Select zero extending and sign extending load for MIPS32.
Use size from MachineMemOperand to determine number of bytes to load.
Differential Revision: https://reviews.llvm.org/D57099
llvm-svn: 352038
Use CombinerHelper to combine extending load instructions.
G_LOAD combined with G_ZEXT, G_SEXT or G_ANYEXT gives G_ZEXTLOAD,
G_SEXTLOAD or G_LOAD with same type as def of extending instruction
respectively.
Similarly G_ZEXTLOAD combined with G_ZEXT gives G_ZEXTLOAD and
G_SEXTLOAD combined with G_SEXT gives G_SEXTLOAD with same type
as def of extending instruction.
Differential Revision: https://reviews.llvm.org/D56914
llvm-svn: 352037
This reapplies commit r351987 with a failed test fix. Now the test
accepts both DW_OP_GNU_push_tls_address and DW_OP_form_tls_address
opcode.
Original commit message:
```
This is a fix for a regression introduced by the rL348194 commit. In
that change new type (MEK_DTPREL) of MipsMCExpr expression was added,
but in some places of the code this type of expression considered as
unexpected.
This change fixes the bug. The MEK_DTPREL type of expression is used for
marking TLS DIEExpr only and contains a regular sub-expression. Where we
need to handle the expression, we retrieve the sub-expression and
handle it in a common way.
```
llvm-svn: 352034
After creating new PHI instructions during isel pseudo expansion, the NoPHIs
property of MF should be reset in case it was previously set.
Review: Ulrich Weigand
llvm-svn: 352030
This is a fix for a regression introduced by the rL348194 commit. In
that change new type (MEK_DTPREL) of MipsMCExpr expression was added,
but in some places of the code this type of expression considered as
unexpected.
This change fixes the bug. The MEK_DTPREL type of expression is used for
marking TLS DIEExpr only and contains a regular sub-expression. Where we
need to handle the expression, we retrieve the sub-expression and
handle it in a common way.
llvm-svn: 351987
Enable full support for the debug info. Recommit to fix the emission of
the not required closing brace.
Differential revision: https://reviews.llvm.org/D46189
llvm-svn: 351972
This patch adds a new ReadAdvance definition named ReadInt2Fpu.
ReadInt2Fpu allows x86 scheduling models to accurately describe delays caused by
data transfers from the integer unit to the floating point unit.
ReadInt2Fpu currently defaults to a delay of zero cycles (i.e. no delay) for all
x86 models excluding BtVer2. That means, this patch is only a functional change
for the Jaguar cpu model only.
Tablegen definitions for instructions (V)PINSR* have been updated to account for
the new ReadInt2Fpu. That read is mapped to the the GPR input operand.
On Jaguar, int-to-fpu transfers are modeled as a +6cy delay. Before this patch,
that extra delay was added to the opcode latency. In practice, the insert opcode
only executes for 1cy. Most of the actual latency is actually contributed by the
so-called operand-latency. According to the AMD SOG for family 16h, (V)PINSR*
latency is defined by expression f+1, where f is defined as a forwarding delay
from the integer unit to the fpu.
When printing instruction latency from MCA (see InstructionInfoView.cpp) and LLC
(only when flag -print-schedule is speified), we now need to account for any
extra forwarding delays. We do this by checking if scheduling classes declare
any negative ReadAdvance entries. Quoting a code comment in TargetSchedule.td:
"A negative advance effectively increases latency, which may be used for
cross-domain stalls". When computing the instruction latency for the purpose of
our scheduling tests, we now add any extra delay to the formula. This avoids
regressing existing codegen and mca schedule tests. It comes with the cost of an
extra (but very simple) hook in MCSchedModel.
Differential Revision: https://reviews.llvm.org/D57056
llvm-svn: 351965
Summary:
With XNACK, an smem load whose result is coalesced with an operand (thus
it overwrites its own operand) cannot appear in a clause, because some
other instruction might XNACK and restart the whole clause.
The clause breaker already realized that an smem that overwrites an
operand cannot appear in a clause, and broke the clause. The problem
that this commit fixes is that the SIFormMemoryClauses optimization
formed a bundle with early clobber, which caused the earlier code that
set up the coalesced operand to be removed as dead.
Differential Revision: https://reviews.llvm.org/D57008
Change-Id: I703c4d5b0bf7d6060222bec491f45c18bb3c0016
llvm-svn: 351950
Currently in Arm code, we allocate LR first, under the assumption that
it needs to be saved anyway. Unfortunately this has the disadvantage
that it will require any instructions using it to be the longer thumb2
instructions, not the shorter thumb1 ones.
This switches the order when we are optimising for minsize, returning to
the default order so that more lower registers can be used. It can end
up requiring more pushed registers, but on average produces smaller code.
Differential Revision: https://reviews.llvm.org/D56008
llvm-svn: 351938
In the last stage of type promotion, we replace any zext that uses a
new trunc with the operand of the trunc. This is okay when we only
allowed one type to be optimised, but now its the case that the trunc
maybe needed to produce a more narrow type than the one we were
optimising for. So we need to check this before doing the replacement.
Differential Revision: https://reviews.llvm.org/D57041
llvm-svn: 351935
As part of speculation hardening, the stack pointer gets masked with the
taint register (X16) before a function call or before a function return.
Since there are no instructions that can directly mask writing to the
stack pointer, the stack pointer must first be transferred to another
register, where it can be masked, before that value is transferred back
to the stack pointer.
Before, that temporary register was always picked to be x17, since the
ABI allows clobbering x17 on any function call, resulting in the
following instruction pattern being inserted before function calls and
returns/tail calls:
mov x17, sp
and x17, x17, x16
mov sp, x17
However, x17 can be live in those locations, for example when the call
is an indirect call, using x17 as the target address (blr x17).
To fix this, this patch looks for an available register just before the
call or terminator instruction and uses that.
In the rare case when no register turns out to be available (this
situation is only encountered twice across the whole test-suite), just
insert a full speculation barrier at the start of the basic block where
this occurs.
Differential Revision: https://reviews.llvm.org/D56717
llvm-svn: 351930
Two backend optimizations failed to handle cases when compiled with -g, due
to failing to consider DBG_VALUE instructions. This was in
SystemZTargetLowering::emitSelect() and
SystemZElimCompare::getRegReferences().
This patch makes sure that DBG_VALUEs are recognized so that they do not
affect these optimizations.
Tests for branch-on-count, load-and-trap and consecutive selects.
Review: Ulrich Weigand
https://reviews.llvm.org/D57048
llvm-svn: 351928
Each hwasan check requires emitting a small piece of code like this:
https://clang.llvm.org/docs/HardwareAssistedAddressSanitizerDesign.html#memory-accesses
The problem with this is that these code blocks typically bloat code
size significantly.
An obvious solution is to outline these blocks of code. In fact, this
has already been implemented under the -hwasan-instrument-with-calls
flag. However, as currently implemented this has a number of problems:
- The functions use the same calling convention as regular C functions.
This means that the backend must spill all temporary registers as
required by the platform's C calling convention, even though the
check only needs two registers on the hot path.
- The functions take the address to be checked in a fixed register,
which increases register pressure.
Both of these factors can diminish the code size effect and increase
the performance hit of -hwasan-instrument-with-calls.
The solution that this patch implements is to involve the aarch64
backend in outlining the checks. An intrinsic and pseudo-instruction
are created to represent a hwasan check. The pseudo-instruction
is register allocated like any other instruction, and we allow the
register allocator to select almost any register for the address to
check. A particular combination of (register selection, type of check)
triggers the creation in the backend of a function to handle the check
for specifically that pair. The resulting functions are deduplicated by
the linker. The pseudo-instruction (really the function) is specified
to preserve all registers except for the registers that the AAPCS
specifies may be clobbered by a call.
To measure the code size and performance effect of this change, I
took a number of measurements using Chromium for Android on aarch64,
comparing a browser with inlined checks (the baseline) against a
browser with outlined checks.
Code size: Size of .text decreases from 243897420 to 171619972 bytes,
or a 30% decrease.
Performance: Using Chromium's blink_perf.layout microbenchmarks I
measured a median performance regression of 6.24%.
The fact that a perf/size tradeoff is evident here suggests that
we might want to make the new behaviour conditional on -Os/-Oz.
But for now I've enabled it unconditionally, my reasoning being that
hwasan users typically expect a relatively large perf hit, and ~6%
isn't really adding much. We may want to revisit this decision in
the future, though.
I also tried experimenting with varying the number of registers
selectable by the hwasan check pseudo-instruction (which would result
in fewer variants being created), on the hypothesis that creating
fewer variants of the function would expose another perf/size tradeoff
by reducing icache pressure from the check functions at the cost of
register pressure. Although I did observe a code size increase with
fewer registers, I did not observe a strong correlation between the
number of registers and the performance of the resulting browser on the
microbenchmarks, so I conclude that we might as well use ~all registers
to get the maximum code size improvement. My results are below:
Regs | .text size | Perf hit
-----+------------+---------
~all | 171619972 | 6.24%
16 | 171765192 | 7.03%
8 | 172917788 | 5.82%
4 | 177054016 | 6.89%
Differential Revision: https://reviews.llvm.org/D56954
llvm-svn: 351920
It might be a bit nicer to use the fancy .legalIf and co. predicates,
but this was requiring more boilerplate and disables the coverage
assertions.
llvm-svn: 351886
For AMDGPU the shift amount is never 64-bit, and
this needs to use a 32-bit shift.
X86 uses i8, but seemed to be hacking around this before.
llvm-svn: 351882
Summary: Enable full support for the debug info.
Reviewers: echristo
Subscribers: jholewinski, aprantl, JDevlieghere, llvm-commits
Differential Revision: https://reviews.llvm.org/D46189
llvm-svn: 351846
Summary: Initial function labels must follow the debug location for the correct relocation info generation.
Reviewers: tra, jlebar, echristo
Subscribers: jholewinski, llvm-commits
Differential Revision: https://reviews.llvm.org/D45784
llvm-svn: 351843
Previously we had names like 'Call' or 'Tail'. This potentially clashes with
the naming scheme used elsewhere in RISCVInstrInfo.td. Many other backends
would use names like AArch64call or PPCtail. I prefer the SystemZ approach,
which uses prefixed all-lowercase names. This matches the naming scheme used
for target-independent SelectionDAG nodes.
llvm-svn: 351823
For constant bit select patterns, replace one AND with a ANDNP, allowing us to reuse the constant mask. Only do this if the mask has multiple uses (to avoid losing load folding) or if we have XOP as its VPCMOV can handle most folding commutations.
This also requires computeKnownBitsForTargetNode support for X86ISD::ANDNP and X86ISD::FOR to prevent regressions in fabs/fcopysign patterns.
Differential Revision: https://reviews.llvm.org/D55935
llvm-svn: 351819
Similar to horizontal ops on D56777, the sse2 (but not mmx) bit shift ops has local forwarding disabled, adding +1cy to the use latency for the result.
Differential Revision: https://reviews.llvm.org/D57026
llvm-svn: 351817
Similar to horizontal ops on D56777, the vpermilpd/vpermilps variable mask ops has local forwarding disabled, adding +1cy to the use latency for the result.
Differential Revision: https://reviews.llvm.org/D57022
llvm-svn: 351815
First step towards PR40376, this patch adds support for getCmpSelInstrCost to use the (optional) Instruction CmpInst predicate to indicate the type of integer comparison we're performing and alter the costs accordingly.
Differential Revision: https://reviews.llvm.org/D57013
llvm-svn: 351810
When we are inserting 1 "inline" element, and zeroing 2 of the other elements then we can safely commute the insertps source inputs to improve memory folding.
Differential Revision: https://reviews.llvm.org/D56843
llvm-svn: 351807
Avoid the infinite loop caused by the target DAG combine converting ANYEXT to
SIGNEXT and the target-independent DAG combine logic converting back to
ANYEXT. Do this by not adding the new node to the worklist.
Committing directly as this definitely doesn't make the problem any worse, and
I intend to follow-up with a patch that avoids this custom combiner logic
altogether and just lowers the i32 operations to a target-specific
SelectionDAG node. This should be easier to reason about and improve codegen
quality in some cases (though may miss out on some later DAG combines).
llvm-svn: 351806
This broke the RISCV build, and even with that fixed, one of the RISCV
tests behaves surprisingly differently with asserts than without,
leaving there no clear test pattern to use. Generally it seems bad for
hte IR to differ substantially due to asserts (as in, an alloca is used
with asserts that isn't needed without!) and nothing I did simply would
fix it so I'm reverting back to green.
This also required reverting the RISCV build fix in r351782.
llvm-svn: 351796
The break isn't strictly needed yet as there is no subsequent entry in the
case. But adding to prevent mistakes further down the road.
llvm-svn: 351785
This patch may seem familiar... but my previous patch handled the
equivalent lsls+and, not this case. Usually instcombine puts the
"and" after the shift, so this case doesn't come up. However, if the
shift comes out of a GEP, it won't get canonicalized by instcombine,
and DAGCombine doesn't have an equivalent transform.
This also modifies isDesirableToCommuteWithShift to suppress DAGCombine
transforms which would make the overall code worse.
I'm not really happy adding a bunch of code to handle this, but it would
probably be tricky to substantially improve the behavior of DAGCombine
here.
Differential Revision: https://reviews.llvm.org/D56032
llvm-svn: 351776
Not sure this is the best fix, but it saves an instruction for certain
constructs involving variable shifts.
Differential Revision: https://reviews.llvm.org/D55572
llvm-svn: 351768
Summary:
Use X86ISD::VFPROUND in the instruction isel patterns. Add new patterns for ISD::FP_ROUND to maintain support for fptrunc in IR.
In the process I found a couple duplicate isel patterns which I also deleted in this patch.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D56991
llvm-svn: 351762
Summary:
For compress, a select node doesn't semantically reflect the behavior of the instruction. The mask would have holes in it, but the resulting write is to contiguous elements at the bottom of the vector.
Furthermore, as far as the compressing and expanding is concerned the behavior is depended on the mask. You can't just have an expand/compress node that only reads the input vector. That node would have no meaning by itself.
This all only works because we pattern match the compress/expand+select back to the instruction. But conceivably an optimization of the select could break the pattern and leave something meaningless.
This patch modifies the expand and compress node to take the mask and passthru as additional inputs and gets rid of the select all together.
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D57002
llvm-svn: 351761
Fixes two problems with GCNHazardRecognizer:
1. It only scans up to 5 instructions emitted earlier.
2. It does not take control flow into account. An earlier instruction
from the previous basic block is not necessarily a predecessor.
At the same time a real predecessor block is not scanned.
The patch provides a way to distinguish between scheduler and
hazard recognizer mode. It is OK to work with emitted instructions
in the scheduler because we do not really know what will be emitted
later and its order. However, when pass works as a hazard recognizer
the schedule is already finalized, and we have full access to the
instructions for the whole function, so we can properly traverse
predecessors and their instructions.
Differential Revision: https://reviews.llvm.org/D56923
llvm-svn: 351759
D56777 added +1cy local forwarding penalty for horizontal operations, but this penalty only affects sse2/xmm variants, the mmx variants don't suffer the penalty.
Confirmed with @andreadb
llvm-svn: 351755
r327630 introduced new write definitions for float/vector loads.
Before that revision, WriteLoad was used by both integer/float (scalar/vector)
load. So, WriteLoad had to conservatively declare a latency to 5cy. That is
because the load-to-use latency for float/vector load is 5cy.
Now that we have dedicated writes for float/vector loads, there is no reason why
we should keep the latency of WriteLoad to 5cy. At the moment, WriteLoad is only
used by scalar integer loads only; we can assume an optimstic 3cy latency for
them.
This patch changes that latency from 5cy to 3cy, and regenerates the affected
scheduling/mca tests.
Differential Revision: https://reviews.llvm.org/D56922
llvm-svn: 351742
This updates the AVR Select8/Select16 expansion code so that, when
inserting the two basic blocks for true and false conditions, any
existing fallthrough on the previous block is preserved.
Prior to this patch, if the block before the Select pseudo fell through
to the subsequent block, two new basic blocks would be inserted at the
prior fallthrough point, changing the fallthrough destination.
The predecessor or successor lists were not updated, causing the
BranchFolding pass at -O1 and above the rearrange basic blocks, causing
an infinite loop. Not to mention the unconditional fallthrough to the
true block is incorrect in of itself.
This patch modifies the Select8/16 expansion so that, if inserting true
and false basic blocks at a fallthrough point, the implicit branch is
preserved by means of an explicit, unconditional branch to the previous
fallthrough destination.
Thanks to Carl Peto for reporting this bug.
This fixes avr-rust bug https://github.com/avr-rust/rust/issues/123.
llvm-svn: 351721
Prior to this, the code was missing AVR-specific relocation logic in
RelocVisitor.h.
This patch teaches RelocVisitor about R_AVR_16 and R_AVR_32.
Debug information is emitted in the final object file, and understood by
'avr-readelf --debug-dump' from AVR-GCC.
llvm-dwarfdump is yet to understand how to dump AVR DWARF symbols.
llvm-svn: 351720
This reverts commit r351718.
Carl pointed out that the unit test could be improved.
This patch will be recommitted once the test is made more resilient.
llvm-svn: 351719
This updates the AVR Select8/Select16 expansion code so that, when
inserting the two basic blocks for true and false conditions, any
existing fallthrough on the previous block is preserved.
Prior to this patch, if the block before the Select pseudo fell through
to the subsequent block, two new basic blocks would be inserted at the
prior fallthrough point, changing the fallthrough destination.
The predecessor or successor lists were not updated, causing the
BranchFolding pass at -O1 and above the rearrange basic blocks, causing
an infinite loop. Not to mention the unconditional fallthrough to the
true block is incorrect in of itself.
This patch modifies the Select8/16 expansion so that, if inserting true
and false basic blocks at a fallthrough point, the implicit branch is
preserved by means of an explicit, unconditional branch to the previous
fallthrough destination.
Thanks to Carl Peto for reporting this bug.
This fixes avr-rust bug https://github.com/avr-rust/rust/issues/123.
llvm-svn: 351718
There is a combine that was hiding these tests
not actually testing what they should be, although
they were producing the expected end result.
llvm-svn: 351698
This causes a couple of changes in the upgrade tests as signed/unsigned eq/ne are equivalent and we constant fold true/false codes, these changes are the same as what we already do for avx512 cmp/ucmp.
Noticed while cleaning up vector integer comparison costs for PR40376.
llvm-svn: 351697
This was crashing in the predicate function assuming the value
is a vector.
Copy more of what AArch64 uses. This probably needs more refinement
later, but I don't exactly understand what it means in some cases,
particularly since any legalization for these seems to be missing.
llvm-svn: 351693
Prior to SSE41 (and sometimes on AVX1), vector select has to be performed as a ((X & C)|(Y & ~C)) bit select.
Exposes a couple of issues with the min/max reduction costs (which only go down to SSE42 for some reason).
The increase pre-SSE41 selection costs also prevent a couple of tests from firing any longer, so I've either tweaked the target or added AVX tests as well to the existing SSE2 tests.
llvm-svn: 351685
These were originally introduced in a copy-paste committed in r351526.
The reference to 't2_so_imm' have been updated to 'imm_com8' so the
comment is now accurate.
Thanks to Eli Friedman for noticing this.
llvm-svn: 351674
Prior to this patch, the AVR::LDWRdPtr instruction was always lowered to
instructions of this pattern:
ld $GPR8, [PTR:XYZ]+
ld $GPR8, [PTR]+1
This has a problem; the [PTR] is incremented in-place once, but never
decremented.
Future uses of the same pointer will use the now clobbered value,
leading to the pointer being incorrect by an offset of one.
This patch modifies the expansion code of the LDWRdPtr pseudo
instruction so that the pointer variable is not silently clobbered in
future uses in the same live range.
Bug first reported by Keshav Kini.
Patch by Kaushik Phatak.
llvm-svn: 351673
This reverts commit r351544.
In that commit, I had mistakenly misattributed the issue submitter as
the patch author, Kaushik Phatak.
The patch will be recommitted immediately with the correct attribution.
llvm-svn: 351672
to reflect the new license. These used slightly different spellings that
defeated my regular expressions.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351648
to reflect the new license.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351636
Summary:
Right now we include ${TGT}GenCallingConv.inc once per each instruction
selection method implemented by ${TGT}:
- ${TGT}ISelLowering.cpp
- ${TGT}CallLowering.cpp
- ${TGT}FastISel.cpp
Instead, add a mechanism to tablegen for marking a particular convention
as "External", which causes tablegen to emit into the ::llvm namespace,
instead of as a static helper. This allows us to provide a header to
forward declare it, so we can simply call the function from all the
places it is referenced. Typically the calling convention analyzer is
called indirectly, so it doesn't benefit from inlining.
This saves a bit of final binary size, but mostly just saves object file
size:
before after diff artifact
12852K 12492K -360K X86ISelLowering.cpp.obj
4640K 4280K -360K X86FastISel.cpp.obj
1704K 2092K +388K X86CallingConv.cpp.obj
52448K 52336K -112K llc.exe
I didn't collect before numbers for X86CallLowering.cpp.obj, which is
for GlobalISel, but we should save 360K there as well.
This patch applies the strategy to the X86 backend, but there is no
reason it couldn't be applied to the other backends that implement
multiple ISel strategies, like AArch64.
Reviewers: craig.topper, hfinkel, efriedma
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D56883
llvm-svn: 351616
This sends these intrinsics through isel in a much more normal way. This should allow addressing mode matching in isel to make better use of the displacement field.
llvm-svn: 351583
This sends these intrinsics through isel in a much more normal way. This should allow addressing mode matching in isel to make better use of the displacement field.
Differential Revision: https://reviews.llvm.org/D56827
llvm-svn: 351570
This commit adds some missing intrinsics into the isAlwaysUniform list
for the AMDGPU backend.
Differential Revision: https://reviews.llvm.org/D56845
llvm-svn: 351562
Prior to this patch, the AVR::LDWRdPtr instruction was always lowered to
instructions of this pattern:
ld $GPR8, [PTR:XYZ]+
ld $GPR8, [PTR]+1
This has a problem; the [PTR] is incremented in-place once, but never
decremented.
Future uses of the same pointer will use the now clobbered value,
leading to the pointer being incorrect by an offset of one.
This patch modifies the expansion code of the LDWRdPtr pseudo
instruction so that the pointer variable is not silently clobbered in
future uses in the same live range.
Patch by Keshav Kini.
llvm-svn: 351544
The CBR instruction is just an ANDI instruction with the immediate
complemented.
Because of this, prior to this change TableGen would warn due to a
decoding conflict.
This commit fixes the existing compilation warning:
===============
[423/492] Building AVRGenDisassemblerTables.inc...
Decoding Conflict:
0111............
01..............
................
ANDIRdK 0111____________
CBRRdK 0111____________
================
After this commit, there are no more decoding conflicts in the AVR
backend's instruction definitions.
Thanks to Eli F for pointing me torward `t2_so_imm_not` as an example of
how to perform a complement in an instruction alias.
Fixes BugZilla PR38802.
llvm-svn: 351526
This change modifies the LLVM ISel lowering settings so that
8-bit/16-bit multiplication is expanded to calls into the compiler
runtime library if the MCU being targeted does not support
multiplication in hardware.
Before this, MUL instructions would be generated on CPUs like the
ATtiny85, triggering a CPU reset due to an illegal instruction at
runtime.
First raised in https://github.com/avr-rust/rust/issues/124.
llvm-svn: 351523
The callee address is added as an optional operand (MCSymbol) in
AdjustInstrPostInstrSelection() and then used by asm printer to insert:
'.reloc tmplabel, R_MIPS_JALR, symbol
tmplabel:'.
Controlled with '-mips-jalr-reloc', default is true.
Differential revision: https://reviews.llvm.org/D56694
llvm-svn: 351485
EXPENSIVE_CHECKS buildbots are failing due to r351404.
Add x1 as live in to the funclet basic block for SEH funclets, as well as
-verify-machineinstrs to the test case that triggered the failure.
llvm-svn: 351472
Summary:
objdump was interpreting the function header containing the locals
declaration as instructions. To parse these without injecting target
specific code in objdump, MCDisassembler::onSymbolStart was added to
be implemented by the WebAssembly implemention.
WasmObjectFile now returns a code offset for the "address" of a symbol,
rather than the index. This is also more in-line with what other
targets do.
Also ensured that the AsmParser correctly puts each function
in its own segment to enable this test case.
Reviewers: sbc100, dschuff
Subscribers: jgravelle-google, aheejin, sunfish, rupprecht, llvm-commits
Differential Revision: https://reviews.llvm.org/D56684
llvm-svn: 351460
Allow varargs functions to be called, both in arm and thumb mode. This
boils down to choosing the correct calling convention, which we can
easily test by making sure arm_aapcscc is used instead of
arm_aapcs_vfpcc when the callee is variadic.
llvm-svn: 351424
In order to support codegen RV64A, this patch:
* Introduces masked atomics intrinsics for atomicrmw operations and cmpxchg
that use the i64 type. These are ultimately lowered to masked operations
using lr.w/sc.w, but we need to use these alternate intrinsics for RV64
because i32 is not legal
* Modifies RISCVExpandPseudoInsts.cpp to handle PseudoAtomicLoadNand64 and
PseudoCmpXchg64
* Modifies the AtomicExpandPass hooks in RISCVTargetLowering to sext/trunc as
needed for RV64 and to select the i64 intrinsic IDs when necessary
* Adds appropriate patterns to RISCVInstrInfoA.td
* Updates test/CodeGen/RISCV/atomic-*.ll to show RV64A support
This ends up being a fairly mechanical change, as the logic for RV32A is
effectively reused.
Differential Revision: https://reviews.llvm.org/D53233
llvm-svn: 351422
Summary:
Everything before the word "version" is the tool, and everything after
the word "version" is the version.
Reviewers: aheejin, dschuff
Subscribers: sbc100, jgravelle-google, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D56742
llvm-svn: 351399
Previously we used ISD::SHL and ISD::SRL to represent these in SelectionDAG. ISD::SHL/SRL interpret an out of range shift amount as undefined behavior and will constant fold to undef. While the intrinsics are defined to return 0 for out of range shift amounts. A previous patch added a special node for VPSRAV to produce all sign bits.
This was previously believed safe because undefs frequently get turned into 0 either from the constant pool or a desire to not have a false register dependency. But undef is treated specially in some optimizations. For example, its ignored in detection of vector splats. So if the ISD::SHL/SRL can be constant folded and all of the elements with in bounds shift amounts are the same, we might fold it to single element broadcast from the constant pool. This would not put 0s in the elements with out of bounds shift amounts.
We do have an existing InstCombine optimization to use shl/lshr when the shift amounts are all constant and in bounds. That should prevent some loss of constant folding from this change.
Patch by zhutianyang and Craig Topper
Differential Revision: https://reviews.llvm.org/D56695
llvm-svn: 351381
This cleans up the duplication we have with both intrinsic isel patterns and vselect isel patterns. This should also allow the intrinsics to get SimplifyDemandedBits support for the condition.
I've switched the canonical pattern in isel to use the X86ISD::BLENDV node instead of VSELECT. Since it always seemed weird to move from BLENDV with its relaxed rules on condition bits to VSELECT which has strict rules about all bits of the condition element being the same. Its more correct to go from VSELECT to BLENDV.
Differential Revision: https://reviews.llvm.org/D56771
llvm-svn: 351380
Summary:
For these loads that write to the HI part of a register, we should chain them to the op that writes to the LO part
of the register to maintain the appropriate order.
Reviewers:
rampitec, arsenm
Differential Revision:
https://reviews.llvm.org/D56454
llvm-svn: 351379
If we're going to generate a new inverted setcc, we should make sure we will be able to remove the old setcc.
Differential Revision: https://reviews.llvm.org/D56765
llvm-svn: 351378
Summary:
This patch supports MS SEH extensions __try/__except/__finally. The intrinsics localescape and localrecover are responsible for communicating escaped static allocas from the try block to the handler.
We need to preserve frame pointers for SEH. So we create a new function/property HasLocalEscape.
Reviewers: rnk, compnerd, mstorsjo, TomTan, efriedma, ssijaric
Reviewed By: rnk, efriedma
Subscribers: smeenai, jrmuizel, alex, majnemer, ssijaric, ehsan, dmajor, kristina, javed.absar, kristof.beyls, chrib, llvm-commits
Differential Revision: https://reviews.llvm.org/D53540
llvm-svn: 351370
On Jaguar, horizontal adds/subs have local forwarding disable.
That means, we pay a compulsory extra cycle of write-back stage, and the value
is not available until the end of that stage.
This patch changes the latency of horizontal operations by adding an extra
cycle. With this patch, latency numbers now match what is reported by perf.
I plan to send another patch to also 'fix' the latency of shuffle operations (on
Jaguar, local forwarding is disabled for vector shuffles too).
Differential Revision: https://reviews.llvm.org/D56777
llvm-svn: 351366
Remove the existing assertion and just return false for unexpected shuffle value types (<X x i1> mainly....).
Found while updating combineX86ShufflesRecursively to run within SimplifyDemandedVectorElts/SimplifyDemandedBits.
llvm-svn: 351365
combineX86ShufflesRecursively is pretty cumbersome with a lot of arguments that only matter later in recursion.
This commit adds a wrapper version that only takes the initial root Op to simplify calls that don't need to worry about these.
An early, cleanup step towards merging combineX86ShufflesRecursively into SimplifyDemandedVectorElts/SimplifyDemandedBits.
llvm-svn: 351352
I was trying to prevent shuffle regressions while matching more horizontal ops
and ended up here:
shuf (extract X, 0), (extract X, 4), Mask --> extract (shuf X, undef, Mask'), 0
The affected tests were added for:
https://bugs.llvm.org/show_bug.cgi?id=34380
This patch won't change the examples in the bug report itself, but we should be
able to extend this to catch more types.
Differential Revision: https://reviews.llvm.org/D56756
llvm-svn: 351346
This is LLVM part of D56663
Linker scripts shipped by TI require to have every
interrupt vector in a separate section with a specific name:
SECTIONS
{
__interrupt_vector_XX : { KEEP (*(__interrupt_vector_XX )) } > VECTXX
...
}
Follow the requirement emit the section for every vector
which contain address of interrupt handler:
.section __interrupt_vector_XX,"ax",@progbits
.word %isr%
Patch by Kristina Bessonova!
Differential Revision: https://reviews.llvm.org/D56664
llvm-svn: 351345
https://reviews.llvm.org/D52803
This patch adds support to continuously CSE instructions during
each of the GISel passes. It consists of a GISelCSEInfo analysis pass
that can be used by the CSEMIRBuilder.
llvm-svn: 351283
Summary:
Make recoverfp intrinsic target-independent so that it can be implemented for AArch64, etc.
Refer D53541 for the context. Clang counterpart D56748.
Reviewers: rnk, efriedma
Reviewed By: rnk, efriedma
Subscribers: javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D56747
llvm-svn: 351281
That's really what it is. If we didn't use intrinsics for BLENDVPS/BLENDVPD/PBLENDVB all the way to isel, this is the node we would use.
llvm-svn: 351278
We're trying to have the vXi1 types in IR as much as possible. This prevents the need for bitcasts when the producer of the mask was already a vXi1 value like an icmp. The bitcasts can be subject to code motion and interfere with basic block at a time isel in bad ways.
llvm-svn: 351275
Summary:
We have seen performance regression when v_add3 is generated. The major reason is that the v_mad pattern
is broken when v_add3 is generated. We also see the register pressure increased. While we could not properly
estimate register pressure during instruction selection, we can give mad a higher priority.
In this work, we raise the priority for mad24 in selection and resolve the performance regression.
Reviewers:
rampitec
Differential Revision:
https://reviews.llvm.org/D56745
llvm-svn: 351273
In keeping with our general direction of having the vXi1 type present in IR, this patch converts the mask argument for avx512 gather to vXi1. This can avoid k-register to GPR to k-register transitions late in codegen.
I left the existing intrinsics behind because they have many out of tree users such as ISPC. They generate their own code and don't go through the autoupgrade path which only works for bitcode and ll parsing. Ideally we will get them to migrate to target independent intrinsics, but it might be easier for them to migrate to these new intrinsics.
I'll work on scatter and gatherpf/scatterpf next.
Differential Revision: https://reviews.llvm.org/D56527
llvm-svn: 351234
Summary:
As described in PR40209, there can be issues in DBG_VALUEs handling when multiple defs present in a BB. This patch
adds logic for detection of related to def DBG_VALUEs and localizes register update and movement to found DBG_VALUEs.
Reviewers: aheejin
Subscribers: mgorny, dschuff, sbc100, jgravelle-google, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D56401
llvm-svn: 351216
Modify getRegForInlineAsmConstraint to return special singleton
register class when a constraint references ST(7) not RFP80 for which
ST(7) is not a member.
llvm-svn: 351206
If we're shuffling with a zero vector, then we are better off not doing VECTOR_SHUFFLE(UNPCK()) as we lose track of those zero elements.
We were already doing this for SSSE3 targets as we have PSHUFB, but its worth doing for all targets.
llvm-svn: 351203
Summary:
V8 currently implements SIMD shifts as taking an immediate operation,
which disagrees with the spec proposal and the toolchain
implementation. As a stopgap measure to get things working, unroll all
vector shifts. Since this is a temporary measure, there are no tests.
Reviewers: aheejin, dschuff
Subscribers: sbc100, jgravelle-google, sunfish, dmgreen, llvm-commits
Differential Revision: https://reviews.llvm.org/D56520
llvm-svn: 351151
Summary:
This allows moving the condition from the intrinsic to the standard ICmp
opcode, so that LLVM can do simplifications on it. The icmp.i1 intrinsic
is an identity for retrieving the SGPR mask.
And we can also get the mask from and i1, or i1, xor i1.
Reviewers: arsenm, nhaehnle
Subscribers: kzhuravl, jvesely, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D52060
llvm-svn: 351150
Summary:
In r345197 ESP and RSP were added to GR32_TC/GR64_TC, allowing them to
be used for tail calls, but this also caused `findDeadCallerSavedReg` to
think they were acceptable targets for clobbering. Filter them out.
Fixes PR40289.
Patch by Geoffry Song!
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D56617
llvm-svn: 351146
Otherwise, with D56544, the intrinsic will be expanded to an integer
csel, which is probably not what the user expected. This matches the
general convention of using "v1" types to represent scalar integer
operations in vector registers.
While I'm here, also add some error checking so we don't generate
illegal ABS nodes.
Differential Revision: https://reviews.llvm.org/D56616
llvm-svn: 351141
This feature enables the fusion of some arithmetic and logic instructions
together.
Differential revision: https://reviews.llvm.org/D56572
llvm-svn: 351139
If we have PSHUFB and we're shuffling with a zero vector, then we are better off not doing VECTOR_SHUFFLE(UNPCK()) as we lose track of those zero elements.
llvm-svn: 351103
add (extractelt (X, 0), extractelt (X, 1)) --> extractelt (hadd X, X), 0
This is the integer sibling to D56011.
There's an additional restriction to only to do this transform in the
case where we don't have extra extracts from the source vector. Without
that, we can fail to match larger horizontal patterns that are more
beneficial than this minimal case. An improvement to the more general
h-op lowering may allow us to remove the restriction here in a follow-up.
llvm-svn: 351093
This removes the old grow_memory and mem.grow-style intrinsics, leaving just
the memory.grow-style intrinsics.
Differential Revision: https://reviews.llvm.org/D56645
llvm-svn: 351084
With this patch, shifts are lowered to optimal number of instructions
necessary to shift types larger than the general purpose register size.
This resolves PR/32293.
Thanks to Kyle Butt for reporting the issue!
Differential Revision: https://reviews.llvm.org/D56320
llvm-svn: 351059
Make it possible for TableGen to produce code for selecting MOVi32imm.
This allows reasonably recent ARM targets to select a lot more constants
than before.
We achieve this by adding GISelPredicateCode to arm_i32imm. It's
impossible to use the exact same code for both DAGISel and GlobalISel,
since one uses "Subtarget->" and the other "STI." to refer to the
subtarget. Moreover, in GlobalISel we don't have ready access to the
MachineFunction, so we need to add a bit of code for obtaining it from
the instruction that we're selecting. This is also the reason why it
needs to remain a PatLeaf instead of the more specific IntImmLeaf.
llvm-svn: 351056
TFE and LWE support requires extra result registers that are written in the
event of a failure in order to detect that failure case.
The specific use-case that initiated these changes is sparse texture support.
This means that if image intrinsics are used with either option turned on, the
programmer must ensure that the return type can contain all of the expected
results. This can result in redundant registers since the vector size must be a
power-of-2.
This change takes roughly 6 parts:
1. Modify the instruction defs in tablegen to add new instruction variants that
can accomodate the extra return values.
2. Updates to lowerImage in SIISelLowering.cpp to accomodate setting TFE or LWE
(where the bulk of the work for these instruction types is now done)
3. Extra verification code to catch cases where intrinsics have been used but
insufficient return registers are used.
4. Modification to the adjustWritemask optimisation to account for TFE/LWE being
enabled (requires extra registers to be maintained for error return value).
5. An extra pass to zero initialize the error value return - this is because if
the error does not occur, the register is not written and thus must be zeroed
before use. Also added a new (on by default) option to ensure ALL return values
are zero-initialized that is required for sparse texture support.
6. Disable the inst_combine optimization in the presence of tfe/lwe (later TODO
for this to re-enable and handle correctly).
There's an additional fix now to avoid a dmask=0
For an image intrinsic with tfe where all result channels except tfe
were unused, I was getting an image instruction with dmask=0 and only a
single vgpr result for tfe. That is incorrect because the hardware
assumes there is at least one vgpr result, plus the one for tfe.
Fixed by forcing dmask to 1, which gives the desired two vgpr result
with tfe in the second one.
The TFE or LWE result is returned from the intrinsics using an aggregate
type. Look in the test code provided to see how this works, but in essence IR
code to invoke the intrinsic looks as follows:
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.1d.v4f32i32.i32(i32 15,
i32 %s, <8 x i32> %rsrc, i32 1, i32 0)
%v.vec = extractvalue {<4 x float>, i32} %v, 0
%v.err = extractvalue {<4 x float>, i32} %v, 1
This re-submit of the change also includes a slight modification in
SIISelLowering.cpp to work-around a compiler bug for the powerpc_le
platform that caused a buildbot failure on a previous submission.
Differential revision: https://reviews.llvm.org/D48826
Change-Id: If222bc03642e76cf98059a6bef5d5bffeda38dda
Work around for ppcle compiler bug
Change-Id: Ie284cf24b2271215be1b9dc95b485fd15000e32b
llvm-svn: 351054
Part of the effort to refactoring frame pointer code generation. We used
to use two function attributes "no-frame-pointer-elim" and
"no-frame-pointer-elim-non-leaf" to represent three kinds of frame
pointer usage: (all) frames use frame pointer, (non-leaf) frames use
frame pointer, (none) frame use frame pointer. This CL makes the idea
explicit by using only one enum function attribute "frame-pointer"
Option "-frame-pointer=" replaces "-disable-fp-elim" for tools such as
llc.
"no-frame-pointer-elim" and "no-frame-pointer-elim-non-leaf" are still
supported for easy migration to "frame-pointer".
tests are mostly updated with
// replace command line args ‘-disable-fp-elim=false’ with ‘-frame-pointer=none’
grep -iIrnl '\-disable-fp-elim=false' * | xargs sed -i '' -e "s/-disable-fp-elim=false/-frame-pointer=none/g"
// replace command line args ‘-disable-fp-elim’ with ‘-frame-pointer=all’
grep -iIrnl '\-disable-fp-elim' * | xargs sed -i '' -e "s/-disable-fp-elim/-frame-pointer=all/g"
Patch by Yuanfang Chen (tabloid.adroit)!
Differential Revision: https://reviews.llvm.org/D56351
llvm-svn: 351049
Introduce GlobalISel pre legalizer pass for MIPS.
It will be used to cope with instructions that require
combining before legalization.
Differential Revision: https://reviews.llvm.org/D56269
llvm-svn: 351046
We can't represent this properly with vselect like we normally do. We also have to update the instruction definition to use a VK2WM mask instead of VK4WM to represent this.
Fixes another case from PR34877
llvm-svn: 351018
We can't represent this properly with vselect like we normally do. We also have to update the instruction definition to use a VK2WM mask instead of VK4WM to represent this.
Fixes another case from PR34877.
llvm-svn: 351017
This patch takes some of the code from D49837 to allow us to enable ISD::ABS support for all SSE vector types.
Differential Revision: https://reviews.llvm.org/D56544
llvm-svn: 350998
The 128-bit input produces 64-bits of output and fills the upper 64-bits with 0. The mask only applies to the lower elements. But we can't represent this with a vselect like we normally do.
This also avoids the need to have a special X86ISD::SELECT when avx512bw isn't enabled since vselect v8i16 isn't legal there.
Fixes another instruction for PR34877.
llvm-svn: 350994
As discussed on llvm-dev
<http://lists.llvm.org/pipermail/llvm-dev/2018-December/128497.html>, we have
to be careful when trying to select the *w RV64M instructions. i32 is not a
legal type for RV64 in the RISC-V backend, so operations have been promoted by
the time they reach instruction selection. Information about whether the
operation was originally a 32-bit operations has been lost, and it's easy to
write incorrect patterns.
Similarly to the variable 32-bit shifts, a DAG combine on ANY_EXTEND will
produce a SIGN_EXTEND if this is likely to result in sdiv/udiv/urem being
selected (and so save instructions to sext/zext the input operands).
Differential Revision: https://reviews.llvm.org/D53230
llvm-svn: 350993
This restores support for selecting the SLLW/SRLW/SRAW instructions, which was
removed in rL348067 as the previous patterns made some unsafe assumptions.
Also see the related llvm-dev discussion
<http://lists.llvm.org/pipermail/llvm-dev/2018-December/128497.html>
Ultimately I didn't introduce a custom SelectionDAG node, but instead added a
DAG combine that inserts an AssertZext i5 on the shift amount for an i32
variable-length shift and also added an ANY_EXTEND DAG-combine which will
instead produce a SIGN_EXTEND for an i32 variable-length shift, increasing the
opportunity to safely select SLLW/SRLW/SRAW.
There are obviously different ways of addressing this (a number discussed in
the llvm-dev thread), so I'd welcome further feedback and comments.
Note that there are now some cases in
test/CodeGen/RISCV/rv64i-exhaustive-w-insts.ll where sraw/srlw/sllw is
selected even though sra/srl/sll could be used without any extra instructions.
Given both are semantically equivalent, there doesn't seem a good reason to
prefer one vs the other. Given that would require more logic to still select
sra/srl/sll in those cases, I've left it preferring the *w variants.
Differential Revision: https://reviews.llvm.org/D56264
llvm-svn: 350992
We no longer need to extend mask scalars before bitcasting them to vXi1. This was only needed for the truncate intrinsics. And was really a bug in our lowering of them.
llvm-svn: 350991
We still use i8 for the load/store type. So we need to convert to/from i16 to around the mask type.
By doing this we get an i8->i16 extload which we can then pattern match to a KMOVW if the access is aligned.
llvm-svn: 350989
We can't properly represent this with a vselect since the upper elements of the result are supposed to be zeroed regardless of the mask.
This also reuses the new nodes even when the result type fits in 128 bits if the input is q/d and the result is w/b since vselect w/b using k-register condition isn't legal without avx512bw. Currently we're doing this even when avx512bw is enabled, but I might change that.
This fixes some of PR34877
llvm-svn: 350985
Teach x86 assembly operand parsing to distinguish between assembler
variable assigned to named registers and those assigned to immediate
values.
Reviewers: rnk, nickdesaulniers, void
Subscribers: hiraditya, jyknight, llvm-commits
Differential Revision: https://reviews.llvm.org/D56287
llvm-svn: 350966
Previously, we limited this transform to cases where the
extraction into the build vector happens from vectors of
the same type as the build vector, but that's not required.
There's a slight potential regression seen in the AVX512
result for phadd -- we're using the 256-bit flavor of the
instruction now even though the 128-bit subset is sufficient.
The same problem could already be seen in the AVX2 result.
Follow-up patches will attempt to narrow that back down.
llvm-svn: 350928
We were lowering the last step extract_vector_elt to a bitcast+truncate. Change it to use an extract_vector_elt of index 0 instead. Add isel patterns to do the equivalent of what the bitcast would have done. Plus an isel pattern for an any_extend+extract to prevent some regressions.
Finally add a DAG combine to turn v1i1 scalar_to_vector+extract_vector_elt of 0 into an extract_subvector.
This fixes some of the regressions from D350800.
llvm-svn: 350918
Summary:
We now use __stack_pointer global and global.get/global.set instruction.
This fixes the checking routine for stack_pointer writes accordingly.
This also fixes the existing __stack_pointer test in reg-stackify.ll:
That test used to pass not because of __stack_pointer clashes but
because the function `stackpointer_callee` was not marked as `readnone`,
so it was assumed to possibly write to memory arbitraily, and
`global.set` instruction was marked as `mayStore` in the .td definition,
so they were identified as intervening writes. After we added `readnone`
to its attribute, this test fails without this patch.
Reviewers: dschuff, sunfish
Subscribers: jgravelle-google, sbc100, llvm-commits
Differential Revision: https://reviews.llvm.org/D56094
llvm-svn: 350906
* Teach AsmParser to recognize @rn in distination operand as 0(rn).
* Do not allow Disassembler decoding instructions that have size more
than a number of input bytes.
* Fix UB in MSP430MCCodeEmitter.
Patch by Kristina Bessonova!
Differential Revision: https://reviews.llvm.org/D56547
llvm-svn: 350903
Summary:
This is a third attempt, but this time we have vetted it on Windows
first. The previous errors were due to an uninitialized class member.
Reviewers: aheejin
Subscribers: dschuff, sbc100, jgravelle-google, sunfish, jfb, llvm-commits
Differential Revision: https://reviews.llvm.org/D56560
llvm-svn: 350901
This extends to combineVSelectToShrunkBlend to be able to resimplify SHRUNKBLENDS that have already been created.
This should help some of the regressions from D56387
Differential Revision: https://reviews.llvm.org/D56421
llvm-svn: 350875
Despite what the comment says, FCMP_UNE would be an OR not an AND. In the lowering code the first branch created still goes to the original destination. The second branch was exchanged to go to where the subsequent unconditional branch went. This is different than what we do for FCMP_OEQ where both branches that we create go to the original unconditional branch.
As far as I can tell, I think this means we don't need to exchange the branch target with the unconditional branch for FCMP_UNE at all.
Differential Revision: https://reviews.llvm.org/D56309
llvm-svn: 350873
This commit fixes the dwordx3/southern-islands failures that were found
in bugzilla https://bugs.llvm.org/show_bug.cgi?id=40129, by not
generating the dwordx3 variants of load/store instructions that were
added to the ISA after southern islands.
Differential Revision: https://reviews.llvm.org/D56434
llvm-svn: 350838
That is, remove many of the calls to Type::getNumContainedTypes(),
Type::subtypes(), and Type::getContainedType(N).
I'm not intending to remove these accessors -- they are
useful/necessary in some cases. However, removing the pointee type
from pointers would potentially break some uses, and reducing the
number of calls makes it easier to audit.
llvm-svn: 350835
This further improves compatibility with GNU as, allowing input such as the
following to be assembled:
.equ CONST, 0x123456
li a0, CONST
addi a0, a0, %lo(CONST)
.equ CONST, 1
slli a0, a0, CONST
Note that we don't have perfect compatibility with gas, as it will avoid
emitting a relocation in this case:
addi a0, a0, %lo(CONST2)
.equ CONST2, 0x123456
Thanks to Shiva Chen for suggesting a better way to approach this during review.
Differential Revision: https://reviews.llvm.org/D52298
llvm-svn: 350831
When we use the partial-matching function on a 128-bit chunk, we must
account for the possibility that we've matched undef halves of the
original source vectors, so the outputs may need to be reset.
This should allow closing PR40243:
https://bugs.llvm.org/show_bug.cgi?id=40243
llvm-svn: 350830
This is a partial fix for:
https://bugs.llvm.org/show_bug.cgi?id=40243
...as seen in the integer test, we still need to correct the result when using the
existing (old) horizontal op matching function because it does not model the way
x86 256-bit horizontal ops return results (each 128-bit half is its own horizontal-op).
A potential follow-up change for that is discussed in the bug report - see also D56490.
This generally duplicates a lot of the existing matching code, but we can't just remove
that without introducing regressions, so the existing code is renamed and used less often.
Follow-ups may try to reduce that overlap.
Differential Revision: https://reviews.llvm.org/D56450
llvm-svn: 350826
Summary:
This patch changes the legalization action for some half-precision floating-
point vector intrinsics (FSIN, FLOG, etc.) from Promote to Expand. These ops
are not supported in hardware for half-precision vectors, but promotion is
not always possible (for v8f16 operands). Changing the action to Expand fixes
an assertion failure in the legalizer when the frontend produces such ops.
In addition, a quick microbenchmark shows that, in the v4f16 case,
expanding introduces fewer spills and is therefore slightly faster than
promoting.
Reviewers: t.p.northover, SjoerdMeijer
Reviewed By: SjoerdMeijer
Subscribers: javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D56296
llvm-svn: 350825
Add t2TEQrr to the map of instructions with can be reduced down into
a T1 instruction. This is a special case because TEQ just sets the
CPSR and doesn't write to a GPR, which is not the case for EOR. So,
we need to ensure that the EOR can write to the first operand.
Differential Revision: https://reviews.llvm.org/D56255
llvm-svn: 350801
Summary:
This pass replaces GR8/GR16/GR32/GR64 with their equivalent sized mask register classes. But VK32/VK64 aren't legal without AVX512BW. Apparently this mostly appears to work if the register coalescer is able to remove the VK32/VK64 register class reference. Or if we don't ever spill it. But there's no guarantee of that.
Another Intel employee managed to trigger a crash due to this with ISPC. Unfortunately, I've lost the test case he sent me at the time. I'm trying to get him to reproduce it for me. I'd like to get this in before 8.0 branches since its a little scary.
The regressions here are unfortunate, but I think we can make some improvements to DAG combine, load folding, etc. to fix them. Just not sure if we can get that done for 8.0.
Fixes PR39741
Reviewers: RKSimon, spatel
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D56460
llvm-svn: 350800
Summary:
D55896 and D56029 add support to emit fixups for :abs_g0: , :abs_g1_s: , etc.
This patch adds the necessary enums and MCExpr needed for lowering these.
Reviewers: rnk, mstorsjo, efriedma
Reviewed By: efriedma
Subscribers: javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D56037
llvm-svn: 350798
This is a second attempt at r350778, which was reverted in
r350789. The only change is that the unimplemented-simd128 feature has
been renamed simd128-unimplemented, since naming it
unimplemented-simd128 somehow made the simd128 feature flag enable the
unimplemented-simd128 feature on Windows.
llvm-svn: 350791
Found while trying to figure out why my second version of D56421 worked better than the first version. We weren't deleting the vselect in a timely fashion and that caused SimplfyDemandedBit to see an additional user.
The new version doesn't have this problem so this fix isn't needed there, but seemed like the right thing to do.
llvm-svn: 350781
Summary:
This replaces the old ad-hoc -wasm-enable-unimplemented-simd
flag. Also makes the new unimplemented-simd128 feature imply the
simd128 feature.
Reviewers: aheejin, dschuff
Subscribers: sbc100, jgravelle-google, sunfish, llvm-commits, alexcrichton
Differential Revision: https://reviews.llvm.org/D56501
llvm-svn: 350778
Summary:
Looks like many passes print its pass description as a debug message at
the start of each pass, so added that to (mostly newly added) other
passes as well.
Reviewers: dschuff
Subscribers: jgravelle-google, sbc100, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D56142
llvm-svn: 350771
Summary:
For some reason the backend assumed that the condition mask would be
the first argument to the LLVM intrinsic, but everywhere else the
condition mask is the third argument.
Reviewers: aheejin
Subscribers: dschuff, sbc100, jgravelle-google, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D56412
llvm-svn: 350746
This is an initial implementation for Speculative Load Hardening for
AArch64. It builds on top of the recently introduced
AArch64SpeculationHardening pass.
This doesn't implement (yet) some of the optimizations implemented for
the X86SpeculativeLoadHardening pass. I thought introducing the
optimizations incrementally in follow-up patches should make this easier
to review.
Differential Revision: https://reviews.llvm.org/D55929
llvm-svn: 350729
Fixed issue with identity values and other cases, f32/f16 identity values to be added later. fma/mac instructions is disabled for now.
Test is fully reworked, added comments. Other fixes:
1. dpp move with uses and old reg initializer should be in the same BB.
2. bound_ctrl:0 is only considered when bank_mask and row_mask are fully enabled (0xF). Othervise the old register value is checked for identity.
3. Added add, subrev, and, or instructions to the old folding function.
4. Kill flag is cleared for the src0 (DPP register) as it may be copied into more than one user.
Differential revision: https://reviews.llvm.org/D55444
llvm-svn: 350721
Follow up patch of rL350385, for adding predres
command line option. This patch renames the
feature as to keep it aligned with the option
passed by/to clang
Differential Revision: https://reviews.llvm.org/D56484
llvm-svn: 350702
When the result type is v2i64/v2f64 and the index element size is i32, the index vector has two unused elements making the type v4i32. The mask VT should match the number of memory accesses that will be made.
This is consistent with the isel patterns used for the target independent gather/scatter intrinsic.
llvm-svn: 350687
Bad machine code: Illegal virtual register for instruction
function: TestULE
basic block: %bb.0 entry (0x1000a39b158)
instruction: %2:crrc = FCMPUD %1:vsfrc, %3:f8rc
operand 1: %1:vsfrc
Fix assert about missing match between fcmp instruction and register class.
We should use vsx related cmp instruction xvcmpudp instead of fcmpu when vsx is opened.
add -verifymachineinstrs option into related test cases to enable the verify pass.
Differential Revision: https://reviews.llvm.org/D55686
llvm-svn: 350685
This removes check for single use from general ShrinkDemandedConstant
to the BE because of the AArch64 regression after D56289/rL350475.
After several hours of experiments I did not come up with a testcase
failing on any other targets if check is not performed.
Moreover, direct call to ShrinkDemandedConstant is not really needed
and superceed by SimplifyDemandedBits.
Differential Revision: https://reviews.llvm.org/D56406
llvm-svn: 350684
Summary:
StoreResults pass does not optimize store instructions anymore because
store instructions don't return results values anymore. Now this pass is
used solely for memory intrinsics, so update the pass name accordingly
and fix outdated pass descriptions as well.
This patch does not change any meaningful behavior, but not marked as
NFC because it changes a comment check line in a test case.
Reviewers: dschuff
Subscribers: mgorny, sbc100, jgravelle-google, sunfiish, llvm-commits
Differential Revision: https://reviews.llvm.org/D56093
llvm-svn: 350669
Commit f1db33c5c1a9 ("[BPF] Disable relocation for .BTF.ext section")
assigned relocation type R_BPF_NONE if the fixup type
is FK_Data_4 and the symbol is temporary.
The reason is we use FK_Data_4 as a fixup type
for insn offsets in .BTF.ext section.
Just checking whether the symbol is temporary is not enough.
For example, .debug_info may reference some strings whose
fixup is FK_Data_4 with a temporary symbol as well.
To truely reflect the case for .BTF.ext section,
this patch further checks that the section associateed with the symbol
must be SHF_ALLOC and SHF_EXECINSTR, i.e., in the text section.
This fixed the above-mentioned problem.
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 350637
For stack frames on the size of a register in x86, a code size optimization
emits "push rax/eax" instead of "sub" for stack allocation. For example:
foo:
.cfi_startproc
BB#0:
pushq %rax
Ltmp0:
.cfi_def_cfa_offset 16
...
.cfi_endproc
However, we are falling back to DWARF in this case because we cannot
encode %rax as a saved register.
This requirement is wrong, since we don't care about the contents of
%rax, it is the equivalent of a sub.
In order to specify that we care about the contents of %rax, we would
need a .cfi_offset %rax, <offset>.
It's also overzealous in the case where there are pushes for callee saved
registers followed by a "push rax/eax" instead of "sub", in which case we should
also be able to encode the callee saved regs and everything else using compact
unwind.
Patch authored by Bruno Cardoso Lopes.
Differential Revision: https://reviews.llvm.org/D13793
llvm-svn: 350623
We have code to split vector splats (of zero and non-zero) for performance
reasons, but it ignores the fact that a store might be truncating.
Actually, truncating stores are formed for vNi8 and vNi16 types. Since the
truncation is from a legal type, the size of the store is always <= 64-bits and
so they don't actually benefit from being split up anyway, so this patch just
disables that transformation.
llvm-svn: 350620
Using a PatLeaf for sext_16_node allowed matching smulbb and smlabb
instructions once the operands had been sign extended. But we also
need to use sext_inreg operands along with sext_16_node to catch a
few more cases that enable use to remove the unnecessary sxth.
Differential Revision: https://reviews.llvm.org/D55992
llvm-svn: 350613
I'm not entirely sure this is the correct thing
to do with the global isel philosophy, but I think
this is necessary to handle how differently SGPRs
are used normally vs. from a condition.
For example, it makes sense to allow a copy
from a VGPR to an SGPR, but it makes no sense
to allow a copy from VGPRs to SGPRs used as
select mask.
This avoids regbankselecting strange code with
a truncate feeding directly into a condition field.
Now a copy is forced from sgpr(s1) to vcc, which is
more sensible to handle.
Some of these issues could probably avoided with making enough
operations resulting in i1 illegal. I think we can't avoid
this register bank for legality.
For example, an i1 and where one source is from a truncate, and
one source is a compare needs some kind of copy inserted to
make sure both are in condition registers.
llvm-svn: 350611
Summary:
FixIrreducibleControlFlow and LateEHPrepare both possibly modify CFG and
create new registers. There seems to be no reason these passes go after
register-related optimization passes (PrepareForLiveIntervals,
OptimizeLiveIntervals, StoreResults, RegStackify, and RegColoring), and
this also possibly create new optimization opportunities. I think we
should put all current and future optimization passes before RegStackify
(and related passes) unless there's a reason not to.
Reviewers: kripken
Subscribers: dschuff, sbc100, sunfish, jgravelle-google, llvm-commits
Differential Revision: https://reviews.llvm.org/D56356
llvm-svn: 350596
If a copy was needed to handle the condition of brcond, it was being
inserted before the defining instruction. Add tests for iterator edge
cases.
I find the existing code here suspect for the case where it's looking
for terminators that modify the register. It's going to insert a copy
in the middle of the terminators, which isn't allowed (it might be
necessary to have a COPY_terminator if anybody actually needs this).
Also legalize brcond for AMDGPU.
llvm-svn: 350595
Summary:
We don't need to explicitly use `NI` anymore because we now don't use
`let` statements within the definitions.
Reviewers: aardappel
Subscribers: dschuff, sbc100, jgravelle-google, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D56376
llvm-svn: 350594
As we saw in D56057 when we tried to use this function on X86, it's unsafe. It allows the operand node to have multiple users, but doesn't prevent recursing past the first node when it does have multiple users. This can cause other simplifications earlier in the graph without regard to what bits are needed by the other users of the first node. Ideally all we should do to the first node if it has multiple uses is bypass it when its not needed by the user we started from. Doing any other transformation that SimplifyDemandedBits can do like turning ZEXT/SEXT into AEXT would result in an increase in instructions.
Fortunately, we already have a function that can do just that, GetDemandedBits. It will only make transformations that involve bypassing a node.
This patch changes AMDGPU's simplifyI24, to use a combination of GetDemandedBits to handle the multiple use simplifications. And then uses the regular SimplifyDemandedBits on each operand to handle simplifications allowed when the operand only has a single use. Unfortunately, GetDemandedBits simplifies constants more aggressively than SimplifyDemandedBits. This caused the -7 constant in the changed test to be simplified to remove the upper bits. I had to modify computeKnownBits to account for this by ignoring the upper 8 bits of the input.
Differential Revision: https://reviews.llvm.org/D56087
llvm-svn: 350560
This patch adds the sign/zero extension done by
vgetlane to ARM computeKnownBitsForTargetNode.
Differential revision: https://reviews.llvm.org/D56098
llvm-svn: 350553
Summary:
If a divergent branch instruction is marked as divergent by propagation
rule 2 in DivergencePropagator::exploreSyncDependency() and its condition
is uniform, that branch would incorrectly be assumed to be uniform.
Reviewers: arsenm, tstellar
Reviewed By: arsenm
Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D56331
llvm-svn: 350532
Summary: AVX512VBMI2 supports a funnel shift by immediate and a funnel shift by a variable vector.
Reviewers: spatel, RKSimon
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D56361
llvm-svn: 350498
There are no test changes here in the existing cost model
regression tests because integer add/sub have a default
legal cost of 1 already. This would break, however, if
we custom lower those ops because the default cost model
assumes that custom-lowered ops are more expensive.
This is similar to the change in rL350403. See discussion
in D56011 for more details. When we enhance that patch to
handle integer ops, we need this cost model change to avoid
unintended diffs here from the custom lowering.
llvm-svn: 350496
This was here because out and in instructions allow '(%dx)' even though its not a memory reference. To handle this we build a special operand for the DX register reference before we get to the call to CheckBaseRegAndIndexRegAndScale. So we no longer need this special case.
llvm-svn: 350483
This is especially helpful on targets without avx512bw since we don't have a good way to convert from v16i8/v32i8 to v16i1/v32i1 for the truncate anyway. If we're just going to convert it to a GPR we might as well use pmovmskb to accomplish both.
llvm-svn: 350480
We don't need to require the first operand to be an integer because we already said it was the same type as the result which we also constrained to an integer.
llvm-svn: 350455
Simon Pilgrim