar result, whether the value is accumulated or not.
- For non-load instructions, the predicate flags determine if the
destination register byte is updated with the new value or if the
previous value is preserved.
- For vector store instructions, whether the store occurs or not.
- For vector load instructions, whether the value that is loaded or
whether zeros are written to that element of the destination
register.
This patch implements a pass that takes a hardware loop, containing
masked vector instructions, and converts it something that resembles
an MVE tail predicated loop. Currently, if we had code generation,
we'd generate a loop in which the VCTP would generate the predicate
and VPST would then setup the value of VPR.PO. The loads and stores
would be placed in VPT blocks so this is not tail predication, but
normal VPT predication with the predicate based upon a element
counting induction variable. Further work needs to be done to finally
produce a true tail predicated loop.
Because only the loads and stores are predicated, in both the LLVM IR
and MIR level, we will restrict support to only lane-wise operations
(no horizontal reductions). We will perform a final check on MIR
during loop finalisation too.
Another restriction, specific to MVE, is that all the vector
instructions need operate on the same number of elements. This is
because predication is performed at the byte level and this is set
on entry to the loop, or by the VCTP instead.
Differential Revision: https://reviews.llvm.org/D65884
llvm-svn: 371179
A number of inline assembly constraints are currently supported by LLVM, but rejected as invalid by Clang:
Target independent constraints:
s: An integer constant, but allowing only relocatable values
ARM specific constraints:
j: An immediate integer between 0 and 65535 (valid for MOVW)
x: A 32, 64, or 128-bit floating-point/SIMD register: s0-s15, d0-d7, or q0-q3
N: An immediate integer between 0 and 31 (Thumb1 only)
O: An immediate integer which is a multiple of 4 between -508 and 508. (Thumb1 only)
This patch adds support to Clang for the missing constraints along with some checks to ensure that the constraints are used with the correct target and Thumb mode, and that immediates are within valid ranges (at least where possible). The constraints are already implemented in LLVM, but just a couple of minor corrections to checks (V8M Baseline includes MOVW so should work with 'j', 'N' and 'O' shouldn't be valid in Thumb2) so that Clang and LLVM are in line with each other and the documentation.
Differential Revision: https://reviews.llvm.org/D65863
Change-Id: I18076619e319bac35fbb60f590c069145c9d9a0a
llvm-svn: 371079
This attempts to just fix the creation of VPT blocks, fixing up the iterating,
which instructions are considered in the bundle, and making sure that we do not
overrun the end of the block.
Differential Revision: https://reviews.llvm.org/D67219
llvm-svn: 371064
Summary:
This patch renames functions that takes or returns alignment as log2, this patch will help with the transition to llvm::Align.
The renaming makes it explicit that we deal with log(alignment) instead of a power of two alignment.
A few renames uncovered dubious assignments:
- `MirParser`/`MirPrinter` was expecting powers of two but `MachineFunction` and `MachineBasicBlock` were using deal with log2(align). This patch fixes it and updates the documentation.
- `MachineBlockPlacement` exposes two flags (`align-all-blocks` and `align-all-nofallthru-blocks`) supposedly interpreted as power of two alignments, internally these values are interpreted as log2(align). This patch updates the documentation,
- `MachineFunctionexposes` exposes `align-all-functions` also interpreted as power of two alignment, internally this value is interpreted as log2(align). This patch updates the documentation,
Reviewers: lattner, thegameg, courbet
Subscribers: dschuff, arsenm, jyknight, dylanmckay, sdardis, nemanjai, jvesely, nhaehnle, javed.absar, hiraditya, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, dexonsmith, PkmX, jocewei, jsji, Jim, s.egerton, llvm-commits, courbet
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65945
llvm-svn: 371045
For any unpaired muls, we accumulate them as an input to the
reduction. Check the type of the mul and perform a sext if the
existing accumlator input type is not the same.
Differential Revision: https://reviews.llvm.org/D66993
llvm-svn: 370851
On AArch64, s128 types have to be split into s64 GPRs when passed as arguments.
This change adds the generic support in call lowering for dealing with multiple
registers, for incoming and outgoing args.
Support for splitting for return types not yet implemented.
Differential Revision: https://reviews.llvm.org/D66180
llvm-svn: 370822
These flags should simply be passed through to the target, which will do
the right thing. Add an MC/X86 test that uses these directives with the
three primary object file formats and shows that they disassemble the
same everywhere.
There is a missing test for .code32 on Windows ARM, since I'm not sure
exactly how to construct one.
Fixes PR43203
llvm-svn: 370805
The code here seems to date back to r134705, when tablegen lowering was first
being added. I don't believe that we need to include CPSR implicit operands on
the MCInst. This now works more like other backends (like AArch64), where all
implicit registers are skipped.
This allows the AliasInst for CSEL's to match correctly, as can be seen in the
test changes.
Differential revision: https://reviews.llvm.org/D66703
llvm-svn: 370745
This moves ConstantMaterializationCost into ARMBaseInstrInfo so that it can
also be used in ISel Lowering, adding codesize values to the computed costs, to
be able to compare either approximate instruction counts or codesize costs.
It also adds a HasLowerConstantMaterializationCost, which compares the
ConstantMaterializationCost of two values, returning true if the first is
smaller either in instruction count/codesize, or falling back to the other in
the case that they are equal.
This is used in constant CSEL lowering to invert the predicate if the opposite
is easier to materialise.
Differential revision: https://reviews.llvm.org/D66701
llvm-svn: 370741
Arm 8.1-M adds a number of related CSEL instructions, including CSINC, CSNEG and CSINV. These choose between two values given the content in CPSR and a condition, performing an increment, negation or inverse of the false value.
This adds some selection for them, either from constant values or patterns. It does not include CSEL directly, which is currently not always making code better. It is still useful, but we will have to check more carefully where it should and shouldn't be used.
Code by Ranjeet Singh and Simon Tatham, with some modifications from me.
Differential revision: https://reviews.llvm.org/D66483
llvm-svn: 370739
We were using isShiftedInt<7, Shift>(RHSC) to detect the ranges of offsets to
fold into MVE loads/stores. The instructions actually take a 7 bit unsigned
integer which is either added or subtracted. So something more like
isShiftedUInt<7, Shift>(abs(RHSC)).
Instead I've changes this to use the isScaledConstantInRange method, same as in
SelectT2AddrModeImm7Offset used by pre/post inc, which seemed to already be
getting this correct.
Differential revision: https://reviews.llvm.org/D66997
llvm-svn: 370731
Decoding of VMSR doesn't diagnose some unpredictable encodings, as the unpredictable bits are not correctly set.
Diff-reduce this instruction's internals WRT VMRS so I can see the differences better. Mostly this is s/src/Rt/g.
Fill in the "should-be-(0)" bits.
Designate the Unpredictable{} bits for both VMRS and VMSR.
Patch by Mark Murray!
Differential revision: https://reviews.llvm.org/D66938
llvm-svn: 370729
To save a 'add sp,#val' instruction by adding registers to the final pop instruction,
the first register transferred by this pop instruction need to be found.
If the function to be optimized has a non-void return value, the operand list contains
r0 (implicit) which prevents the optimization to take place.
Therefore implicit register references should be skipped in the search loop,
because this registers are never popped from the stack.
Patch by Rainer Herbertz (rOptimizer)!
Differential revision: https://reviews.llvm.org/D66730
llvm-svn: 370728
We should be using MQPR, and if we don't we can get COPYs and PHIs created for
QPR. These get folded into instructions, failing verification checks.
Differential revision: https://reviews.llvm.org/D66214
llvm-svn: 370676
These were never enabled correctly and are causing other problems. Taking them
out for the moment, whilst we work on the issues.
This reverts r370329.
llvm-svn: 370607
Masked loads and store fit naturally with MVE, the instructions being easily
predicated. This adds lowering for the simple cases of masked loads and stores.
It does not yet deal with widening/narrowing or pre/post inc.
The llvm masked load intrinsic will accept a "passthru" value, dictating the
values used for the zero masked lanes. In MVE the instructions write 0 to the
zero predicated lanes, so we need to match a passthru that isn't 0 (or undef)
with a select instruction to pull in the correct data after the load.
We also need to do something with unaligned loads/stores. Currently this uses a
similar method used in big endian, using an VLDRB.8 (and potentially a VREV in
BE). This does mean that the predicate mask is converted from, for example, a
v4i1 to a v16i1. The VLDR instructions are defined as using the first bit of
the relevant mask lane, so this could potentially load different results if the
predicate is little odd. As the input is a v4i1 however, I believe this is OK
and all the bits required should be set in the predicate, making the VLDRB.8
load the same data.
Differential Revision: https://reviews.llvm.org/D66534
llvm-svn: 370329
The patch fixed the issue that RV64 didn't clear the upper bits
when return complex floating value with lp64 ABI.
float _Complex
complex_add(float _Complex a, float _Complex b)
{
return a + b;
}
RealResult = zero_extend(RealA + RealB)
ImageResult = ImageA + ImageB
Return (RealResult | (ImageResult << 32))
The patch introduces shouldExtendTypeInLibCall target hook to suppress
the AssertZext generation when lowering floating LibCall.
Thanks to Eli's comments from the Bugzilla
https://bugs.llvm.org/show_bug.cgi?id=42820
Differential Revision: https://reviews.llvm.org/D65497
llvm-svn: 370275
Summary: There are at least 2 ways to express the same shuffle. Various pieces of code explicit check for both option, but other places do not when they would benefit from doing it. This patches refactor the codebase to use buildLegalVectorShuffle in order to make that behavior more consistent.
Reviewers: craig.topper, efriedma, RKSimon, lebedev.ri
Subscribers: javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66804
llvm-svn: 370190
This just pulls the MVEVPTBlockPass into a separate file, as opposed to being
wrapped up in Thumb2ITBlockPass.
Differential revision: https://reviews.llvm.org/D66579
llvm-svn: 370187
This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some
adjustments for MVE. It allows us to move fp16 registers without going into and
out of gprs.
VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits
of another register, zeroing the rest. This can be used for odd MVE register
lanes. The top bits are not read by fp16 instructions, so no move is required
there if we are dealing with even lanes.
Differential revision: https://reviews.llvm.org/D66793
llvm-svn: 370184
rL369567 reverted a couple of recent changes made to ARMParallelDSP
because of a miscompilation error: PR43073.
The issue stemmed from an underlying bug that was caused by adding
muls into a reduction before it was proved that they could be executed
in parallel with another mul.
Most of the changes here are from the previously reverted commits.
The additional changes have been made area:
1) The Search function now doesn't insert any muls into the Reduction
object. That now happens once the search has successfully finished.
2) For any muls added into the reduction but that weren't paired, we
accumulate their values as an input into the smlad.
Differential Revision: https://reviews.llvm.org/D66660
llvm-svn: 370171
Prefer `MCFixupKind` where possible and add getTargetKind() to
convert to `unsigned` when needed rather than scattering cast
operators around the place.
Differential Revision: https://reviews.llvm.org/D59890
llvm-svn: 369720
The CodeGen/Thumb2/mve-vaddv.ll test needed to be amended to reflect the
changes from the above patch.
This reverts commit cd53ff6, reapplying 7c6b229.
llvm-svn: 369638
It broke the bots, see e.g. http://lab.llvm.org:8011/builders/clang-cuda-build/builds/36275/
> This patch fixes shifts by a 128/256 bit shift amount. It also fixes
> codegen for shifts of 32 by delegating to LLVM's default optimisation
> instead of emitting a long shift.
>
> Tests that used to generate long shifts of 32 are updated to check for the
> more optimised codegen.
>
> Differential revision: https://reviews.llvm.org/D66519
>
> llvm-svn: 369626
llvm-svn: 369636
This patch fixes shifts by a 128/256 bit shift amount. It also fixes
codegen for shifts of 32 by delegating to LLVM's default optimisation
instead of emitting a long shift.
Tests that used to generate long shifts of 32 are updated to check for the
more optimised codegen.
Differential revision: https://reviews.llvm.org/D66519
llvm-svn: 369626
The patch introduces MakeLibCallOptions struct as suggested by @efriedma on D65497.
The struct contain argument flags which will pass to makeLibCall function.
The patch should not has any functionality changes.
Differential Revision: https://reviews.llvm.org/D65795
llvm-svn: 369622
This patch adds vecreduce_add and the relevant instruction selection for
vaddv.
Differential revision: https://reviews.llvm.org/D66085
llvm-svn: 369245
This adds some sext costs for MVE, taken from the length of assembly sequences
that we currently generate.
Differential Revision: https://reviews.llvm.org/D66010
llvm-svn: 369244
We currently don't use liveness information after this point, but it can
be useful to catch bugs using -verify-machineinstrs, and optimizations
could potentially use this information in the future.
Differential Revision: https://reviews.llvm.org/D66319
llvm-svn: 369162
Push LR register before calling __gnu_mcount_nc as it expects the value of LR register to be the top value of
the stack on ARM32.
Differential Revision: https://reviews.llvm.org/D65019
llvm-svn: 369147
MVE also has some sext of loads, which will be free just as scalar
instructions are.
Differential Revision: https://reviews.llvm.org/D66008
llvm-svn: 369118
The widening and narrowing MVE instructions like VLDRH.32 are only permitted to
use low tGPR registers. This means that if they are used for a stack slot,
where the register used is only decided during frame setup, we need to be able
to correctly pick a thumb1 register over a normal GPR.
This attempts to add the required logic into eliminateFrameIndex and
rewriteT2FrameIndex, only picking the FrameReg if it is a valid register for
the operands register class, and picking a valid scratch register for the
register class.
Differential Revision: https://reviews.llvm.org/D66285
llvm-svn: 369108
We don't yet know how to generate these instructions for MVE. And in the case
of VLD3, we don't even have the instruction. For the moment don't tell the
vectoriser that we have VLD4, just to end up serialising the results.
Differential Revision: https://reviews.llvm.org/D66009
llvm-svn: 369101
Two issues:
1. t2CMPri shouldn't use CPSR if it isn't predicated. This doesn't
really have any visible effect at the moment, but it might matter in the
future.
2. The t2CMPri generated for t2WhileLoopStart might need to use a
register that isn't LR.
My team found this because we have a patch to track register liveness
late in the pass pipeline. I'll look into upstreaming it to help catch
issues like this earlier.
Differential Revision: https://reviews.llvm.org/D66243
llvm-svn: 369069
Summary:
This clang-tidy check is looking for unsigned integer variables whose initializer
starts with an implicit cast from llvm::Register and changes the type of the
variable to llvm::Register (dropping the llvm:: where possible).
Partial reverts in:
X86FrameLowering.cpp - Some functions return unsigned and arguably should be MCRegister
X86FixupLEAs.cpp - Some functions return unsigned and arguably should be MCRegister
X86FrameLowering.cpp - Some functions return unsigned and arguably should be MCRegister
HexagonBitSimplify.cpp - Function takes BitTracker::RegisterRef which appears to be unsigned&
MachineVerifier.cpp - Ambiguous operator==() given MCRegister and const Register
PPCFastISel.cpp - No Register::operator-=()
PeepholeOptimizer.cpp - TargetInstrInfo::optimizeLoadInstr() takes an unsigned&
MachineTraceMetrics.cpp - MachineTraceMetrics lacks a suitable constructor
Manual fixups in:
ARMFastISel.cpp - ARMEmitLoad() now takes a Register& instead of unsigned&
HexagonSplitDouble.cpp - Ternary operator was ambiguous between unsigned/Register
HexagonConstExtenders.cpp - Has a local class named Register, used llvm::Register instead of Register.
PPCFastISel.cpp - PPCEmitLoad() now takes a Register& instead of unsigned&
Depends on D65919
Reviewers: arsenm, bogner, craig.topper, RKSimon
Reviewed By: arsenm
Subscribers: RKSimon, craig.topper, lenary, aemerson, wuzish, jholewinski, MatzeB, qcolombet, dschuff, jyknight, dylanmckay, sdardis, nemanjai, jvesely, wdng, nhaehnle, sbc100, jgravelle-google, kristof.beyls, hiraditya, aheejin, kbarton, fedor.sergeev, javed.absar, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, tpr, PkmX, jocewei, jsji, Petar.Avramovic, asbirlea, Jim, s.egerton, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65962
llvm-svn: 369041
Now that we've moved to C++14, we no longer need the llvm::make_unique
implementation from STLExtras.h. This patch is a mechanical replacement
of (hopefully) all the llvm::make_unique instances across the monorepo.
llvm-svn: 369013
We need to allow any alignment at least 2, not just exactly 2, so that the big
endian loads and stores can be selected successfully. I've also added extra BE
testing for the load and store tests.
Thanks to Oliver for the report.
Differential Revision: https://reviews.llvm.org/D66222
llvm-svn: 368996
Stack loads and stores were already working, but direct stores were not. This
adds the patterns for them, same as predicate loads.
Differential Revision: https://reviews.llvm.org/D66213
llvm-svn: 368988
This adds patterns for selecting trunc instructions from full vectors to i1's
vectors.
Differential Revision: https://reviews.llvm.org/D66201
llvm-svn: 368981
The MVE architecture has the idea of "beats", where a vector instruction can be
executed over several ticks of the architecture. This adds a similar system
into the Arm backend cost model, multiplying the cost of all vector
instructions by a factor.
This factor essentially becomes the expected difference between scalar code
and vector code, on average. MVE Vector instructions can also overlap so the a
true cost of them is often lower. But equally scalar instructions can in some
situations be dual issued, or have other optimisations such as unrolling or
make use of dsp instructions. The default is chosen as 2. This should not
prevent vectorisation is a most cases (as the vector instructions will still be
doing at least 4 times the work), but it will help prevent over vectorising in
cases where the benefits are less likely.
This adds things so far to the obvious places in ARMTargetTransformInfo, and
updates a few related costs like not treating float instructions as cost 2 just
because they are floats.
Differential Revision: https://reviews.llvm.org/D66005
llvm-svn: 368733
Currently shufflemasks get emitted as any other constant, and you end
up with a bunch of virtual registers of G_CONSTANT with a
G_BUILD_VECTOR. The AArch64 selector then asserts on anything that
doesn't fit this pattern. This isn't an ideal representation, and
should avoid legalization and have fewer opportunities for a
representational error.
Rather than invent a new shuffle mask operand type, similar to what
ShuffleVectorSDNode does, just track the original IR Constant mask
operand. I don't completely like the idea of adding another link to
the IR, but MIR is already quite dependent on IR constants already,
and this will allow sharing the shuffle mask utility functions with
the IR.
llvm-svn: 368704
Currently we can't keep any state in the selector object that we get from
subtarget. As a result we have to plumb through all our variables through
multiple functions. This change makes it non-const and adds a virtual init()
method to allow further state to be captured for each target.
AArch64 makes use of this in this patch to cache a call to hasFnAttribute()
which is expensive to call, and is used on each selection of G_BRCOND.
Differential Revision: https://reviews.llvm.org/D65984
llvm-svn: 368652
This teaches the cost model that the sext or zext of a load is going to be
free.
Differential Revision: https://reviews.llvm.org/D66006
llvm-svn: 368593
A VDUP will perform a vector broadcast in a single instruction. Update the cost
model for MVE accordingly.
Code originally by David Sherwood.
Differential Revision: https://reviews.llvm.org/D63448
llvm-svn: 368589
This puts some of the calls in ARMTargetTransformInfo.cpp behind hasNeon()
checks, now that we have MVE, and updates all the tests accordingly.
Differential Revision: https://reviews.llvm.org/D63447
llvm-svn: 368587
Due to the nature of the beat system in the MVE architecture, along with tail
predication and low-overhead loops, unrolling has less benefit compared to
normal loops. You can not, for example, hide the latency of a load with other
instructions as you can for scalar code. Preventing unrolling also makes the
code easier to read and reason about.
So if a loop contains vector code, don't enable the runtime unrolling. At least
for the time being.
Differential Revision: https://reviews.llvm.org/D65803
llvm-svn: 368530
With enough codegen complete, we can now correctly report the number and size
of vector registers for MVE, allowing auto vectorisation. This also allows FP
auto-vectorization for MVE without -Ofast/-ffast-math, due to support for IEEE
FP arithmetic and parity between scalar and vector FP behaviour.
Patch by David Sherwood.
Differential Revision: https://reviews.llvm.org/D63728
llvm-svn: 368529
Summary:
Targets often have instructions that can sign-extend certain cases faster
than the equivalent shift-left/arithmetic-shift-right. Such cases can be
identified by matching a shift-left/shift-right pair but there are some
issues with this in the context of combines. For example, suppose you can
sign-extend 8-bit up to 32-bit with a target extend instruction.
%1:_(s32) = G_SHL %0:_(s32), i32 24 # (I've inlined the G_CONSTANT for brevity)
%2:_(s32) = G_ASHR %1:_(s32), i32 24
%3:_(s32) = G_ASHR %2:_(s32), i32 1
would reasonably combine to:
%1:_(s32) = G_SHL %0:_(s32), i32 24
%2:_(s32) = G_ASHR %1:_(s32), i32 25
which no longer matches the special case. If your shifts and extend are
equal cost, this would break even as a pair of shifts but if your shift is
more expensive than the extend then it's cheaper as:
%2:_(s32) = G_SEXT_INREG %0:_(s32), i32 8
%3:_(s32) = G_ASHR %2:_(s32), i32 1
It's possible to match the shift-pair in ISel and emit an extend and ashr.
However, this is far from the only way to break this shift pair and make
it hard to match the extends. Another example is that with the right
known-zeros, this:
%1:_(s32) = G_SHL %0:_(s32), i32 24
%2:_(s32) = G_ASHR %1:_(s32), i32 24
%3:_(s32) = G_MUL %2:_(s32), i32 2
can become:
%1:_(s32) = G_SHL %0:_(s32), i32 24
%2:_(s32) = G_ASHR %1:_(s32), i32 23
All upstream targets have been configured to lower it to the current
G_SHL,G_ASHR pair but will likely want to make it legal in some cases to
handle their faster cases.
To follow-up: Provide a way to legalize based on the constant. At the
moment, I'm thinking that the best way to achieve this is to provide the
MI in LegalityQuery but that opens the door to breaking core principles
of the legalizer (legality is not context sensitive). That said, it's
worth noting that looking at other instructions and acting on that
information doesn't violate this principle in itself. It's only a
violation if, at the end of legalization, a pass that checks legality
without being able to see the context would say an instruction might not be
legal. That's a fairly subtle distinction so to give a concrete example,
saying %2 in:
%1 = G_CONSTANT 16
%2 = G_SEXT_INREG %0, %1
is legal is in violation of that principle if the legality of %2 depends
on %1 being constant and/or being 16. However, legalizing to either:
%2 = G_SEXT_INREG %0, 16
or:
%1 = G_CONSTANT 16
%2:_(s32) = G_SHL %0, %1
%3:_(s32) = G_ASHR %2, %1
depending on whether %1 is constant and 16 does not violate that principle
since both outputs are genuinely legal.
Reviewers: bogner, aditya_nandakumar, volkan, aemerson, paquette, arsenm
Subscribers: sdardis, jvesely, wdng, nhaehnle, rovka, kristof.beyls, javed.absar, hiraditya, jrtc27, atanasyan, Petar.Avramovic, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61289
llvm-svn: 368487
I've now needed to add an extra parameter to this call twice recently. Not only
is the signature getting extremely unwieldy, but just updating all of the
callsites and implementations is a pain. Putting the parameters in a struct
sidesteps both issues.
llvm-svn: 368408
As loads are combined and widened, we replaced their sext users
operands whereas we should have been replacing the uses of the sext.
I've added a load of tests, with only a few of them originally
causing assertion failures, the rest improve pattern coverage.
Differential Revision: https://reviews.llvm.org/D65740
llvm-svn: 368404
This adds pre- and post- increment and decrements for MVE loads and stores. It
uses the builtin pre and post load/store detection, unlike Neon. Loads are
selected with the code in tryT2IndexedLoad, stores are selected with tablegen
patterns. The immediates have a +/-7bit range, multiplied by the size of the
element.
Differential Revision: https://reviews.llvm.org/D63840
llvm-svn: 368305
This adds some missing patterns for big endian loads/stores, allowing unaligned
loads/stores to also be selected with an extra VREV, which produces better code
than aligning through a stack. Also moves VLDR_P0 to not be LE only, and
adjusts some of the tests to show all that working.
Differential Revision: https://reviews.llvm.org/D65583
llvm-svn: 368304
VLDRH needs to have an alignment of at least 2, including the
widening/narrowing versions. This tightens up the ISel patterns for it and
alters allowsMisalignedMemoryAccesses so that unaligned accesses are expanded
through the stack. It also fixed some incorrect shift amounts, which seemed to
be passing a multiple not a shift.
Differential Revision: https://reviews.llvm.org/D65580
llvm-svn: 368256
Currently we check whether LR is stored/loaded to/from inbetween the
loop decrement and loop end pseudo instructions. There's two problems
here:
- It relies on all load/store instructions being labelled as such in
tablegen.
- Actually any use of loop decrement is troublesome because the value
doesn't exist!
So we need to check for any read/write of LR that occurs between the
two instructions and revert if we find anything.
Differential Revision: https://reviews.llvm.org/D65792
llvm-svn: 368130
Summary:
This is patch is part of a serie to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet, jfb, jakehehrlich
Reviewed By: jfb
Subscribers: wuzish, jholewinski, arsenm, dschuff, nemanjai, jvesely, nhaehnle, javed.absar, sbc100, jgravelle-google, hiraditya, aheejin, kbarton, asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, dexonsmith, PkmX, jocewei, jsji, s.egerton, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65514
llvm-svn: 367828
Add an explicit construction of the ArrayRef, gcc 5 and earlier don't
seem to select the ArrayRef constructor which takes a C array when the
construction is implicit.
Original commit message:
- Avoid a crash when IPRA calls ARMFrameLowering::determineCalleeSaves
with a null RegScavenger. Simply not updating the register scavenger
is fine because IPRA only cares about the SavedRegs vector, the acutal
code of the function has already been generated at this point.
- Add a new hook to TargetRegisterInfo to get the set of registers which
can be clobbered inside a call, even if the compiler can see both
sides, by linker-generated code.
Differential revision: https://reviews.llvm.org/D64908
llvm-svn: 367819
This adds big endian MVE patterns for bitcasts. They are defined in llvm as
being the same as a store of the existing type and the load into the new. This
means that they have to become a VREV between the two types, working in the
same way that NEON works in big-endian. This also adds some example tests for
bigendian, showing where code is and isn't different.
The main difference, especially from a testing perspective is that vectors are
passed as v2f64, and so are VREV into and out of call arguments, and the
parameters are passed in a v2f64 format. Same happens for inline assembly where
the register class is used, so it is VREV to a v16i8.
So some of this is probably not correct yet, but it is (mostly) self-consistent
and seems to be consistent with how llvm treats vectors. The rest we can
hopefully fix later. More details about big endian neon can be found in
https://llvm.org/docs/BigEndianNEON.html.
Differential Revision: https://reviews.llvm.org/D65581
llvm-svn: 367780
Fix for https://bugs.llvm.org/show_bug.cgi?id=42760. A tBR_JTr
instruction is duplicated by tail duplication, which results in
the same jumptable with the same label being emitted twice.
Fix this by marking tBR_JTr as not duplicable. The corresponding
ARM/Thumb instructions are already marked as not duplicable.
Additionally also mark tTBB_JT and tTBH_JT to be consistent with
Thumb2, even though this shouldn't be strictly necessary.
Differential Revision: https://reviews.llvm.org/D65606
llvm-svn: 367753
This optimisation isn't generally profitable for ARM, because we can
save/restore many registers in the prologue and epilogue using the PUSH
and POP instructions, but mostly use individual LDR/STR instructions for
other spills.
Differential revision: https://reviews.llvm.org/D64910
llvm-svn: 367670
- Avoid a crash when IPRA calls ARMFrameLowering::determineCalleeSaves
with a null RegScavenger. Simply not updating the register scavenger
is fine because IPRA only cares about the SavedRegs vector, the acutal
code of the function has already been generated at this point.
- Add a new hook to TargetRegisterInfo to get the set of registers which
can be clobbered inside a call, even if the compiler can see both
sides, by linker-generated code.
Differential revision: https://reviews.llvm.org/D64908
llvm-svn: 367669
The VREV64 instruction is apparently unpredictable if Qd == Qm, due to the
cross-beat nature of the instruction. This adds an earlyclobber to Qd, which
seems to be the same way we deal with this on other instructions like the
write-back on loads and stores.
Differential Revision: https://reviews.llvm.org/D65502
llvm-svn: 367544
This is extremely specific, but saves three instructions when it's
legal. I don't think the code can be usefully generalized.
Differential Revision: https://reviews.llvm.org/D65351
llvm-svn: 367492
Thumb1 has very limited immediate modes, so turning an "and" into a
shift can save multiple instructions.
It's possible to simplify the generated code for test2 and test3 in
cmp-and-fold.ll a little more, but I'll implement that as a followup.
Differential Revision: https://reviews.llvm.org/D65175
llvm-svn: 367491
Summary:
This will make it possible to improve IPRA by taking into account
register usage in indirect calls.
NFC yet; this is just laying the groundwork to start building
up patches to take advantage of the information for improved register
allocation.
Reviewers: aditya_nandakumar, volkan, qcolombet, arsenm, rovka, aemerson, paquette
Subscribers: sdardis, wdng, javed.absar, hiraditya, jrtc27, atanasyan, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65488
llvm-svn: 367476
Summary:
According to the Armv8.1-M manual CSEL, CSINC, CSINV and CSNEG are
"constrained unpredictable" when SP is used as the source register Rn.
The assembler should diagnose this case.
Reviewers: momchil.velikov, dmgreen, ostannard, simon_tatham, t.p.northover
Reviewed By: ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65505
llvm-svn: 367433
Use a switch instead of many isa<> while checking for supported
values. Also be explicit about which cast instructions are supported;
This allows the removal of SIToFP from GenerateSignBits.
llvm-svn: 367402
The code is now in a good enough state to pass the bunch of tests that
I have run (after fixing the bugs), so let's enable it by default.
Differential Revision: https://reviews.llvm.org/D65277
llvm-svn: 367297
Revert the hardware loop upon finding a LoopEnd that doesn't target
the loop header, instead of asserting a failure.
Differential Revision: https://reviews.llvm.org/D65268
llvm-svn: 367296
- Remove some unused typedefs.
- Rename BinOpChain struct to MulCandidate.
- Remove the size method of MulCandidate.
- Store only the first input of the ValueList provided to
MulCandidate, as it's the only value we care about. This means we
don't have to perform any ugly (and unnecessary) iterations of the
list later on.
llvm-svn: 367208
This adds the patterns required to transform xor P0, -1 to a VPNOT. The
instruction operands have to change a little for this, adding an in and an out
VCCR reg and using a custom DecodeMVEVPNOT for the decode.
Differential Revision: https://reviews.llvm.org/D65133
llvm-svn: 367192
These are some better patterns for converting between predicates and floating
points. Much like the extends, we select "1"/"-1" or "0" depending on the
predicate value. Or we perform a compare against 0 to convert to a predicate.
Differential Revision: https://reviews.llvm.org/D65103
llvm-svn: 367191
Both WhileLoopStart and LoopEnd may get turned into a cmp and br pair,
so add an implicit def to these pseudo instructions in case that WLS
and LE aren't generated.
Differential Revision: https://reviews.llvm.org/D65275
llvm-svn: 367089
This removes the VCEQ/VCNE/VCGE/VCEQZ/etc nodes, just using two called VCMP and
VCMPZ with an extra operand as the condition code. I believe this will make
some combines simpler, allowing us to just look at these codes and not the
operands. It also helps fill in a missing VCGTUZ MVE selection without adding
extra nodes for it.
Differential Revision: https://reviews.llvm.org/D65072
llvm-svn: 366934
The prevents us from trying to convert an i1 predicate vector to a float, or
vice-versa. Better patterns are possible, which will follow in a subsequent
commit. For now we just expand them.
Differential Revision: https://reviews.llvm.org/D65066
llvm-svn: 366931
MVE VCMP instructions can use a general purpose register as the second operand.
This adds the combines for it, selecting from a compare of a vdup.
Differential Revision: https://reviews.llvm.org/D65061
llvm-svn: 366924
This adds a DeMorgan combine for OR's of compares to turn them into AND's,
helping prevent them from going into and out of gpr registers. It also fills in
the VCLE and VCLT nodes that MVE can select, allowing it to invert more
compares.
Differential Revision: https://reviews.llvm.org/D65059
llvm-svn: 366920
Add a number of folds to convert and(vcmp, vcmp) into a single VPT block, where
the second vcmp becomes predicated on the first.
The VCMP; VPST; VCMP will eventually be converted to VPT; VCMP in the
VPTBlockPass.
Differential Revision: https://reviews.llvm.org/D65058
llvm-svn: 366910
Much like integers, this adds MVE floating point compares and select. It
requires a lot more buildvector/shuffle code because we may need to expand the
compares without mve.fp, and requires support for and/or because of the way we
lower llvm condition codes.
Some original code by David Sherwood
Differential Revision: https://reviews.llvm.org/D65054
llvm-svn: 366909
This adds some basic, "worst case" handling for MVE predicate Or/And/Xor. It
does this by going into and out of GPRs, doing the operation on scalars.
Code by David Sherwood.
Differential Revision: https://reviews.llvm.org/D65053
llvm-svn: 366907
This change make sure that llvm does not emit an invalid IT block
by putting the constant pool in the middle of an IT block.
We have code to try to avoid putting a constant island in the middle of an
IT block, but it only works if we see an IT between the one currently
referencing CPE and possible insertion point. If the first instruction
we look at is the VLDRD after the IT , we never see the IT and does not
realize that the instruction doing the load could be in an IT block itself.
Differential Revision: https://reviews.llvm.org/D64621
Change-Id: I24cecb37cded75e8992870bd997f6226853bd920
llvm-svn: 366905
This adds support code for building and shuffling i1 predicate registers. It
generally uses two basic principles, either converting the predicate into an
scalar (through a PREDICATE_CAST) and doing scalar operations on it there, or
by converting the register to an full vector register and back.
Some of the code here is a not super efficient but will hopefully cover most
cases of moving i1 vectors around and can be improved in subsequent patches.
Some code by David Sherwood.
Differential Revision: https://reviews.llvm.org/D65052
llvm-svn: 366890
This adds the very basics for MVE vector predication, adding integer VCMP and
VSEL instruction support. This is done through predicate registers (MVT::v16i1,
MVT::v8i1, MVT::v4i1), but otherwise using same mechanics as NEON to custom
lower setcc's through ARMISD::VCXX nodes (VCEQ, VCGT, VCEQZ, etc).
An extra VCNE was added, as this can be handled sensibly by MVE's expanded
number of VCMP condition codes. (There are also VCLE and VCLT which are added
later).
VPSEL is also added here, simply selecting on the vselect.
Original code by David Sherwood.
Differential Revision: https://reviews.llvm.org/D65051
llvm-svn: 366885
While combining two loads into a single load, we often need to
reorder the pointer operands for the new load. This reordering was
broken in the cases where there was a chain of values that built up
the pointer.
Differential Revision: https://reviews.llvm.org/D65193
llvm-svn: 366881
While lowering test.set.loop.iterations, it wasn't checked how the
brcond was using the result and so the wls could branch to the loop
preheader instead of not entering it. The same was true for
loop.decrement.reg.
So brcond and br_cc and now lowered manually when using the hwloop
intrinsics. During this we now check whether the result has been
negated and whether we're using SETEQ or SETNE and 0 or 1. We can
then figure out which basic block the WLS and LE should be targeting.
Differential Revision: https://reviews.llvm.org/D64616
llvm-svn: 366809
ARMLowOverheadLoops would assert a failure if it did not find all the
pseudo instructions that comprise the hardware loop. Instead of doing
this, iterate through all the instructions of the function and revert
any remaining pseudo instructions that haven't been converted.
Differential Revision: https://reviews.llvm.org/D65080
llvm-svn: 366691
We need to ensure that the number of T's is correct when adding multiple
instructions into the same VPT block.
Differential revision: https://reviews.llvm.org/D65049
llvm-svn: 366684
ARM has code to recognise uses of the "returned" function parameter
attribute which guarantee that the value passed to the function in r0
will be returned in r0 unmodified. IPRA replaces the regmask on call
instructions, so needs to be told about this to avoid reverting the
optimisation.
Differential revision: https://reviews.llvm.org/D64986
llvm-svn: 366669
Summary:
According to the new Armv8-M specification
https://static.docs.arm.com/ddi0553/bh/DDI0553B_h_armv8m_arm.pdf the
instructions SQRSHRL and UQRSHLL now have an additional immediate
operand <saturate>. The new assembly syntax is:
SQRSHRL<c> RdaLo, RdaHi, #<saturate>, Rm
UQRSHLL<c> RdaLo, RdaHi, #<saturate>, Rm
where <saturate> can be either 64 (the existing behavior) or 48, in
that case the result is saturated to 48 bits.
The new operand is encoded as follows:
#64 Encoded as sat = 0
#48 Encoded as sat = 1
sat is bit 7 of the instruction bit pattern.
This patch adds a new assembler operand class MveSaturateOperand which
implements parsing and encoding. Decoding is implemented in
DecodeMVEOverlappingLongShift.
Reviewers: ostannard, simon_tatham, t.p.northover, samparker, dmgreen, SjoerdMeijer
Reviewed By: simon_tatham
Subscribers: javed.absar, kristof.beyls, hiraditya, pbarrio, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64810
llvm-svn: 366555
Summary:
PerformVMOVRRDCombine ommits adding a offset
of 4 to the PointerInfo, when converting a
f64 = load[M]
to
{i32, i32} = {load[M], load[M + 4]}
Which would allow the machine scheduller
to break dependencies with the second load.
- pr42638
Reviewers: eli.friedman, dmgreen, ostannard
Reviewed By: ostannard
Subscribers: ostannard, javed.absar, kristof.beyls, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64870
llvm-svn: 366423
Migrate CallLowering::lowerReturnVal to use the same infrastructure as
lowerCall/FormalArguments and remove the now obsolete code path from
splitToValueTypes.
Forgot to push this earlier.
llvm-svn: 366308
We need to make sure that we are sensibly dealing with vectors of types v2i64
and v2f64, even if most of the time we cannot generate native operations for
them. This mostly adds a lot of testing, plus fixes up a couple of the issues
found. And, or and xor can be legal for v2i64, and shifts combining needs a
slight fixup.
Differential Revision: https://reviews.llvm.org/D64316
llvm-svn: 366106
This adds basic lowering for MVE shifts. There are many shifts in MVE, but the
instructions handled here are:
VSHL (imm)
VSHRu (imm)
VSHRs (imm)
VSHL (vector)
VSHL (register)
MVE, like NEON before it, doesn't have shift right by a vector (or register).
We instead have to negate the amount and shift in the opposite direction. This
means we have to convert any SHR's into a form of SHL (that is still signed or
unsigned) with a negated condition and selecting from there. MVE still does
have shifting by an immediate for SHL, ASR and LSR.
This adds lowering for these and for register forms, which work well for shift
lefts but may require an extra fold of neg(vdup(x)) -> vdup(neg(x)) to potentially
work optimally for right shifts.
Differential Revision: https://reviews.llvm.org/D64212
llvm-svn: 366056
This just moves the shift instruction definitions further down the
ARMInstrMVE.td file, to make positioning patterns slightly more natural.
llvm-svn: 366054
This adjusts the way that we lower NEON shifts to use a DAG target node, not
via a neon intrinsic. This is useful for handling MVE shifts operations in the
same the way. It also renames some of the immediate shift nodes for
consistency, and moves some of the processing of immediate shifts into
LowerShift allowing it to capture more cases.
Differential Revision: https://reviews.llvm.org/D64426
llvm-svn: 366051
The vmovlb instructions can be uses to sign or zero extend vector registers
between types. This adds some patterns for them and relevant testing. The
VBICIMM generation is also put behind a hasNEON check (as is already done for
VORRIMM).
Code originally by David Sherwood.
Differential Revision: https://reviews.llvm.org/D64069
llvm-svn: 366008
This selects integer VNEG instructions, which can be especially useful with shifts.
Differential Revision: https://reviews.llvm.org/D64204
llvm-svn: 366006
This simply makes the MVE integer min and max instructions legal and adds the
relevant patterns for them.
Differential Revision: https://reviews.llvm.org/D64026
llvm-svn: 366004
This adds support for the floor/ceil/trunc/... series of instructions,
converting to various forms of VRINT. They use the same suffixes as their
floating point counterparts. There is not VTINTR, so nearbyint is expanded.
Also added a copysign test, to show it is expanded.
Differential Revision: https://reviews.llvm.org/D63985
llvm-svn: 366003
This adds the patterns for minnm and maxnm from the fminnum and fmaxnum nodes,
similar to scalar types.
Original patch by Simon Tatham
Differential Revision: https://reviews.llvm.org/D63870
llvm-svn: 366002
This patch addresses a couple of problems:
1) The maximum supported offset of LE is -4094.
2) The offset of WLS also needs to be checked, this uses a
maximum positive offset of 4094.
The use of BasicBlockUtils has been changed because the block offsets
weren't being initialised, but the isBBInRange checks both positive
and negative offsets.
ARMISelLowering has been tweaked because the test case presented
another pattern that we weren't supporting.
llvm-svn: 365749
The VQDMLAH.U8, VQDMLAH.U16 and VQDMLAH.U32 instructions don't
actually exist: the Armv8.1-M architecture spec only lists signed
forms of that instruction. The unsigned ones were added in error: they
existed in an early draft of the spec, but they were removed before
the public version, and we missed that particular spec change.
Also affects the variant forms VQDMLASH, VQRDMLAH and VQRDMLASH.
Reviewers: miyuki
Subscribers: javed.absar, kristof.beyls, hiraditya, dmgreen, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64502
llvm-svn: 365747
Two functional changes have been made here:
- Now search up from any add instruction to find the chains of
operations that we may turn into a smlad. This allows the
generation of a smlad which doesn't accumulate into a phi.
- The search function has been corrected to stop it falsely searching
up through an invalid path.
The bulk of the changes have been making the Reduction struct a class
and making it more C++y with getters and setters.
Differential Revision: https://reviews.llvm.org/D61780
llvm-svn: 365740
Summary:
Use the same predicates as VSTMDB/VLDMIA since VPUSH/VPOP alias to
these.
Patch by Momchil Velikov.
Reviewers: ostannard, simon_tatham, SjoerdMeijer, samparker, t.p.northover, dmgreen
Reviewed By: dmgreen
Subscribers: javed.absar, kristof.beyls, hiraditya, dmgreen, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64413
llvm-svn: 365604
Summary:
According to a recently updated Armv8-M spec
(https://static.docs.arm.com/ddi0553/bh/DDI0553B_h_armv8m_arm.pdf) the
32-bit width versions of the following instructions:
* VQDMLADH
* VQDMLADHX
* VQRDMLADH
* VQRDMLADHX
* VQDMLSDH
* VQDMLSDHX
* VQRDMLSDH
* VQRDMLSDHX
are no longer unpredictable when their output register is the same as
one of the input registers.
This patch updates the assembler parser and the corresponding tests
and also removes @earlyclobber from the instruction constraints.
Reviewers: simon_tatham, ostannard, dmgreen, SjoerdMeijer, samparker
Reviewed By: simon_tatham
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64250
llvm-svn: 365306
This adds some handling for VMOVimm, using the same method that NEON uses. We
create VMOVIMM/VMVNIMM/VMOVFPIMM nodes based on the immediate, and select them
using the now renamed ARMvmovImm/etc. There is also an extra 64bit immediate
mode that I have not yet added here.
Code by David Sherwood
Differential Revision: https://reviews.llvm.org/D63884
llvm-svn: 365178
The arm condition codes for GE is N==V (and for LT is N!=V). If the source of
flags cannot set V (overflow), such as a cmp against #0, then we can use the
simpler PL and MI conditions that only check N. As these PL/MI conditions are
simpler than GE/LT, other passes like the peephole optimiser can have a better
time optimising away the redundant CMPs.
The exception is the VSEL instruction, which cannot take the PL code, so there
the transform favours GE.
Differential Revision: https://reviews.llvm.org/D64160
llvm-svn: 365117
This adds patterns for the simpler VAND, VORR and VEOR bitwise vector
instructions. It also adjusts the top16Zero PatLeaf to not match on vector
instructions, which can otherwise cause problems.
Code written by David Sherwood.
Differential Revision: https://reviews.llvm.org/D63867
llvm-svn: 365113
For Thumb2, we prefer low regs (costPerUse = 0) to allow narrow
encoding. However, current allocation order is like:
R0-R3, R12, LR, R4-R11
As a result, a lot of instructs that use R12/LR will be wide instrs.
This patch changes the allocation order to:
R0-R7, R12, LR, R8-R11
for thumb2 and -Osize.
In most cases, there is no extra push/pop instrs as they will be folded
into existing ones. There might be slight performance impact due to more
stack usage, so we only enable it when opt for min size.
https://reviews.llvm.org/D30324
llvm-svn: 365014
Summary:
This is the backend part of [[ https://bugs.llvm.org/show_bug.cgi?id=42457 | PR42457 ]].
In middle-end, we'd want to prefer the form with two adds - D63992,
but as this diff shows, not every target will prefer that pattern.
Out of 4 targets for which i added tests all seem to be ok with inc-of-add for scalars,
but only X86 prefer that same pattern for vectors.
Here i'm adding a new TLI hook, always defaulting to the inc-of-add,
but adding AArch64,ARM,PowerPC overrides to prefer inc-of-add only for scalars.
Reviewers: spatel, RKSimon, efriedma, t.p.northover, hfinkel
Reviewed By: efriedma
Subscribers: nemanjai, javed.absar, kristof.beyls, kbarton, jsji, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64090
llvm-svn: 365010
There were two issues here: one, some of the relevant instructions were
missing the expected "FrameSetup" flag, and two,
ARMAsmPrinter::EmitUnwindingInstruction wasn't expecting "mov"
instructions in the prologue.
I'm sticking the additional state into ARMFunctionInfo so it's obvious
it only applies to the current function.
I considered a few alternative approaches where we would compute the
correct unwind information as part of the prologue/epilogue lowering,
but it seems like a lot of work to introduce pseudo-instructions, and
the current code seems to be reliable enough.
Fixes https://bugs.llvm.org/show_bug.cgi?id=42408.
Differential Revision: https://reviews.llvm.org/D63964
llvm-svn: 364970
Passing a vector type over the soft-float ABI involves it being split
into four GPRs, so the first thing that has to happen at the start of
the function is to recombine those into a vector register. The ABI
types all vectors as v2f64, so we need to support BUILD_VECTOR for
that type, which I do in this patch by allowing it to be expanded in
terms of INSERT_VECTOR_ELT, and writing an ISel pattern for that in
turn. Similarly, I provide a rule for EXTRACT_VECTOR_ELT so that a
returned vector can be marshalled back into GPRs.
While I'm here, I've also added ISD::UNDEF to the list of operations
we turn back on in `setAllExpand`, because I noticed that otherwise it
gets expanded into a BUILD_VECTOR with explicit zero inputs, leading
to pointless machine instructions to zero out a vector register that's
about to have every lane overwritten of in any case.
Reviewers: dmgreen, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63937
llvm-svn: 364910
If you compile with `-mattr=+mve` (enabling integer MVE instructions
but not floating-point ones), then the scalar FP //registers// exist
and it's legal to move things in and out of them, load and store them,
but it's not legal to do arithmetic on them.
In D60708, the calls to `addRegisterClass` in ARMISelLowering that
enable use of the scalar FP registers became conditionalised on
`Subtarget->hasFPRegs()` instead of `Subtarget->hasVFP2Base()`, so
that loads, stores and moves of those registers would work. But I
didn't realise that that would also enable all the operations on those
types by default.
Now, if the target doesn't have basic VFP, we follow up those
`addRegisterClass` calls by turning back off all the nontrivial
operations you can perform on f32 and f64. That causes several
knock-on failures, which are fixed by allowing the `VMOVDcc` and
`VMOVScc` instructions to be selected even if all you have is
`HasFPRegs`, and adjusting several checks for 'is this a double in a
single-precision-only world?' to the more general 'is this any FP type
we can't do arithmetic on?'. Between those, the whole of the
`float-ops.ll` and `fp16-instructions.ll` tests can now run in
MVE-without-FP mode and generate correct-looking code.
One odd side effect is that I had to relax the check lines in that
test so that they permit test functions like `add_f` to be generated
as tailcalls to software FP library functions, instead of ordinary
calls. Doing that is entirely legal, but the mystery is why this is
the first RUN line that's needed the relaxation: on the usual kind of
non-FP target, no tailcalls ever seem to be generated. Going by the
llc messages, I think `SoftenFloatResult` must be perturbing the code
generation in some way, but that's as much as I can guess.
Reviewers: dmgreen, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63938
llvm-svn: 364909
Summary:
According to the ARMARM, the VQDMLADH, VQRDMLADH, VQDMLSDH and
VQRDMLSDH instructions handle their results as follows: "The base
variant writes the results into the lower element of each pair of
elements in the destination register, whereas the exchange variant
writes to the upper element in each pair". I.e., the initial content
of the output register affects the result, as usual, we model this
with an additional input.
Also, for 32-bit variants Qd is not allowed to be the same register as
Qm and Qn, we use @earlyclobber to indicate this.
This patch also changes vpred_r to vpred_n because the instructions
don't have an explicit 'inactive' operand.
Reviewers: dmgreen, ostannard, simon_tatham
Reviewed By: simon_tatham
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D64007
llvm-svn: 364796
Backend changes to enable WLS/LE low-overhead loops for armv8.1-m:
1) Use TTI to communicate to the HardwareLoop pass that we should try
to generate intrinsics that guard the loop entry, as well as setting
the loop trip count.
2) Lower the BRCOND that uses said intrinsic to an Arm specific node:
ARMWLS.
3) ISelDAGToDAG the node to a new pseudo instruction:
t2WhileLoopStart.
4) Add support in ArmLowOverheadLoops to handle the new pseudo
instruction.
Differential Revision: https://reviews.llvm.org/D63816
llvm-svn: 364733
MVE adds the lsll, lsrl and asrl instructions, which perform a shift on a 64 bit value separated into two 32 bit registers.
The Expand64BitShift function is modified to accept ISD::SHL, ISD::SRL and ISD::SRA and convert it into the appropriate opcode in ARMISD. An SHL is converted into an lsll, an SRL is converted into an lsrl for the immediate form and a negation and lsll for the register form, and SRA is converted into an asrl.
test/CodeGen/ARM/shift_parts.ll is added to test the logic of emitting these instructions.
Differential Revision: https://reviews.llvm.org/D63430
llvm-svn: 364654
This simply adds integer and floating point VMUL patterns for MVE, same as we
have add and sub.
Differential Revision: https://reviews.llvm.org/D63866
llvm-svn: 364643
This adds handling and tests for a number of floating point math routines,
which have no MVE instructions.
Differential Revision: https://reviews.llvm.org/D63725
llvm-svn: 364641
MVE has instructions to widen as it loads, and narrow as it stores. This adds
the required patterns and legalisation to make them work including specifying
that they are legal, patterns to select them and test changes.
Patch by David Sherwood.
Differential Revision: https://reviews.llvm.org/D63839
llvm-svn: 364636
This fills in the gaps for basic MVE loads and stores, allowing unaligned
access and adding far too many tests. These will become important as
narrowing/expanding and pre/post inc are added. Big endian might still not be
handled very well, because we have not yet added bitcasts (and I'm not sure how
we want it to work yet). I've included the alignment code anyway which maps
with our current patterns. We plan to return to that later.
Code written by Simon Tatham, with additional tests from Me and Mikhail Maltsev.
Differential Revision: https://reviews.llvm.org/D63838
llvm-svn: 364633
We don't have vector operations for these, so they need to be expanded for both
integer and float.
Differential Revision: https://reviews.llvm.org/D63595
llvm-svn: 364631
The same as integer arithmetic, we can add simple floating point MVE addition and
subtraction patterns.
Initial code by David Sherwood
Differential Revision: https://reviews.llvm.org/D63257
llvm-svn: 364629
This adds the first few patterns for MVE code generation, adding simple integer
add and sub patterns.
Initial code by David Sherwood
Differential Revision: https://reviews.llvm.org/D63255
llvm-svn: 364627
This patch adds necessary shuffle vector and buildvector support for ARM MVE.
It essentially adds support for VDUP, VREVs and some VMOVs, which are often
required by other code (like upcoming patches).
This mostly uses the same code from Neon that already generated
NEONvdup/NEONvduplane/NEONvrev's. These have been renamed to ARMvdup/etc and
moved to ARMInstrInfo as they are common to both architectures. Most of the
selection code seems to be applicable to both, but NEON does have some more
instructions making some parts specific.
Most code originally by David Sherwood.
Differential Revision: https://reviews.llvm.org/D63567
llvm-svn: 364626
The code to generate register move instructions in and out of VPR and
FPSCR_NZCV had assertions checking that the other register involved
was a GPR _pair_, instead of a single GPR as it should have been.
Reviewers: miyuki, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63865
llvm-svn: 364534
The BF and WLS/WLSTP instructions have various branch-offset fields
occupying different positions and lengths in the instruction encoding,
and all of them were decoded at disassembly time by the function
DecodeBFLabelOffset() which returned SoftFail if the offset was zero.
In fact, it's perfectly fine and not even a SoftFail for most of those
offset fields to be zero. The only one that can't be zero is the 4-bit
field labelled `boff` in the architecture spec, occupying bits {26-23}
of the BF instruction family. If that one is zero, the encoding
overlaps other instructions (WLS, DLS, LETP, VCTP), so it ought to be
a full Fail.
Fixed by adding an extra template parameter to DecodeBFLabelOffset
which controls whether a zero offset is accepted or rejected. Adjusted
existing tests (only in error messages for bad disassemblies); added
extra tests to demonstrate zero offsets being accepted in all the
right places, and a few demonstrating rejection of zero `boff`.
Reviewers: DavidSpickett, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63864
llvm-svn: 364533
Different versions of the Arm architecture disallow the use of generic
coprocessor instructions like MCR and CDP on different sets of
coprocessors. This commit centralises the check of the coprocessor
number so that it's consistent between assembly and disassembly, and
also updates it for the new restrictions in Arm v8.1-M.
New tests added that check all the coprocessor numbers; old tests
updated, where they used a number that's now become illegal in the
context in question.
Reviewers: DavidSpickett, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63863
llvm-svn: 364532
In the `CSEL Rd,Rm,Rn` instruction family (also including CSINC, CSINV
and CSNEG), the architecture lists it as CONSTRAINED UNPREDICTABLE
(i.e. SoftFail) to use SP in the Rd or Rm slot, but outright illegal
to use it in the Rn slot, not least because some encodings of that
form are used by MVE instructions such as UQRSHLL.
MC was treating all three slots the same, as SoftFail. So the only
reason UQRSHLL was disassembled correctly at all was because the MVE
decode table is separate from the Thumb2 one and takes priority; if
you turned off MVE, then encodings such as `[0x5f,0xea,0x0d,0x83]`
would disassemble as spurious CSELs.
Fixed by inventing another version of the `GPRwithZR` register class,
which disallows SP completely instead of just SoftFailing it.
Reviewers: DavidSpickett, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63862
llvm-svn: 364531
Change the interface of CallLowering::lowerCall to accept several
virtual registers for each argument, instead of just one. This is a
follow-up to D46018.
CallLowering::lowerReturn was similarly refactored in D49660 and
lowerFormalArguments in D63549.
With this change, we no longer pack the virtual registers generated for
aggregates into one big lump before delegating to the target. Therefore,
the target can decide itself whether it wants to handle them as separate
pieces or use one big register.
ARM and AArch64 have been updated to use the passed in virtual registers
directly, which means we no longer need to generate so many
merge/extract instructions.
NFCI for AMDGPU, Mips and X86.
Differential Revision: https://reviews.llvm.org/D63551
llvm-svn: 364512
Change the interface of CallLowering::lowerCall to accept several
virtual registers for the call result, instead of just one. This is a
follow-up to D46018.
CallLowering::lowerReturn was similarly refactored in D49660 and
lowerFormalArguments in D63549.
With this change, we no longer pack the virtual registers generated for
aggregates into one big lump before delegating to the target. Therefore,
the target can decide itself whether it wants to handle them as separate
pieces or use one big register.
ARM and AArch64 have been updated to use the passed in virtual registers
directly, which means we no longer need to generate so many
merge/extract instructions.
NFCI for AMDGPU, Mips and X86.
Differential Revision: https://reviews.llvm.org/D63550
llvm-svn: 364511
Change the interface of CallLowering::lowerFormalArguments to accept
several virtual registers for each formal argument, instead of just one.
This is a follow-up to D46018.
CallLowering::lowerReturn was similarly refactored in D49660. lowerCall
will be refactored in the same way in follow-up patches.
With this change, we forward the virtual registers generated for
aggregates to CallLowering. Therefore, the target can decide itself
whether it wants to handle them as separate pieces or use one big
register. We also copy the pack/unpackRegs helpers to CallLowering to
facilitate this.
ARM and AArch64 have been updated to use the passed in virtual registers
directly, which means we no longer need to generate so many
merge/extract instructions.
AArch64 seems to have had a bug when lowering e.g. [1 x i8*], which was
put into a s64 instead of a p0. Added a test-case which illustrates the
problem more clearly (it crashes without this patch) and fixed the
existing test-case to expect p0.
AMDGPU has been updated to unpack into the virtual registers for
kernels. I think the other code paths fall back for aggregates, so this
should be NFC.
Mips doesn't support aggregates yet, so it's also NFC.
x86 seems to have code for dealing with aggregates, but I couldn't find
the tests for it, so I just added a fallback to DAGISel if we get more
than one virtual register for an argument.
Differential Revision: https://reviews.llvm.org/D63549
llvm-svn: 364510
Allow CallLowering::ArgInfo to contain more than one virtual register.
This is useful when passes split aggregates into several virtual
registers, but need to also provide information about the original type
to the call lowering. Used in follow-up patches.
Differential Revision: https://reviews.llvm.org/D63548
llvm-svn: 364509
The current implementation of ThumbRegisterInfo::saveScavengerRegister
is bad for two reasons: one, it's buggy, and two, it blocks using R12
for other optimizations. So this patch gets rid of it, and adds the
necessary support for using an ordinary emergency spill slot on Thumb1.
(Specifically, I think saveScavengerRegister was broken by r305625, and
nobody noticed for two years because the codepath is almost never used.
The new code will also probably not be used much, but it now has better
tests, and if we fail to emit a necessary emergency spill slot we get a
reasonable error message instead of a miscompile.)
A rough outline of the changes in the patch:
1. Gets rid of ThumbRegisterInfo::saveScavengerRegister.
2. Modifies ARMFrameLowering::determineCalleeSaves to allocate an
emergency spill slot for Thumb1.
3. Implements useFPForScavengingIndex, so the emergency spill slot isn't
placed at a negative offset from FP on Thumb1.
4. Modifies the heuristics for allocating an emergency spill slot to
support Thumb1. This includes fixing ExtraCSSpill so we don't try to
use "lr" as a substitute for allocating an emergency spill slot.
5. Allocates a base pointer in more cases, so the emergency spill slot
is always accessible.
6. Modifies ARMFrameLowering::ResolveFrameIndexReference to compute the
right offset in the new cases where we're forcing a base pointer.
7. Ensures we never generate a load or store with an offset outside of
its frame object. This makes the heuristics more straightforward.
8. Changes Thumb1 prologue and epilogue emission so it never uses
register scavenging.
Some of the changes to the emergency spill slot heuristics in
determineCalleeSaves affect ARM/Thumb2; hopefully, they should allow
the compiler to avoid allocating an emergency spill slot in cases
where it isn't necessary. The rest of the changes should only affect
Thumb1.
Differential Revision: https://reviews.llvm.org/D63677
llvm-svn: 364490
Summary:
The getFixupKindContainerSizeBytes function returns the size of the
instruction containing a given fixup. Currently fixup_arm_pcrel_9 is
not handled in this function, this causes an assertion failure in
the debug build and incorrect codegen in the release build.
This patch fixes the problem.
Reviewers: ostannard, simon_tatham
Reviewed By: ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, pbarrio, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63778
llvm-svn: 364404
"To" selects an odd-numbered GPR, and "Te" an even one. There are some
8.1-M instructions that have one too few bits in their register fields
and require registers of particular parity, without necessarily using
a consecutive even/odd pair.
Also, the constraint letter "t" should select an MVE q-register, when
MVE is present. This didn't need any source changes, but some extra
tests have been added.
Reviewers: dmgreen, samparker, SjoerdMeijer
Subscribers: javed.absar, eraman, kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D60709
llvm-svn: 364331
This provides the low-level support to start using MVE vector types in
LLVM IR, loading and storing them, passing them to __asm__ statements
containing hand-written MVE vector instructions, and *if* you have the
hard-float ABI turned on, using them as function parameters.
(In the soft-float ABI, vector types are passed in integer registers,
and combining all those 32-bit integers into a q-reg requires support
for selection DAG nodes like insert_vector_elt and build_vector which
aren't implemented yet for MVE. In fact I've also had to add
`arm_aapcs_vfpcc` to a couple of existing tests to avoid that
problem.)
Specifically, this commit adds support for:
* spills, reloads and register moves for MVE vector registers
* ditto for the VPT predication mask that lives in VPR.P0
* make all the MVE vector types legal in ISel, and provide selection
DAG patterns for BITCAST, LOAD and STORE
* make loads and stores of scalar FP types conditional on
`hasFPRegs()` rather than `hasVFP2Base()`. As a result a few
existing tests needed their llc command lines updating to use
`-mattr=-fpregs` as their method of turning off all hardware FP
support.
Reviewers: dmgreen, samparker, SjoerdMeijer
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60708
llvm-svn: 364329
The expensive buildbots highlighted the mir tests were broken, which
I've now updated and added --verify-machineinstrs to them. This also
uncovered a couple of bugs in the backend pass, so these have also
been fixed.
llvm-svn: 364323
Including both 'case ARM_AM::uxtw' and 'default' in the getShiftOp
switch caused a buildbot to fail with
error: default label in switch which covers all enumeration values [-Werror,-Wcovered-switch-default]
llvm-svn: 364300
A minor iteration on the MVE VPT Block pass to enable more efficient VPT Block
code generation: consecutive VPT predicated statements, predicated on the same
condition, will be placed within the same VPT Block. This essentially is also
an exercise to write some more tests for the next step, which should be more
generic also merging instructions when they are not consecutive.
Differential Revision: https://reviews.llvm.org/D63711
llvm-svn: 364298
If an FP_EXTEND or FP_ROUND isel dag node converts directly between
f16 and f32 when the target CPU has no instruction to do it in one go,
it has to be done in two steps instead, going via f32.
Previously, this was done implicitly, because all such CPUs had the
storage-only implementation of f16 (i.e. the only thing you can do
with one at all is to convert it to/from f32). So isel would legalize
the f16 into an f32 as soon as it saw it, by inserting an fp16_to_fp
node (or vice versa), and then the fp_extend would already be f32->f64
rather than f16->f64.
But that technique can't support a target CPU which has full f16
support but _not_ f64, such as some variants of Arm v8.1-M. So now we
provide custom lowering for FP_EXTEND and FP_ROUND, which checks
support for f16 and f64 and decides on the best thing to do given the
combination of flags it gets back.
Reviewers: dmgreen, samparker, SjoerdMeijer
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60692
llvm-svn: 364294
This final batch includes the tail-predicated versions of the
low-overhead loop instructions (LETP); the VPSEL instruction to select
between two vector registers based on the predicate mask without
having to open a VPT block; and VPNOT which complements the predicate
mask in place.
Reviewers: dmgreen, samparker, SjoerdMeijer, t.p.northover
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62681
llvm-svn: 364292
This adds the rest of the vector memory access instructions. It
includes contiguous loads/stores, with an ordinary addressing mode
such as [r0,#offset] (plus writeback variants); gather loads and
scatter stores with a scalar base address register and a vector of
offsets from it (written [r0,q1] or similar); and gather/scatters with
a vector of base addresses (written [q0,#offset], again with
writeback). Additionally, some of the loads can widen each loaded
value into a larger vector lane, and the corresponding stores narrow
them again.
To implement these, we also have to add the addressing modes they
need. Also, in AsmParser, the `isMem` query function now has
subqueries `isGPRMem` and `isMVEMem`, according to which kind of base
register is used by a given memory access operand.
I've also had to add an extra check in `checkTargetMatchPredicate` in
the AsmParser, without which our last-minute check of `rGPR` register
operands against SP and PC was failing an assertion because Tablegen
had inserted an immediate 0 in place of one of a pair of tied register
operands. (This matches the way the corresponding check for `MCK_rGPR`
in `validateTargetOperandClass` is guarded.) Apparently the MVE load
instructions were the first to have ever triggered this assertion, but
I think only because they were the first to have a combination of the
usual Arm pre/post writeback system and the `rGPR` class in particular.
Reviewers: dmgreen, samparker, SjoerdMeijer, t.p.northover
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62680
llvm-svn: 364291
Introduce three pseudo instructions to be used during DAG ISel to
represent v8.1-m low-overhead loops. One maps to set_loop_iterations
while loop_decrement_reg is lowered to two, so that we can separate
the decrement and branching operations. The pseudo instructions are
expanded pre-emission, where we can still decide whether we actually
want to generate a low-overhead loop, in a new pass:
ARMLowOverheadLoops. The pass currently bails, reverting to an sub,
icmp and br, in the cases where a call or stack spill/restore happens
between the decrement and branching instructions, or if the loop is
too large.
Differential Revision: https://reviews.llvm.org/D63476
llvm-svn: 364288
Guillaume Chatelet