v2i8/v4i8/v8i8 + v2i16/v4i16 all show up in vectorizer code and by just using the legalized types (v16i8/v8i16) we're highly exaggerating the actual cost of the shuffle.
This adds the instruction encoding and mnenomics for the proposed
RISC-V Bit Manipulation extension (version 0.92). It is implemented with
each category of instruction as its own target feature, with the 'b'
extension feature enabling all options. Since this extension is not yet
ratified, all target features are prefixed with 'experimental-' to note
their status.
Differential Revision: https://reviews.llvm.org/D65649
Summary:
This patch adds support for handling of variadic functions for AIX.
This includes ensuring that use and consume correct type of
va_list (char *va_list) for AIX.
Authored by: ZarkoCA
Reviewers: cebowleratibm, sfertile, jasonliu
Reviewed by: jasonliu
Differential Revision: https://reviews.llvm.org/D76130
Add initial support for PC Relative addressing for constant pool loads.
This includes adding a new relocation for @pcrel and adding a new PowerPC flag
to identify PC relative addressing.
Differential Revision: https://reviews.llvm.org/D74486
Summary:
VE has special operands to represent 0b000...000111...111 (`(m)0`) and
0b111...111000...000 (`(m)1`) bit sequences. This patch supports those
operands not only in machine instructions but also in DAG lowering.
Reviewed By: simoll
Differential Revision: https://reviews.llvm.org/D77769
InnerLoopVectorizer's code called during VPlan execution still relies on
original IR's def-use relations to decide which vector code to generate,
limiting VPlan transformations ability to modify def-use relations and still
have ILV generate the vector code.
This commit introduces VPValues for VPBlendRecipe to use as the values to
blend. The recipe is generated with VPValues wrapping the phi's incoming values
of the scalar phi. This reduces ingredient def-use usage by ILV as a step
towards full VPlan-based def-use relations.
Differential Revision: https://reviews.llvm.org/D77539
Ensured initialized fields; encapsulad delta calulations and evaluation
of threshold having had changed; assertion for CostThresholdMap
dereference, to indicate design intent.
Differential Revision: https://reviews.llvm.org/D77762
Introduce a new VPWidenCanonicalIVRecipe to generate a canonical vector
induction for use in fold-tail-with-masking, if a primary induction is absent.
The canonical scalar IV having start = 0 and step = VF*UF, created during code
-gen to control the vector loop, is widened into a canonical vector IV having
start = {<Part*VF, Part*VF+1, ..., Part*VF+VF-1> for 0 <= Part < UF} and
step = <VF*UF, VF*UF, ..., VF*UF>.
Differential Revision: https://reviews.llvm.org/D77635
This implements the instruction analysis required to print branch
targets as part of llvm-objdump's disassembly.
Note, this only handles those branches which can be analyzed in a single
instruction, a future patch will handle multiple-instruction patterns,
such as AUIPC/LUI+JALR instruction pairs.
Differential Revision: https://reviews.llvm.org/D77567
Summary:
Although the function had a bool return value, it was always returning
true. Presumably this is because the main type of errors one can
encounter here is running off the end of the stream, and until very
recently, the DataExtractor class made it very difficult to detect that.
The situation has changed now, and we can easily detect errors here,
which this patch does.
Reviewers: dblaikie, aprantl
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77308
When building a VPlan, BasicBlock::instructionsWithoutDebug() is used to
iterate over the instructions in a block. This means that no recipes
should be created for debug info intrinsics already and we can turn the
early exit into an assertion.
Reviewers: Ayal, gilr, rengolin, aprantl
Reviewed By: aprantl
Differential Revision: https://reviews.llvm.org/D77636
The change introduces the usage of physical registers for non-gc deopt values.
This require runtime support to know how to take a value from register.
By default usage is off and can be switched on by option.
The change also introduces additional fix-up patch which forces the spilling
of caller saved registers (clobbered after the call) and re-writes statepoint
to use spill slots instead of caller saved registers.
Reviewers: reames, dantrushin
Reviewed By: reames, dantrushin
Subscribers: mgorny, hiraditya, mgrang, llvm-commits
Differential Revision: https://reviews.llvm.org/D77371
This patch adds VPValue versions for the arguments of the call to
VPWidenCallRecipe and uses them during code-generation.
Similar to D76373 this reduces ingredient def-use usage by ILV as
a step towards full VPlan-based def-use relations.
Reviewers: Ayal, gilr, rengolin
Reviewed By: gilr
Differential Revision: https://reviews.llvm.org/D77655
Summary:
There are at least three clients for KnownBits calculations:
ValueTracking, SelectionDAG and GlobalISel. To reduce duplication the
common logic should be moved out of these clients and into KnownBits
itself.
This patch does this for AND, OR and XOR calculations by implementing
and using appropriate operator overloads KnownBits::operator& etc.
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D74060
Use the simpler BitWidth constructor instead of the copy constructor to
make it clear when we don't actually need to copy an existing KnownBits
value. Split out from D74539. NFC.
Now compiler defines 5 sets of constants to represent rounding mode.
These are:
1. `llvm::APFloatBase::roundingMode`. It specifies all 5 rounding modes
defined by IEEE-754 and is used in `APFloat` implementation.
2. `clang::LangOptions::FPRoundingModeKind`. It specifies 4 of 5 IEEE-754
rounding modes and a special value for dynamic rounding mode. It is used
in clang frontend.
3. `llvm::fp::RoundingMode`. Defines the same values as
`clang::LangOptions::FPRoundingModeKind` but in different order. It is
used to specify rounding mode in in IR and functions that operate IR.
4. Rounding mode representation used by `FLT_ROUNDS` (C11, 5.2.4.2.2p7).
Besides constants for rounding mode it also uses a special value to
indicate error. It is convenient to use in intrinsic functions, as it
represents platform-independent representation for rounding mode. In this
role it is used in some pending patches.
5. Values like `FE_DOWNWARD` and other, which specify rounding mode in
library calls `fesetround` and `fegetround`. Often they represent bits
of some control register, so they are target-dependent. The same names
(not values) and a special name `FE_DYNAMIC` are used in
`#pragma STDC FENV_ROUND`.
The first 4 sets of constants are target independent and could have the
same numerical representation. It would simplify conversion between the
representations. Also now `clang::LangOptions::FPRoundingModeKind` and
`llvm::fp::RoundingMode` do not contain the value for IEEE-754 rounding
direction `roundTiesToAway`, although it is supported natively on
some targets.
This change defines all the rounding mode type via one `llvm::RoundingMode`,
which also contains rounding mode for IEEE rounding direction `roundTiesToAway`.
Differential Revision: https://reviews.llvm.org/D77379
Summary:
New SanitizerCoverage feature `inline-bool-flag` which inserts an
atomic store of `1` to a boolean (which is an 8bit integer in
practice) flag on every instrumented edge.
Implementation-wise it's very similar to `inline-8bit-counters`
features. So, much of wiring and test just follows the same pattern.
Reviewers: kcc, vitalybuka
Reviewed By: vitalybuka
Subscribers: llvm-commits, hiraditya, jfb, cfe-commits, #sanitizers
Tags: #clang, #sanitizers, #llvm
Differential Revision: https://reviews.llvm.org/D77244
Dead constants might be left when a function is replaced, we can
gracefully handle this case and avoid complexity for the users who would
see an assertion otherwise.
Passing a Value * to CreateCall has to call getPointerElementType
to find the type of the pointer.
In this case we can rely on the fact that Intrinsic::getDeclaration
returns a Function * and use that version of CreateCall.
Attributor.cpp became quite big and we need to start provide structure.
The Attributor code is now in Attributor.cpp and the classes derived
from AbstractAttribute are in AttributorAttributes.cpp. Minor changes
were required but no intended functional changes.
We also minimized includes as part of this.
Reviewed By: baziotis
Differential Revision: https://reviews.llvm.org/D76873
Summary:
Remove usages of asserting vector getters in Type in preparation for the
VectorType refactor. The existence of these functions complicates the
refactor while adding little value.
Reviewers: espindola, efriedma, sdesmalen
Reviewed By: efriedma
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77275
Summary:
Remove usages of asserting vector getters in Type in preparation for the
VectorType refactor. The existence of these functions complicates the
refactor while adding little value.
Reviewers: hfinkel, efriedma, sdesmalen
Reviewed By: efriedma
Subscribers: wuzish, nemanjai, hiraditya, kbarton, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77266
Summary:
Remove usages of asserting vector getters in Type in preparation for the
VectorType refactor. The existence of these functions complicates the
refactor while adding little value.
Reviewers: sdesmalen, rriddle, efriedma
Reviewed By: sdesmalen
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77263
This patch introduces the heat coloring of the Control Flow Graph which is based
on the relative "hotness" of each BB. The patch is a part of sequence of three
patches, related to graphs Heat Coloring.
Reviewers: rcorcs, apilipenko, davidxl, sfertile, fedor.sergeev, eraman, bollu
Differential Revision: https://reviews.llvm.org/D77161
Generate PTX using newer versions of PTX and allow using sm_80 with CUDA-11.
None of the new features of CUDA-10.2+ have been implemented yet, so using these
versions will still produce a warning.
Differential Revision: https://reviews.llvm.org/D77670
Summary:
Preserve call site info for duplicated instructions. We copy over the
call site info in CloneMachineInstrBundle to avoid repeated calls to
copyCallSiteInfo in CloneMachineInstr.
(Alternatively, we could copy call site info higher up the stack, e.g.
into TargetInstrInfo::duplicate, or even into individual backend passes.
However, I don't see how that would be safer or more general than the
current approach.)
Reviewers: aprantl, djtodoro, dstenb
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77685
Any or all the argument registers can be used to pass a byval formal
argument, with the limitation that the argument must fit in the
available registers (ie: is not split between registers and stack).
Differential Revision: https://reviews.llvm.org/D76902
callCapturesBefore always returns ModRef , if UseInst isn't a call. As
we only call it if we already know Mod is set, this only destroys the
Must bit for non-calls.
On PowerPC most functions require a valid TOC pointer.
This is the case because either the function itself needs to use this
pointer to access the TOC or because other functions that are called
from that function expect a valid TOC pointer in the register R2.
The main exception to this is leaf functions that do not access the TOC
since they are guaranteed not to need a valid TOC pointer.
This patch introduces a feature that will allow more functions to not
require a valid TOC pointer in R2.
Differential Revision: https://reviews.llvm.org/D73664
Based on the post-commit comments for rG0f56bbc, there might
be a problem with this transform:
(bitcast (fpext/fptrunc X)) to iX) < 0 --> (bitcast X to iY) < 0
...and the ppc_fp128 data type, so conservatively bypass if we
are bitcasting a ppc_fp128.
We might be able to account for endian or other differences to
enable this for PowerPC again if that is useful.
Differential Revision: https://reviews.llvm.org/D77642
This is a performance patch that hoists two conditions in DwarfDebug's
validThroughout to avoid a linear-scan of all instructions in a block. We
now exit early if validThrougout will never return true for the variable
location.
The first added clause filters for the two circumstances where
validThroughout will return true. The second added clause should be
identical to the one that's deleted from after the linear-scan.
Differential Revision: https://reviews.llvm.org/D77639
Summary:
Since D75300 has been landed, I want to support enhanced relaxation when we need to align branches and allow prefix padding. "Enhanced Relaxtion" means we allow an instruction that could not be traditionally relaxed to be emitted into RelaxableFragment so that we increase its length by adding prefixes for optimization.
The motivation is straightforward, RelaxFragment is mostly for relative jumps and we can not increase the length of jumps when we need to align them, so if we need to achieve D75300's purpose (reducing the bytes of nops) when need to align jumps, we have to make more instructions "relaxable".
Reviewers: reames, MaskRay, craig.topper, LuoYuanke, jyknight
Reviewed By: reames
Subscribers: hiraditya, llvm-commits, annita.zhang
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76286
Summary:
This fixes PR45302.
Previously the case
BB1
/ \
| |
TBB FBB
| |
\ /
BB2
was treated as a valid diamond also when TBB and FBB was the same basic
block. This then lead to a failed assertion in IfConvertDiamond.
Since TBB == FBB is quite a degenerated case of a diamond, we now
don't treat it as a valid diamond anymore, and thus we will avoid the
trouble of making IfConvertDiamond handle it correctly.
Reviewers: efriedma, kparzysz
Reviewed By: efriedma
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D77651
This patch adds an analysis of the offset addresses used by gathers
and scatters to the MVEGatherScatterLowering pass to find
multiplications and additions that are loop invariant and thus can
be moved into the loop preheader, avoiding to execute them each time.
Differential Revision: https://reviews.llvm.org/D76681
Summary:
When narrowing G_IMPLICIT_DEF where the original size is not a multiple
of the narrow size, emit a smaller G_IMPLICIT_DEF and use G_ANYEXT.
To prevent a potential endless loop in the legalizer, the condition
to combine G_ANYEXT(G_IMPLICIT_DEF) is changed from isInstUnsupported
to !isInstLegal, since in this case the combine is only valid if
consequent legalization of the newly combined G_IMPLICIT_DEF does not
introduce G_ANYEXT due to narrowing.
Although this legalization for G_IMPLICIT_DEF would also be valid for
the general case, it actually caused a lot of code regressions when
tried due to superfluous COPYs and combines not getting hit anymore.
Reviewers: dsanders, aemerson, volkan, arsenm, aditya_nandakumar
Reviewed By: arsenm
Subscribers: jvesely, nhaehnle, kerbowa, wdng, rovka, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76598
In DWARFv5, type units are stored in .debug_info sections, along with
compilation units, and they are distinguished by the unit_type field
in the header, not by the name of the section. It is impossible to
associate the correct index section of a DWP file with the unit before
the unit's header is read. This patch fixes reading DWARFv5 type units
by parsing the header first and then applying the index entry according
to the actual unit type.
Differential Revision: https://reviews.llvm.org/D77552
By default, all traits in the OpenMP context selector have to match for
it to be acceptable. Though, we sometimes want a single property out of
multiple to match (=any) or no match at all (=none). We offer these
choices as extensions via
`implementation={extension(match_{all,any,none})}`
to the user. The choice will affect the entire context selector not only
the traits following the match property.
The first user will be D75788. There we can replace
```
#pragma omp begin declare variant match(device={arch(nvptx64)})
#define __CUDA__
#include <__clang_cuda_cmath.h>
// TODO: Hack until we support an extension to the match clause that allows "or".
#undef __CLANG_CUDA_CMATH_H__
#undef __CUDA__
#pragma omp end declare variant
#pragma omp begin declare variant match(device={arch(nvptx)})
#define __CUDA__
#include <__clang_cuda_cmath.h>
#undef __CUDA__
#pragma omp end declare variant
```
with the much simpler
```
#pragma omp begin declare variant match(device={arch(nvptx, nvptx64)}, implementation={extension(match_any)})
#define __CUDA__
#include <__clang_cuda_cmath.h>
#undef __CUDA__
#pragma omp end declare variant
```
Reviewed By: mikerice
Differential Revision: https://reviews.llvm.org/D77414
Summary:
ComputeValueKnownInPredecessorsImpl is the main folding mechanism in
JumpThreading.cpp. To avoid potential infinite recursion while
chasing use-def chains, it uses:
DenseSet<std::pair<Value *, BasicBlock *>> &RecursionSet
to keep track of Value-BB pairs that we've processed.
Now, when ComputeValueKnownInPredecessorsImpl recursively calls
itself, it always passes BB as is, so the second element is always BB.
This patch simplifes the function by dropping "BasicBlock *" from
RecursionSet.
Reviewers: wmi, efriedma
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77699
Summary:
Re-used the IR-level debugify for the most part. The MIR-level code then
adds locations to the MachineInstrs afterwards based on the LLVM-IR debug
info.
It's worth mentioning that the resulting locations make little sense as
the range of line numbers used in a Function at the MIR level exceeds that
of the equivelent IR level function. As such, MachineInstrs can appear to
originate from outside the subprogram scope (and from other subprogram
scopes). However, it doesn't seem worth worrying about as the source is
imaginary anyway.
There's a few high level goals this pass works towards:
* We should be able to debugify our .ll/.mir in the lit tests without
changing the checks and still pass them. I.e. Debug info should not change
codegen. Combining this with a strip-debug pass should enable this. The
main issue I ran into without the strip-debug pass was instructions with MMO's and
checks on both the instruction and the MMO as the debug-location is
between them. I currently have a simple hack in the MIRPrinter to
resolve that but the more general solution is a proper strip-debug pass.
* We should be able to test that GlobalISel does not lose debug info. I
recently found that the legalizer can be unexpectedly lossy in seemingly
simple cases (e.g. expanding one instr into many). I have a verifier
(will be posted separately) that can be integrated with passes that use
the observer interface and will catch location loss (it does not verify
correctness, just that there's zero lossage). It is a little conservative
as the line-0 locations that arise from conflicts do not track the
conflicting locations but it can still catch a fair bit.
Depends on D77439, D77438
Reviewers: aprantl, bogner, vsk
Subscribers: mgorny, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77446
dso_local leads to direct access even if the definition is not within this compilation unit (it is
still in the same linkage unit). On ELF, such a relocation (e.g. R_X86_64_PC32) referencing a
STB_GLOBAL STV_DEFAULT object can cause a linker error in a -shared link.
If the linkage is changed to available_externally, the dso_local flag should be dropped, so that no
direct access will be generated.
The current behavior is benign, because -fpic does not assume dso_local
(clang/lib/CodeGen/CodeGenModule.cpp:shouldAssumeDSOLocal).
If we do that for -fno-semantic-interposition (D73865), there will be an
R_X86_64_PC32 linker error without this patch.
Reviewed By: tejohnson
Differential Revision: https://reviews.llvm.org/D74751
Fix the error of show-prof-info.test on some platforms without zlib.
The common profile usage is to collect profile from a target and then use the profile to guide the optimized build for the same target. There are some cases that no profile can be collected for a target. In those cases, although no full profile is available, it is possible to have some partial profile collected from other targets to optimize common libraries and utilities. A flag is needed to tell the partial profile from the full profile apart, so compiler can use different strategy for them.
Differential Revision: https://reviews.llvm.org/D77426
The common profile usage is to collect profile from a target and then use the profile to guide the optimized build for the same target. There are some cases that no profile can be collected for a target. In those cases, although no full profile is available, it is possible to have some partial profile collected from other targets to optimize common libraries and utilities. A flag is needed to tell the partial profile from the full profile apart, so compiler can use different strategy for them.
Differential Revision: https://reviews.llvm.org/D77426
There are a few patterns where we use a superclass for inputs to this
instruction rather than the correct class. This can sometimes lead to
unncessary copies.
This patch extends existing constant folding in logical operations to
handle S_XNOR, S_NAND, S_NOR, S_ANDN2, S_ORN2, V_LSHL_ADD_U32 and
V_AND_OR_B32. Also added a couple of tests for existing folds.
This removes a call to getScalarType from a bunch of call sites.
It also makes the behavior consistent with SIGN_EXTEND_INREG.
Differential Revision: https://reviews.llvm.org/D77631
For implementing "remove obsolete debug info in lld", it is neccesary
to have DWARF generation code implementation. dsymutil uses DwarfStreamer
for that purpose. DwarfStreamer uses AsmPrinter. It is considered OK
to use AsmPrinter based code in lld(D74169). This patch moves
DwarfStreamer implementation into DWARFLinker, so that it could be reused
from lld.
Generally, a better place for such a common DWARF generation code would be
not DWARFLinker but an additional separate library. Such a library could
contain a single version of DWARF generation routines and could also
be independent of AsmPrinter. At the current moment, DwarfStreamer
does not pretend to be such a general implementation of DWARF generation.
So I decided to put it into DWARFLinker since it is the only user
of DwarfStreamer.
Testing: it passes "check-all" lit testing. MD5 checksum for clang .dSYM
bundle matches for the dsymutil with/without that patch.
Reviewed By: JDevlieghere
Differential revision: https://reviews.llvm.org/D77169
7aecf232 fixed the bug where we would miscompile, but we still generate
a crazy amount of code. Turn off the expansion until someone implements
an appropriate heuristic.
Differential Revision: https://reviews.llvm.org/D77599
We can only collapse adjacent SI_END_CF if outer statement
belongs to a simple SI_IF, otherwise correct mask is not in the
register we expect, but is an argument of an S_XOR instruction.
Even if SI_IF is simple it might be lowered using S_XOR because
lowering is dependent on a basic block layout. It is not
considered simple if instruction consuming its output is
not an SI_END_CF. Since that SI_END_CF might have already been
lowered to an S_OR isSimpleIf() check may return false.
This situation is an opportunity for a further optimization
of SI_IF lowering, but that is a separate optimization. In the
meanwhile move SI_END_CF post the lowering when we already know
how the rest of the CFG was lowered since a non-simple SI_IF
case still needs to be handled.
Differential Revision: https://reviews.llvm.org/D77610
These should not be assuming address space 0. Calling getPointerTy is
generally the wrong thing to do, since you should already know the
type from the incoming IR.
RDA sometimes needs to visit blocks twice, to take into account
reaching defs coming in along loop back edges. Currently it handles
repeated visitation the same way as usual, which means that it will
scan through all instructions and their reg unit defs again. Not
only is this very inefficient, it also means that all reaching defs
in loops are going to be inserted twice.
We can do much better than this. The only thing we need to handle
is a new reaching def from a predecessor, which either needs to be
prepended to the reaching definitions (if there was no reaching def
from a predecessor), or needs to replace an existing predecessor
reaching def, if it is more recent. Since D77508 we only store the
most recent predecessor reaching def, so that's the only one that
may need updating.
This also has the nice side-effect that reaching definitions are
now automatically sorted and unique, so drop the llvm::sort() call
in favor of an assertion.
Differential Revision: https://reviews.llvm.org/D77511
An instruction may define the same reg unit multiple times,
avoid inserting the same reaching def multiple times in that case.
Also print the reg unit, rather than the super-register, in the
debug code.
Summary:
- Remove the no longer used Darwin CalleeSavedRegs
- Combine the SVR464 callee saved regs and AIX64 since the two are (and should be) identical into PPC64
- Update tests for 64-bit CSR change
Reviewers: sfertile, ZarkoCA, cebowleratibm, jasonliu, #powerpc
Reviewed By: sfertile
Subscribers: wuzish, nemanjai, hiraditya, kbarton, shchenz, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77235
Shuffle combining can insert zero byte sized elements into the shuffle mask, which combineX86ShufflesConstants will attempt to fold without taking into account whether the byte-sized type is legal (e.g. AVX512F only targets).
If we have a full-zeroable vector then we should just return a zero version of the root type, otherwise if the type isn't valid we should bail.
Fixes PR45443
If the stack pointer is altered for local variables and we are generating
Thumb2 execute-only code the .pad directive is missing.
Usually the size of the adjustment is stored in a PC-relative location
and loaded into a register which is then added to the stack pointer.
However when we are generating execute-only code code the size of the
adjustment is instead generated using the MOVW/MOVT instruction pair.
As a by product of handling the execute-only case this also fixes an
existing issue that in the none execute-only case the .pad directive was
generated against the load of the constant to a register instruction,
instead of the instruction which adds the register to the stack pointer.
Differential Revision: https://reviews.llvm.org/D76849
This patch updates the code that deals with conditions from predicate
info to make use of constant ranges.
For ssa_copy instructions inserted by PredicateInfo, we have 2 ranges:
1. The range of the original value.
2. The range imposed by the linked condition.
1. is known, 2. can be determined using makeAllowedICmpRegion. The
intersection of those ranges is the range for the copy.
With this patch, we get a nice increase in the number of instructions
eliminated by both SCCP and IPSCCP for some benchmarks:
For MultiSource, SPEC2000 & SPEC2006:
Tests: 237
Same hash: 170 (filtered out)
Remaining: 67
Metric: sccp.NumInstRemoved
Program base patch diff
test-suite...Source/Benchmarks/sim/sim.test 10.00 71.00 610.0%
test-suite...CFP2000/177.mesa/177.mesa.test 361.00 1626.00 350.4%
test-suite...encode/alacconvert-encode.test 141.00 602.00 327.0%
test-suite...decode/alacconvert-decode.test 141.00 602.00 327.0%
test-suite...CI_Purple/SMG2000/smg2000.test 1639.00 4093.00 149.7%
test-suite...peg2/mpeg2dec/mpeg2decode.test 75.00 163.00 117.3%
test-suite...T2006/401.bzip2/401.bzip2.test 358.00 513.00 43.3%
test-suite...rks/FreeBench/pifft/pifft.test 11.00 15.00 36.4%
test-suite...langs-C/unix-tbl/unix-tbl.test 4.00 5.00 25.0%
test-suite...lications/sqlite3/sqlite3.test 541.00 667.00 23.3%
test-suite.../CINT2000/254.gap/254.gap.test 243.00 299.00 23.0%
test-suite...ks/Prolangs-C/agrep/agrep.test 25.00 29.00 16.0%
test-suite...marks/7zip/7zip-benchmark.test 1135.00 1304.00 14.9%
test-suite...lications/ClamAV/clamscan.test 1105.00 1268.00 14.8%
test-suite...urce/Applications/lua/lua.test 398.00 436.00 9.5%
Metric: sccp.IPNumInstRemoved
Program base patch diff
test-suite...C/CFP2000/179.art/179.art.test 1.00 3.00 200.0%
test-suite...006/447.dealII/447.dealII.test 429.00 1056.00 146.2%
test-suite...nch/fourinarow/fourinarow.test 3.00 7.00 133.3%
test-suite...CI_Purple/SMG2000/smg2000.test 818.00 1748.00 113.7%
test-suite...ks/McCat/04-bisect/bisect.test 3.00 5.00 66.7%
test-suite...CFP2000/177.mesa/177.mesa.test 165.00 255.00 54.5%
test-suite...ediabench/gsm/toast/toast.test 18.00 27.00 50.0%
test-suite...telecomm-gsm/telecomm-gsm.test 18.00 27.00 50.0%
test-suite...ks/Prolangs-C/agrep/agrep.test 24.00 35.00 45.8%
test-suite...TimberWolfMC/timberwolfmc.test 43.00 62.00 44.2%
test-suite...encode/alacconvert-encode.test 46.00 66.00 43.5%
test-suite...decode/alacconvert-decode.test 46.00 66.00 43.5%
test-suite...langs-C/unix-tbl/unix-tbl.test 12.00 17.00 41.7%
test-suite...peg2/mpeg2dec/mpeg2decode.test 31.00 41.00 32.3%
test-suite.../CINT2000/254.gap/254.gap.test 117.00 154.00 31.6%
Reviewers: efriedma, davide
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D76611
Move the logic whether lowering of deopt value requires a spill slot in
a separate lambda.
Reviewers: reames, dantrushin
Reviewed By: dantrushin
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D77629
From Arm v8 Architecture Reference Manual F5.1.84 LDREXD
The ldrexd instruction in Arm state has the following conditions:
t = UInt(Rt); t2 = t + 1; n = UInt(Rn);
if Rt<0> == '1' || t2 == 15 || n == 15 then UNPREDICTABLE;
In when Rt is odd or if Rt is 14 (making t2 15).
In the implementation when the pair is the UNPREDICTABLE R14_R15 we
would ideally return SOFT_FAIL. We can't because there is no R14_R15
value for us to return so we fail early returning FAIL.
The early return for registers outside the bounds of the table means
the check for Rt == 14 (0xE) redundant which causes a static analyzer
to flag the condition as never being true.
To fix the warning I've removed the check and replaced with a comment
explaining the difference with the specification.
Fixes pr41660
Differential Revision: https://reviews.llvm.org/D77463
Summary:
In `Unix/Process.inc`, we seed a random number generator from
`/dev/urandom` if possible, but if not, we're happy to fall back to
ordinary pseudorandom strategies, like the current time and PID.
The corresponding function on Windows calls `CryptGenRandom`, but it
//doesn't// have a fallback if that strategy fails. But `CryptGenRandom`
//can// fail, if a cryptography provider isn't properly initialized, or
occasionally (by our observation) simply intermittently.
If it's reasonable on Unix to implement traditional pseudorandom-number
seeding as a fallback, then it's surely reasonable to do the same on
Windows. So this patch adds a last-ditch use of ordinary rand(), using
much the same strategy as the Unix fallback code.
Reviewers: hans, sammccall
Reviewed By: hans
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77553
Do not commit the llvm/test/ExecutionEngine/MCJIT/cet-code-model-lager.ll because it will
cause build bot fail(not suitable for window 32 target).
Summary:
This patch comes from H.J.'s 2bd54ce7fa
**This patch fix the failed llvm unit tests which running on CET machine. **(e.g. ExecutionEngine/MCJIT/MCJITTests)
The reason we enable IBT at "JIT compiled with CET" is mainly that: the JIT don't know the its caller program is CET enable or not.
If JIT's caller program is non-CET, it is no problem JIT generate CET code or not.
But if JIT's caller program is CET enabled, JIT must generate CET code or it will cause Control protection exceptions.
I have test the patch at llvm-unit-test and llvm-test-suite at CET machine. It passed.
and H.J. also test it at building and running VNCserver(Virtual Network Console), it works too.
(if not apply this patch, VNCserver will crash at CET machine.)
Reviewers: hjl.tools, craig.topper, LuoYuanke, annita.zhang, pengfei
Reviewed By: LuoYuanke
Subscribers: tstellar, efriedma, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76900
Now that we have scalable vectors, there's a distinction that isn't
getting captured in the original SequentialType: some vectors don't have
a known element count, so counting the number of elements doesn't make
sense.
In some cases, there's a better way to express the commonality using
other methods. If we're dealing with GEPs, there's GEP methods; if we're
dealing with a ConstantDataSequential, we can query its element type
directly.
In the relatively few remaining cases, I just decided to write out
the type checks. We're talking about relatively few places, and I think
the abstraction doesn't really carry its weight. (See thread "[RFC]
Refactor class hierarchy of VectorType in the IR" on llvmdev.)
Differential Revision: https://reviews.llvm.org/D75661
Summary:
In lieu of a proper pass that strips debug info, add a way
to omit debug-locations from the MIR output so that
instructions with MMO's continue to match CHECK's when
mir-debugify is used
Reviewers: aprantl, bogner, vsk
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77575
Summary:
Thanks to Bill Wendling (void) for the report and steps to reproduce. It looks
like this was missed during r350508's cleanup of the CallSite split into
CallBase, CallInst, and CallBrInst.
This was exposed by running pgo on a callbr, which was creating a ptrtoint to
the inline asm thinking it was an indirect call. The relevant callchain looks
like:
IndirectCallPromotionPlugin::run()
-> PGOIndirectCallVisitor::findIndirectCalls()
-> PGOIndirectCallVisitor::visitCallBase()
-> CallBase::isIndirectCall()
Reviewers: void, chandlerc
Reviewed By: void
Subscribers: hiraditya, llvm-commits, craig.topper, srhines
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77600
Summary:
In some cases, ASan may insert instrumentation before function arguments
have been stored into their allocas. This causes two issues:
1) The argument value must be spilled until it can be stored into the
reserved alloca, wasting a stack slot.
2) Until the store occurs in a later basic block, the debug location
will point to the wrong frame offset, and backtraces will show an
uninitialized value.
The proposed solution is to move instructions which initialize allocas
for arguments up into the entry block, before the position where ASan
starts inserting its instrumentation.
For the motivating test case, before the patch we see:
```
| 0033: movq %rdi, 0x68(%rbx) | | DW_TAG_formal_parameter |
| ... | | DW_AT_name ("a") |
| 00d1: movq 0x68(%rbx), %rsi | | DW_AT_location (RBX+0x90) |
| 00d5: movq %rsi, 0x90(%rbx) | | ^ not correct ... |
```
and after the patch we see:
```
| 002f: movq %rdi, 0x70(%rbx) | | DW_TAG_formal_parameter |
| | | DW_AT_name ("a") |
| | | DW_AT_location (RBX+0x70) |
```
rdar://61122691
Reviewers: aprantl, eugenis
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77182
Summary:
To debugify MIR, we need to be able to create metadata and to do that, we
need a non-const Module. However, MachineFunction only had a const reference
to the Function preventing this.
Reviewers: aprantl, bogner
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77439
Summary:
It can be helpful to test behaviour w.r.t locations without having DEBUG_VALUE
around. In particular, because DEBUG_VALUE has the potential to change CodeGen
behaviour (e.g. hasOneUse() vs hasOneNonDbgUse()) while locations generally
don't.
Reviewers: aprantl, bogner
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77438
A global symbol that is defined in a comdat should not generate an alias since
call sites that would've referred to that symbol will refer to their own
independent local aliases rather than the surviving global comdat one. This
could result in something that looks like:
```
ld.lld: error: relocation refers to a discarded section: .text._ZN3fbl8internal18NullFunctionTargetIvJjjPjEED1Ev.stub
>>> defined in user-x64-clang/obj/system/ulib/minfs/libminfs.a(minfs._sources.file.cc.o)
>>> section group signature: _ZN3fbl8internal18NullFunctionTargetIvJjjPjEED1Ev.stub
>>> prevailing definition is in user-x64-clang/obj/system/ulib/minfs/libminfs.a(minfs._sources.vnode.cc.o)
>>> referenced by function.h:169 (../../zircon/system/ulib/fbl/include/fbl/function.h:169)
>>> minfs._sources.file.cc.o:(minfs::File::AllocateAndCommitData(std::__2::unique_ptr<minfs::Transaction, std::__2::default_delete<minfs::Transaction> >)) in archive user-x64-clang/obj/system/ulib/minfs/libminfs.a
```
We ran into this when experimenting with a new C++ ABI for fuchsia
(refer to D72959) which takes relative offsets between comdat'd functions
which is why the normal C++ user wouldn't run into this.
Differential Revision: https://reviews.llvm.org/D77429
Summary:
A bug report mentioned that LLVM was producing jumps off the end of a
function when using "asm goto with outputs". Further digging pointed to
MachineBasicBlocks that had their address taken and were indirect
targets of INLINEASM_BR being removed by BranchFolder, because their
predecessor list was empty, so they appeared to have no entry.
This was a cascading failure caused earlier, during Pre-RA instruction
scheduling. We have a few special cases in Pre-RA instruction scheduling
where we split a MachineBasicBlock in two. This requires careful
handing of predecessor and successor lists for a MachineBasicBlock that
was split, and careful handing of PHI MachineInstrs that referred to the
MachineBasicBlock before it was split.
The clue that led to this fix was the observation that many callers of
MachineBasicBlock::splice() frequently call
MachineBasicBlock::transferSuccessorsAndUpdatePHIs() to update their PHI
nodes after a splice. We don't want to reuse that method, as we have
custom successor transferring logic for this block split.
This patch fixes 2 pre-existing bugs, and adds tests.
The first bug was that MachineBasicBlock::splice() correctly handles
updating most successors and predecessors; we don't need to do anything
more than removing the previous fallthrough block from the first half of
the split block post splice. Previously, we were updating the successor
list incorrectly (updating successors updates predecessors).
The second bug was that PHI nodes that needed registers from the first
half of the split block were not having entries populated. The register
live out information was correct, and the FuncInfo->PHINodesToUpdate was
correct. Specifically, the check in SelectionDAGISel::FinishBasicBlock:
for (unsigned i = 0, e = FuncInfo->PHINodesToUpdate.size(); i != e; ++i) {
MachineInstrBuilder PHI(*MF, FuncInfo->PHINodesToUpdate[i].first);
if (!FuncInfo->MBB->isSuccessor(PHI->getParent()))
continue;
PHI.addReg(FuncInfo->PHINodesToUpdate[i].second).addMBB(FuncInfo->MBB);
was `continue`ing because FuncInfo->MBB tracks the second half of
the post-split block; no one was updating PHI entries for the first half
of the post-split block.
SelectionDAGBuilder::UpdateSplitBlock() already expects to perform
special handling for MachineBasicBlocks that were split post calls to
ScheduleDAGSDNodes::EmitSchedule(), so I'm confident that it's both
correct for ScheduleDAGSDNodes::EmitSchedule() to return the second half
of the split block `CopyBB` which updates `FuncInfo->MBB` (ie. the
current MachineBasicBlock being processed), and perform special handling
for this in SelectionDAGBuilder::UpdateSplitBlock().
Reviewers: void, craig.topper, efriedma
Reviewed By: void, efriedma
Subscribers: hfinkel, fhahn, MatzeB, efriedma, hiraditya, llvm-commits, srhines
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76961
The extracts from control flow intrinsics are already properly handled
by divergence analysis. The inline asm case isn't dead, but has also
never really worked correctly so leave it as-is for now.
This pass is created in d6de5f12d4 and tested
for new and legacy pass manager but never added to new pass manager pipeline.
I am adding it to new pass manager pipeline.
This pass is get used in Vector Function Database (VFDatabase) and without
this pass in new pass manager pipeline, none of the vector libraries are work
ing with new pass manager.
Related passes:
66c120f025https://reviews.llvm.org/D74944
Differential revision: https://reviews.llvm.org/D75354
The previous code used the type of the first field for the VT
passed to getNode for every field.
I've based the implementation here off what is done in visitSelect
as it removes the need to special case aggregates.
Differential Revision: https://reviews.llvm.org/D77093
The patch introduces the system to distinctively store the information
needed for the Control Flow Graph as well as the instrumentary needed for
the follow-up changes: BlockFrequencyInfo and BranchProbabilityInfo.
The patch is a part of sequence of three patches, related to graphs Heat Coloring.
Reviewers: rcorcs, apilipenko, davidxl, sfertile, fedor.sergeev, eraman, bollu
Differential Revision: https://reviews.llvm.org/D76820
So that constant expressions like the following are permitted:
and w0, w0, #~(0xfe<<24)
and w1, w1, #~(0xff<<24)
The behavior matches GNU as (opcodes/aarch64-opc.c:aarch64_logical_immediate_p).
Reviewed By: sdesmalen
Differential Revision: https://reviews.llvm.org/D75885
When entering a basic block, RDA inserts reaching definitions coming
from predecessor blocks (which will be negative numbers) in a rather
peculiar way. If you have incoming reaching definitions -4, -3, -2, -1,
it will insert those. If you have incoming reaching definitions
-1, -2, -3, -4, it will insert -1, -1, -1, -1, as the max is taken
at each step. That's probably not what was intended...
However, RDA only actually cares about the most recent reaching
definition from a predecessor (to calculate clearance), so this
ends up working fine as far as behavior is concerned. It does
waste memory on unnecessary reaching definitions though.
This patch changes the implementation to first compute the most
recent reaching definition in one loop, and then insert only that
one in a separate loop.
Differential Revision: https://reviews.llvm.org/D77508
At the end of a basic block, RDA adjusts all the reaching defs it
found to be relative to the end of the basic block, rather than the
start of it. However, it also does this to registers which don't
have a reaching def, indicated by ReachingDefDefaultVal. This means
that code checking against ReachingDefDefaultVal will not skip them,
and may insert them into the reaching definition list. This is
ultimately harmless, but causes unnecessary work and is logically
not right.
Differential Revision: https://reviews.llvm.org/D77506
The cmyk test is based on the known regression that resulted from:
rGf2fbdf76d8d0
This improves on the equivalent signed min/max change:
rG867f0c3c4d8c
The underlying icmp equivalence is:
~X pred ~Y --> Y pred X
For an icmp with constant, canonicalization results in a swapped pred:
~X < C --> X > ~C
This patch moves calls to their own recipe, to simplify the transition
to VPUser for operands of VPWidenRecipe, as discussed in D76992.
Subsequently additional information can be added to the recipe rather
than computing it during the execute step.
Reviewers: rengolin, Ayal, gilr, hsaito
Reviewed By: gilr
Differential Revision: https://reviews.llvm.org/D77467
Summary:
In D77454 we explain that `LoadInst` and `StoreInst` always have their alignment defined.
This allows to work backward here and to infer that `getNewAlignment` does not need to return `0` in case of failure.
Returning `1` also works since it needs to be greater than the Load/Store alignment which is a least `1`.
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: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77538
SUMMARY:
For the llvm-objdump -D, the symbol name is used as a label in the disassembly for the specific address (when a symbol address is equal to the virtual address in the dump).
In XCOFF, multiple symbols may have the same name, being differentiated by their storage mapping class. It is helpful to print the QualName and not just the name when forming the output label for a csect symbol. The symbol index further removes any ambiguity caused by duplicate names.
To maintain compatibility with the binutils objdump, the XCOFF-specific --symbol-description option is added to enable the enhanced format.
Reviewers: hubert.reinterpretcast, James Henderson, Jason Liu ,daltenty
Subscribers: wuzish, nemanjai, hiraditya
Differential Revision: https://reviews.llvm.org/D72973
ExecutionEngine/MCJIT/cet-code-model-lager.ll is failing on 32-bit
windows, see llvm-commits thread for fef2dab.
This reverts commit 43f031d312
and the follow-ups fef2dab100 and
6a800f6f62.
Summary:
This patch adds support for emission of following DWARFv5 macro forms
in .debug_macro section.
1. DW_MACRO_start_file
2. DW_MACRO_end_file
3. DW_MACRO_define_strp
4. DW_MACRO_undef_strp.
Reviewed By: dblaikie, ikudrin
Differential Revision: https://reviews.llvm.org/D72828
Summary:
This patch adds the optional Error argument, and the Cursor variants to
more DataExtractor methods. The functions now behave the same way as
other error-aware functions (they set the error when they fail, and
don't do anything if the error is already set).
I have merged the LEB128 implementations via a template (similarly to
how fixed-size functions are handled) to reduce code duplication.
Depends on D77304.
Reviewers: dblaikie, aprantl
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77306
Summary:
This patch adds an optional Error argument to DataExtractor functions
for string extraction, and makes them behave like other DataExtractor
functions (set the error if extraction fails, don't do anything if the
error is already set).
I have merged the StringRef and C string versions of the functions to
reduce code duplication.
Reviewers: dblaikie, MaskRay
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77307
Summary:
This is a roll forward of D77394 minus AlignmentFromAssumptions (which needs to be addressed separately)
Differences from D77394:
- DebugStr() now prints the alignment value or `None` and no more `Align(x)` or `MaybeAlign(x)`
- This is to keep Warning message consistent (CodeGen/SystemZ/alloca-04.ll)
- Removed a few unneeded headers from Alignment (since it's included everywhere it's better to keep the dependencies to a minimum)
Reviewers: courbet
Subscribers: sdardis, hiraditya, jrtc27, atanasyan, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77537
Summary:
24 March 2020: LLVM 10.0.0 is out.
I gathered all deprecated function introduced between 9 and 10 and cleaned them up so they will be removed from 11.
> git log -p -S LLVM_ATTRIBUTE_DEPRECATED llvmorg-9.0.0..llvmorg-10.0.0
Reviewers: courbet
Subscribers: hiraditya, jfb, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77409
truncateVectorWithPACK has its own vector length controls, so we can rely on those directly. This helps some existing truncation to subvector tests, which were being combined later during shuffle lowering at which point the sign/zero bit detection had become obscured preventing lowerShuffleWithPACK working as well as it could.
Summary:
Modify lea/load/store instructions to accept `disp(index, base)`
style addressing mode (called ASX format). Also, uniform the
number of DAG nodes to have 3 operands for this ASX format
instructions, and update selectADDR functions to lower
appropriate MI.
Reviewers: arsenm, simoll, k-ishizaka
Reviewed By: simoll
Differential Revision: https://reviews.llvm.org/D76822
This patch adds a -matrix-default-layout option which can be used to
set the default matrix layout to row-major or column-major (default).
The initial patch updates codegen for loads, stores, binary operators
and matrix multiply.
Reviewers: anemet, Gerolf, andrew.w.kaylor, LuoYuanke
Reviewed By: anemet
Differential Revision: https://reviews.llvm.org/D76325
This patch adds initial fusion for load/multiply/store chains of matrix
operations.
The patch contains roughly two parts:
1. Code generation for a fused load/multiply/store chain (LowerMatrixMultiplyFused).
First, we ensure that both loads of the multiply operands do not alias the store.
If they do, we create new non-aliasing copies of the operands. Note that this
may introduce new basic block. Finally we process TileSize x TileSize blocks.
That is: load tiles from the input operands, multiply and store them.
2. Identify fusion candidates & matrix instructions.
As a first step, collect all instructions with shape info and fusion candidates
(currently @llvm.matrix.multiply calls). Next, try to fuse candidates and
collect instructions eliminated by fusion. Finally iterate over all matrix
instructions, skip the ones eliminated by fusion and lower the rest as usual.
Reviewers: anemet, Gerolf, hfinkel, andrew.w.kaylor, LuoYuanke
Reviewed By: anemet
Differential Revision: https://reviews.llvm.org/D75566
In package files, the base offset provided by index sections should be
used to find the contribution of a unit. The patch adds that base
offset when reading range list tables.
Differential revision: https://reviews.llvm.org/D77401
This fixes the reading of location lists headers for compilation units
in package files by adjusting the reading offset according to the
corresponding record in the unit index. This is required for
DW_FORM_loclistx to work.
Differential revision: https://reviews.llvm.org/D77146
Without the patch, all version 5 compile units in a DWP file read
location tables from the beginning of a .debug_loclists.dwo section.
The patch fixes that by adjusting the reading offset the same way as
for pre-v5 units. The section identifier to find the contribution
entry corresponds to the version of the unit.
Differential revision: https://reviews.llvm.org/D77145
DWARFv5 defines index sections in package files in a slightly different
way than the pre-standard GNU proposal, see Section 7.3.5 in the DWARF
standard and https://gcc.gnu.org/wiki/DebugFissionDWP for GNU proposal.
The main concern here is values for section identifiers, which are
partially overlapped with changed meanings. The patch adds support for
v5 index sections and resolves that difficulty by defining a set of
identifiers for internal use which can represent and distinct values
of both standards.
Differential Revision: https://reviews.llvm.org/D75929
This is a preparation for an upcoming patch which adds support for
DWARFv5 unit index sections. The patch adds tag "_EXT_" to identifiers
which reference sections that are deprecated in the DWARFv5 standard.
See D75929 for the discussion.
Differential Revision: https://reviews.llvm.org/D77141
We're ANDing with 1 right after which will cause the SIGN_EXTEND to
be combined to ANY_EXTEND later. Might as well just start with an
ANY_EXTEND.
While there replace create the AND using the getZeroExtendInReg
helper to remove the need to explicitly create the VecOnes constant.
Move the listing of allowed clauses per OpenMP directive to the new
macro file in `llvm/Frontend/OpenMP`. Also, use a single generic macro
that specifies the directive and one allowed clause explicitly instead
of a dedicated macro per directive.
We save 800 loc and boilerplate for all new directives/clauses with no
functional change. We also need to include the macro file only once and
not once per directive.
Depends on D77112.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D77113
This code is replacing a shift with a new shift on an extended type.
If the shift amount type can't represent the maximum shift amount
for the new type, the amount needs to be extended to a type that
can.
Previously, the code just hardcoded a check for 256 bits which
seems to have been an assumption that the original shift amount
was MVT::i8. But that seems more catered to a specific target
like X86 that uses i8 as its legal shift amount type. Other
targets may use different types.
This commit changes the code to look at the real type of the shift
amount and makes sure it has enough bits for the Log2 of the
new type. There are similar checks to this in SelectionDAGBuilder
and LegalizeIntegerTypes.
This is a cleanup and normalization patch that also enables reuse with
Flang later on. A follow up will clean up and move the directive ->
clauses mapping.
Reviewed By: fghanim
Differential Revision: https://reviews.llvm.org/D77112
The new and old pass managers (PassManagerBuilder.cpp and
PassBuilder.cpp) are exposed to an `extern` declaration of
`attributor-disable` option which will guard the addition of the
attributor passes to the pass pipelines.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D76871
Add a new overload of StaticLibraryDefinitionGenerator::Load that takes a triple
argument and supports loading archives from MachO universal binaries in addition
to regular archives.
The LLI tool is updated to use this overload.
We had previously limited the shuffle(HORIZOP,HORIZOP) combine to binary shuffles, but we can often merge unary shuffles just as well, folding in UNDEF/ZERO values into the 64-bit half lanes.
For the (P)HADD/HSUB cases this is limited to fast-horizontal cases but PACKSS/PACKUS combines under all cases.
This patch builds upon D76140 by updating metadata on pointer typed
loads in inlined functions, when the load is the return value, and the
callsite contains return attributes which can be updated as metadata on
the load.
Added test cases show this for nonnull, dereferenceable,
dereferenceable_or_null
Reviewed-By: jdoerfert
Differential Revision: https://reviews.llvm.org/D76792
Previously line table symbol was represented as `DIE::value_iterator`
inside `DwarfCompileUnit` and subsequent function `intStmtList`
was used to create a local `MCSymbol` to initialize it.
This patch removes `DIE::value_iterator` from `DwarfCompileUnit`
and intoduce `MCSymbol` for representing this units symbol for
`debug_line` section. As a result `applyStmtList` is also modified
to utilize this. Further more a helper function `getLineTableStartSym`
is also introduced to get this symbol, this would be used by clients
which need to access this line table, i.e `debug_macro`.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D77489
The newly-created constant zero will need an extra register to hold it
in the current statepoint lowering implementation. Remove it if there
exists one.
Summary:
This patch upstreams support the optional ARMv8.0 Data Gathering Hint (DGH)
extension, which adds the Data Gathering Hint instruction to the hint
space.
See ARMv8.0-DGH in the Arm Architecture Reference Manual Armv8 for more
information.
Reviewers: t.p.northover, rengolin, SjoerdMeijer, ab, danielkiss, samparker
Reviewed By: SjoerdMeijer
Subscribers: LukeGeeson, ostannard, kristof.beyls, hiraditya, danielkiss, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77097
Summary:
This patch upstreams support for the ARMv8.6A Enhanced Counter Virtualization
(ECV) extension, which adds 6 new system registers.
See ARMv8.6-ECV in the Arm Architecture Reference Manual Armv8 for more
information.
Reviewers: t.p.northover, rengolin, SjoerdMeijer, pcc, ab, chill
Reviewed By: SjoerdMeijer
Subscribers: LukeGeeson, ostannard, kristof.beyls, hiraditya, danielkiss, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77094
As discussed in D76983, that patch can turn a chain of insert/extract
with scalar trunc ops into bitcast+extract and existing instcombine
vector transforms end up creating a shuffle out of that (see the
PhaseOrdering test for an example). Currently, that process requires
at least this sequence: -instcombine -early-cse -instcombine.
Before D76983, the sequence of insert/extract would reach the SLP
vectorizer and become a vector trunc there.
Based on a small sampling of public targets/types, converting the
shuffle to a trunc is better for codegen in most cases (and a
regression of that form is the reason this was noticed). The trunc is
clearly better for IR-level analysis as well.
This means that we can induce "spontaneous vectorization" without
invoking any explicit vectorizer passes (at least a vector cast op
may be created out of scalar casts), but that seems to be the right
choice given that we started with a chain of insert/extract, and the
backend would expand back to that chain if a target does not support
the op.
Differential Revision: https://reviews.llvm.org/D77299
Summary:
This patch upstreams support for the ARMv8.6A Fine Grain Traps (FGT)
extension, which adds 5 new system registers.
See ARMv8.6-FGT in the Arm Architecture Reference Manual Armv8 for more
information.
Reviewers: t.p.northover, rengolin, SjoerdMeijer, ab, momchil.velikov
Reviewed By: SjoerdMeijer
Subscribers: LukeGeeson, ostannard, kristof.beyls, hiraditya, danielkiss, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76991
The cmyk tests are based on the known regression that resulted from:
rGf2fbdf76d8d0
So this improvement in analysis might be enough to restore that commit.
Summary:
This patch upstreams v8.6A activity monitors virtualization
assembler support, which consists of 32 new system
registers (two groups, each with 16 numbered registers).
See ARMv8.6-AMU in the Arm Architecture Reference Manual Armv8 for more
information.
Reviewers: t.p.northover, rengolin, SjoerdMeijer, ab, john.brawn, ostannard
Reviewed By: ostannard
Subscribers: LukeGeeson, dnsampaio, ostannard, kristof.beyls, hiraditya, danielkiss, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76998
D51664 added Instruction::comesBefore which should provide better
performance than the manual check.
Reviewers: rnk, nikic, spatel
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D76228
Our existing combine allows to merge the shuffle of 2 similar 64-bit wide 'horizontal ops' (HADD/PACK/etc.) if the shuffle was a UNPCK/MOVSD.
This patch generalizes this to decode any target shuffle mask that can be widened to a 128-bit repeating v2*64 mask, which helps us catch PBLENDW/PBLENDD cases.
If we're packing from 128-bits to 64-bits then we don't need the RHS argument. This helps with register allocation, especially as we avoid repeating a use of the input value.
Query AAValueSimplify on pointers in memory accessing instructions to take
advantage of the constant propagation (or any other value simplification) of such values.
After 49d00824bb, VPWidenRecipe only stores a single instruction.
tryToWiden can simply return the widen recipe, like other helpers in
VPRecipeBuilder.
Summary:
D77423 started using a dominator tree in WasmEHPrepare, but we deleted
BBs in `prepareThrows` before we used the domtree in `prepareEHPads`,
and those CFG changes were not reflected in the domtree. This uses
`DomTreeUpdater` to make sure we update the domtree every time we delete
BBs from the CFG. This fixes ubsan/msan/expensive_check errors caught in
LLVM buildbots.
Reviewers: dschuff
Subscribers: sbc100, jgravelle-google, hiraditya, sunfish, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77465
eraseInstFromFunction() adds the operands of the erased instructions,
as those might now be dead as well. However, this is limited to
instructions with less than 8 operands.
This check doesn't make a lot of sense to me. As the instruction
gets removed afterwards, I don't see a potential for anything
overly pathological happening here (as we can only add those
operands to the worklist once). The impact on CTMark is in
the noise. We also have the same code in instruction sinking
and don't limit the operand count there.
Differential Revision: https://reviews.llvm.org/D77325
Summary:
When we insert a call to the personality function wrapper
(`_Unwind_CallPersonality`) for a catch pad, we store some necessary
info in `__wasm_lpad_context` struct and pass it. One of the info is the
LSDA address for the function. For this, we insert a call to
`wasm.lsda()`, which will be lowered down to the address of LSDA, and
store it in a field in `__wasm_lpad_context`.
There are exceptions to this personality call insertion: catchpads for
`catch (...)` and cleanuppads (for destructors) don't need personality
function calls, because we don't need to figure out whether the current
exception should be caught or not. (They always should.)
There was a little optimization to `wasm.lsda()` call insertion. Because
the LSDA address is the same throughout a function, we don't need to
insert a store of `wasm.lsda()` return value in every catchpad. For
example:
```
try {
foo();
} catch (int) {
// wasm.lsda() call and a store are inserted here, like, in
// pseudocode,
// %lsda = wasm.lsda();
// store %lsda to a field in __wasm_lpad_context
try {
foo();
} catch (int) {
// We don't need to insert the wasm.lsda() and store again, because
// to arrive here, we have already stored the LSDA address to
// __wasm_lpad_context in the outer catch.
}
}
```
So the previous algorithm checked if the current catch has a parent EH
pad, we didn't insert a call to `wasm.lsda()` and its store.
But this was incorrect, because what if the outer catch is `catch (...)`
or a cleanuppad?
```
try {
foo();
} catch (...) {
// wasm.lsda() call and a store are NOT inserted here
try {
foo();
} catch (int) {
// We need wasm.lsda() here!
}
}
```
In this case we need to insert `wasm.lsda()` in the inner catchpad,
because the outer catchpad does not have one.
To minimize the number of inserted `wasm.lsda()` calls and stores, we
need a way to figure out whether we have encountered `wasm.lsda()` call
in any of EH pads that dominates the current EH pad. To figure that
out, we now visit EH pads in BFS order in the dominator tree so that we
visit parent BBs first before visiting its child BBs in the domtree.
We keep a set named `ExecutedLSDA`, which basically means "Do we have
`wasm.lsda()` either in the current EH pad or any of its parent EH
pads in the dominator tree?". This is to prevent scanning the domtree up
to the root in the worst case every time we examine an EH pad: each EH
pad only needs to examine its immediate parent EH pad.
- If any of its parent EH pads in the domtree has `wasm.lsda()`, this
means we don't need `wasm.lsda()` in the current EH pad. We also insert
the current EH pad in `ExecutedLSDA` set.
- If none of its parent EH pad has `wasm.lsda()`
- If the current EH pad is a `catch (...)` or a cleanuppad, done.
- If the current EH pad is neither a `catch (...)` nor a cleanuppad,
add `wasm.lsda()` and the store in the current EH pad, and add the
current EH pad to `ExecutedLSDA` set.
Reviewers: dschuff
Subscribers: sbc100, jgravelle-google, hiraditya, sunfish, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77423
Similar to the lowerV16I8Shuffle implementation, for binary compaction v8i16 shuffles we can avoid the PUNPCKLDQ(PSHUFB,PSHUFB) pattern on SSE41+ targets by using PACKUSDW and PBLENDW. Before SSE41 we would need to use PACKSSDW but that requires sign extension that seems to destroy any gains, even on targets without PSHUFB.
This is a bigger gain on AMD than Intel targets but should never be a regression, and avoiding the shuffle mask load(s) is always useful.
Noticed in codegen while dealing with PR31443.
Since D73835 we no longer need to define the whole IRBuilder
implementation in the header. This patch moves some of the larger
methods out of line, into the C++ file.
Differential Revision: https://reviews.llvm.org/D77332
Forward declare DemandedBits in IVDescriptors, and move include
into the cpp file. Also drop the include from LoopUtils, which
does not need it at all.
After finding all such gadgets in a given function, the pass minimally inserts
LFENCE instructions in such a manner that the following property is satisfied:
for all SOURCE+SINK pairs, all paths in the CFG from SOURCE to SINK contain at
least one LFENCE instruction. The algorithm that implements this minimal
insertion is influenced by an academic paper that minimally inserts memory
fences for high-performance concurrent programs:
http://www.cs.ucr.edu/~lesani/companion/oopsla15/OOPSLA15.pdf
The algorithm implemented in this pass is as follows:
1. Build a condensed CFG (i.e., a GadgetGraph) consisting only of the following components:
-SOURCE instructions (also includes function arguments)
-SINK instructions
-Basic block entry points
-Basic block terminators
-LFENCE instructions
2. Analyze the GadgetGraph to determine which SOURCE+SINK pairs (i.e., gadgets) are already mitigated by existing LFENCEs. If all gadgets have been mitigated, go to step 6.
3. Use a heuristic or plugin to approximate minimal LFENCE insertion.
4. Insert one LFENCE along each CFG edge that was cut in step 3.
5. Go to step 2.
6. If any LFENCEs were inserted, return true from runOnFunction() to tell LLVM that the function was modified.
By default, the heuristic used in Step 3 is a greedy heuristic that avoids
inserting LFENCEs into loops unless absolutely necessary. There is also a
CLI option to load a plugin that can provide even better optimization,
inserting fewer fences, while still mitigating all of the LVI gadgets.
The plugin can be found here: https://github.com/intel/lvi-llvm-optimization-plugin,
and a description of the pass's behavior with the plugin can be found here:
https://software.intel.com/security-software-guidance/insights/optimized-mitigation-approach-load-value-injection.
Differential Revision: https://reviews.llvm.org/D75937
Adds a new data structure, ImmutableGraph, and uses RDF to find LVI gadgets and add them to a MachineGadgetGraph.
More specifically, a new X86 machine pass finds Load Value Injection (LVI) gadgets consisting of a load from memory (i.e., SOURCE), and any operation that may transmit the value loaded from memory over a covert channel, or use the value loaded from memory to determine a branch/call target (i.e., SINK).
Also adds a new target feature to X86: +lvi-load-hardening
The feature can be added via the clang CLI using -mlvi-hardening.
Differential Revision: https://reviews.llvm.org/D75936
Summary:
This patch includes two extensions:
1. It extends the GraphDiff to also keep the original list of updates
after legalization, not just the deletes/insert vectors.
It also provides an API to pop the first update (the updates are store
in reverse, such that the first update is at the end of the list)
2. It adds a bool to mark whether the given updates should be applied as
given, or applied in reverse. This moves the task of reversing the
updates (when the caller needs this) to a functionality inside
GraphDiff, versus having the caller do this.
The two changes could be split into two patches, but they seemed
reasonably small to be reviewed together.
Reviewers: kuhar, dblaikie
Subscribers: hiraditya, george.burgess.iv, mgrang, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77167
Adding a pass that replaces every ret instruction with the sequence:
pop <scratch-reg>
lfence
jmp *<scratch-reg>
where <scratch-reg> is some available scratch register, according to the
calling convention of the function being mitigated.
Differential Revision: https://reviews.llvm.org/D75935
Summary:
Remove usages of asserting vector getters in Type in preparation for the
VectorType refactor. The existence of these functions complicates the
refactor while adding little value.
Reviewers: kparzysz, sdesmalen, efriedma
Reviewed By: kparzysz
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77267
We can fix register class of PHI based on its all AGPR uses.
That leaves behind all PHIs which were already processed
earlier. Propagate RC back to PHI operands of a PHI.
Differential Revision: https://reviews.llvm.org/D77344
As detailed on PR45043, static analysis was warning that the StringRef::iterator Position argument was being ignored and the function was hardwired to use the Current iterator.
This patch ensures we use the provided iterator and removes the (barely necessary) setError wrapper that always used Current.
Differential Revision: https://reviews.llvm.org/D76512
Extracting to the same index that we are going to insert back into
allows forming select ("blend") shuffles and enables further transforms.
Admittedly, this is a quick-fix for a more general problem that I'm
hoping to solve by adding transforms for patterns that start with an
insertelement.
But this might resolve some regressions known to be caused by the
extract-extract transform (although I have not gotten more details on
those yet).
In the motivating case from PR34724:
https://bugs.llvm.org/show_bug.cgi?id=34724
The combination of subsequent instcombine and codegen transforms gets us this improvement:
vmovshdup %xmm0, %xmm2 ## xmm2 = xmm0[1,1,3,3]
vhaddps %xmm1, %xmm1, %xmm4
vmovshdup %xmm1, %xmm3 ## xmm3 = xmm1[1,1,3,3]
vaddps %xmm0, %xmm2, %xmm0
vaddps %xmm1, %xmm3, %xmm1
vshufps $200, %xmm4, %xmm0, %xmm0 ## xmm0 = xmm0[0,2],xmm4[0,3]
vinsertps $177, %xmm1, %xmm0, %xmm0 ## xmm0 = zero,xmm0[1,2],xmm1[2]
-->
vmovshdup %xmm0, %xmm2 ## xmm2 = xmm0[1,1,3,3]
vhaddps %xmm1, %xmm1, %xmm1
vaddps %xmm0, %xmm2, %xmm0
vshufps $200, %xmm1, %xmm0, %xmm0 ## xmm0 = xmm0[0,2],xmm1[0,3]
Differential Revision: https://reviews.llvm.org/D76623
Added unit tests for 2 scenarios that were failing.
Made replace_path_prefix back to 3 parameters instead of 5, simplifying the implementation. The other 2 were always used with the default value.
This commit is intended to be the first of 3:
1) simplify/fix replace_path_prefix.
2) use it in the context of -fdebug-prefix-map and -fmacro-prefix-map (see D76869).
3) Make Windows version of replace_path_prefix insensitive to both case and separators (slash vs backslash).
Differential Revision: https://reviews.llvm.org/D77223
Extend lowerShuffleWithPACK/matchShuffleWithPACK/createPackShuffleMask to handle compaction style shuffle masks that can be lowered to chains of PACKSS/PACKUS if their inputs are suitably sign/zero extended.
This helps avoid PSHUFB (and its mask load) for short shuffle chains, shuffle combining will still replace with a PSHUFB if we have enough shuffles as getFauxShuffleMask should recognise the PACKSS/PACKUS chains.
As discussed in post-commit review in https://reviews.llvm.org/D73501
if the goal of this is to help vectorizer, then we should actually
be teaching vectorizer to do this, because right now this rewrite
is still budget-limited, which isn't what we'd want.
Additionally, while the rest of the patch series was universally profitable,
this particular patch is reportedly (https://reviews.llvm.org/D73501#1905171)
exposing cost-modeling issues on ARM.
So let's just back this particular patch out. Once there's an undo transform,
this could be considered for reintegration.
This reverts commit 44edc6fd2c.
Currently when the target is big-endian vmov.i64 reverses the order of the two
words of the vector. This is correct only when the underlying element type is
32-bit, as actually what it should be doing is considering it a vector of the
underlying type and reversing the elements of that.
Differential Revision: https://reviews.llvm.org/D76515
If we have an element-wise vmov immediate instruction then a subsequent vrev
with width greater or equal to the vmov element width, then that vrev won't do
anything. Add a DAG combine to convert bitcasts that would become such vrevs
into vector_reg_casts instead.
Differential Revision: https://reviews.llvm.org/D76514
Summary:
For current architect, we always require setContainingCsect to be
called on every MCSymbol got used in XCOFF context.
This is very hard to achieve because symbols gets created everywhere
and other MCSymbol types(ELF, COFF) do not have similar rules.
It's very easy to miss setting the containing csect, and we would
need to add a lot of XCOFF specialized code around some common code area.
This patch intendeds to do
1. Rely on getFragment().getParent() to get csect from labels.
2. Only use get/setRepresentedCsect (was get/setContainingCsect)
if symbol itself represents a csect.
Reviewers: DiggerLin, hubert.reinterpretcast, daltenty
Differential Revision: https://reviews.llvm.org/D77080
Summary: I think it would be better to require the alignment to be >= 1. It is currently confusing to allow both values.
Reviewers: courbet
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77372
This pass replaces each indirect call/jump with a direct call to a thunk that looks like:
lfence
jmpq *%r11
This ensures that if the value in register %r11 was loaded from memory, then
the value in %r11 is (architecturally) correct prior to the jump.
Also adds a new target feature to X86: +lvi-cfi
("cfi" meaning control-flow integrity)
The feature can be added via clang CLI using -mlvi-cfi.
This is an alternate implementation to https://reviews.llvm.org/D75934 That merges the thunk insertion functionality with the existing X86 retpoline code.
Differential Revision: https://reviews.llvm.org/D76812
The old name was a bit misleading because the functions actually return
contributions to the corresponding sections.
Differential revision: https://reviews.llvm.org/D77302
Summary:
This patch adds parsing and dumping DWARFv5 .debug_macro section in llvm-dwarfdump,
it does not introduce any new switch. Existing switch "--debug-macro"
should be used to dump macinfo or macro section.
Reviewed By: dblaikie, ikudrin, jhenderson
Differential Revision: https://reviews.llvm.org/D73086
isGCValue should detect whether the deopt value is a GC pointer.
Currently it checks by finding the value in SI.Bases and SI.Ptrs.
However these data structures contain only those values which
have corresponding gc.relocate call. So we can miss GC value if it
does not have gc.relocate call (dead after the call).
Check GC strategy whether pointer is GC one or consider any pointer
to be GC one conservatively.
Reviewers: reames, dantrushin
Reviewed By: reames
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D77130
Summary:
A recent change in the instruction simplifier enables a call to a function that just returns one of its parameter to be simplified as simply loading the parameter. This exposes a bug in the inliner where double inlining may be involved which in turn may cause compiler ICE when an already-inlined callsite is reused for further inlining.
To put it simply, in the following-like C program, when the function call second(t) is inlined, its code t = third(t) will be reduced to just loading the return value of the callsite first(). This causes the inliner internal data structure to register the first() callsite for the call edge representing the third() call, therefore incurs a double inlining when both call edges are considered an inline candidate. I'm making a fix to break the inliner from reusing a callsite for new call edges.
```
void top()
{
int t = first();
second(t);
}
void second(int t)
{
t = third(t);
fourth(t);
}
void third(int t)
{
return t;
}
```
The actual failing case is much trickier than the example here and is only reproducible with the legacy inliner. The way the legacy inliner works is to process each SCC in a bottom-up order. That means in reality function first may be already inlined into top, or function third is either inlined to second or is folded into nothing. To repro the failure seen from building a large application, we need to figure out a way to confuse the inliner so that the bottom-up inlining is not fulfilled. I'm doing this by making the second call indirect so that the alias analyzer fails to figure out the right call graph edge from top to second and top can be processed before second during the bottom-up. We also need to tweak the test code so that when the inlining of top happens, the function body of second is not that optimized, by delaying the pass of function attribute deducer (i.e, which tells function third has no side effect and just returns its parameter). Since the CGSCC pass is iterative, additional calls are added to top to postpone the inlining of second to the second round right after the first function attribute deducing pass is done. I haven't been able to repro the failure with the new pass manager since the processing order of ininlined callsites is a bit different, but in theory the issue could happen there too.
Note that this fix could introduce a side effect that blocks the simplification of inlined code, specifically for a call site that can be folded to another call site. I hope this can probably be complemented by subsequent inlining or folding, as shown in the attached unit test. The ideal fix should be to separate the use of VMap. However, in reality this failing pattern shouldn't happen often. And even if it happens, there should be a good chance that the non-folded call site will be refolded by iterative inlining or subsequent simplification.
Reviewers: wenlei, davidxl, tejohnson
Reviewed By: wenlei, davidxl
Subscribers: eraman, nikic, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76248
Summary:
This patch comes from H.J.'s 2bd54ce7fa
**This patch fix the failed llvm unit tests which running on CET machine. **(e.g. ExecutionEngine/MCJIT/MCJITTests)
The reason we enable IBT at "JIT compiled with CET" is mainly that: the JIT don't know the its caller program is CET enable or not.
If JIT's caller program is non-CET, it is no problem JIT generate CET code or not.
But if JIT's caller program is CET enabled, JIT must generate CET code or it will cause Control protection exceptions.
I have test the patch at llvm-unit-test and llvm-test-suite at CET machine. It passed.
and H.J. also test it at building and running VNCserver(Virtual Network Console), it works too.
(if not apply this patch, VNCserver will crash at CET machine.)
Reviewers: hjl.tools, craig.topper, LuoYuanke, annita.zhang, pengfei
Subscribers: tstellar, efriedma, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76900
This was causing a machine verifier failure on the test suite.
Make sure that we don't end up with a weird register class here.
Failure for reference:
*** Bad machine code: Illegal virtual register for instruction ***
- function: check_constrain
- basic block: %bb.1 (0x7f8b70839f80)
- instruction: early-clobber %6:gpr64, early-clobber %7:gpr64sp =
JumpTableDest32 %5:gpr64, %1:gpr64sp, %jump-table.0
- operand 3: %1:gpr64sp
Expected a GPR64 register, but got a GPR64sp register
Differential Revision: https://reviews.llvm.org/D77349
MI peephole will remove unnecessary FRSP instructions. This patch
removes such unnecessary XSRSP.
Reviewed By: steven.zhang
Differential Revision: https://reviews.llvm.org/D77208
Summary:
This fixes a few issues related to SMRD offsets. On gfx9 and gfx10 we have a
signed byte offset immediate, however we can overflow into a negative since we
treat it as unsigned.
Also, the SMRD SOFFSET sgpr is an unsigned offset on all subtargets. We
sometimes tried to use negative values here.
Third, S_BUFFER instructions should never use a signed offset immediate.
Differential Revision: https://reviews.llvm.org/D77082
This will likely introduce catastrophic performance regressions on
older subtargets, but should be correct. A follow up change will
remove the old fp32-denormals subtarget features, and switch to using
the new denormal-fp-math/denormal-fp-math-f32 attributes. Frontends
should be making sure to add the denormal-fp-math-f32 attribute when
appropriate to avoid performance regressions.
Summary:
[llvm][TextAPI] adding inlining reexported libraries support
* this patch adds reader/writer support for MachO tbd files.
The usecase is to represent reexported libraries in top level library
that won't need to exist for linker indirection because all of the
needed content will be inlined in the same document.
Reviewers: ributzka, steven_wu, jhenderson
Reviewed By: ributzka
Subscribers: JDevlieghere, hiraditya, mgrang, dexonsmith, rupprecht, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67646
Consider a callee function that has a call (C) within it which feeds
into the return. When we inline that callee into a callsite that has
return attributes, we can backward propagate valid attributes to the
call (C) within that inlined callee body.
This is safe to do so only if we can guarantee transfer of execution to
successor in the window of instructions between return value (i.e. the
call C) and the return instruction.
Also, this is valid only for attributes which are a property of a
callsite and not those that are not dependent on the ABI, or a property
of the call itself.
Reviewed-By: reames, jdoerfert
Differential Revision: https://reviews.llvm.org/D76140
This is a workaround for clang adding noinline to all functions at
-O0. Previously, we would just add alwaysinline, and the verifier
would complain about having both noinline and alwaysinline. We
currently can't truly codegen this case as a freestanding function, so
override the user forcing noinline.
bitcast (shuf V, MaskC) --> shuf (bitcast V), MaskC'
We do not attempt this in InstCombine because we do not want to change
types and create new shuffle ops that are potentially not lowered as
well as the original code. Here, we can check the cost model to see if
it is worthwhile.
I've aggressively enabled this transform even if the types are the same
size and/or equal cost because moving the bitcast allows InstCombine to
make further simplifications.
In the motivating cases from PR35454:
https://bugs.llvm.org/show_bug.cgi?id=35454
...this is enough to let instcombine and the backend eliminate the
redundant shuffles, but we probably want to extend VectorCombine to
handle the inverse pattern (shuffle-of-bitcast) to get that
simplification directly in IR.
Differential Revision: https://reviews.llvm.org/D76727
SILoadStoreOptimizer::checkAndPrepareMerge() expects base and
paired instruction to come in order and scans MBB from base to
the paired instruction. An original order can be changed if
there were a dependent instruction in between and base instruction
was moved.
Fixed by bailing the optimization. In theory it might be possible
still to perform a merge by swapping instructions, but on practice
it bails anyway because it finds dependency on that same instruction
which has resulted in the base move.
Differential Revision: https://reviews.llvm.org/D77245
These are versions of a function that regressed with:
rGf2fbdf76d8d0
That particular problem occurs with an instcombine-simplifycfg-instcombine
sequence, but we can show that it exists within instcombine only with
other variations of the pattern.
This does not change much in code generation, but in rare cases MachineCSE
can figure out that an instruction is redundant after commuting it.
Review: Ulrich Weigand
This reverts commit f2fbdf76d8.
As noted in the post-commit thread:
https://reviews.llvm.org/rGf2fbdf76d8d0
...this can obscure a min/max pattern where the components
have extra uses. We can show that the problem is independent
of this change with a slightly modified source example, so
this revert just delays/reduces the need to fix the real
problem.
We need to improve our analysis of negation or -- more
generally -- subtraction using patches like D77230 or D68408.
Summary:
Splitting Knowledge retention into Queries in Analysis and Builder into Transform/Utils
allows Queries and Transform/Utils to use Analysis.
Reviewers: jdoerfert, sstefan1
Reviewed By: jdoerfert
Subscribers: mgorny, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77171
This patch adds
- New arguments to getMinPrefetchStride() to let the target decide on a
per-loop basis if software prefetching should be done even with a stride
within the limit of the hw prefetcher.
- New TTI hook enableWritePrefetching() to let a target do write prefetching
by default (defaults to false).
- In LoopDataPrefetch:
- A search through the whole loop to gather information before emitting any
prefetches. This way the target can get information via new arguments to
getMinPrefetchStride() and emit prefetches more selectively. Collected
information includes: Does the loop have a call, how many memory
accesses, how many of them are strided, how many prefetches will cover
them. This is NFC to before as long as the target does not change its
definition of getMinPrefetchStride().
- If a previous access to the same exact address was 'read', and the
current one is 'write', make it a 'write' prefetch.
- If two accesses that are covered by the same prefetch do not dominate
each other, put the prefetch in a block that dominates both of them.
- If a ConstantMaxTripCount is less than ItersAhead, then skip the loop.
- A SystemZ implementation of getMinPrefetchStride().
Review: Ulrich Weigand, Michael Kruse
Differential Revision: https://reviews.llvm.org/D70228
This adds MVE vmull patterns, which are conceptually the same as
mul(vmovl, vmovl), and so the tablegen patterns follow the same
structure.
For i8 and i16 this is simple enough, but in the i32 version the
multiply (in 64bits) is illegal, meaning we need to catch the pattern
earlier in a dag fold. Because bitcasts are involved in the zext
versions and the patterns are a little different in little and big
endian. I have only added little endian support in this patch.
Differential Revision: https://reviews.llvm.org/D76740
The unpredictable/hasSideEffects flag is usually inferred by tablegen
from whether the instruction has a tablegen pattern (and that pattern
only has a single output instruction). Now that the MVE intrinsics are
all committed and producing code, the remaining instructions still
marked as unpredictable need to be specially handled. This adds the flag
directly to instructions that need it, notably the V*MLAL instructions
and some of the MOV's.
Differential Revision: https://reviews.llvm.org/D76910
Summary:
This allows doing `memcmp(p, q, 7)` with 2 loads instead of a call to
memcmp.
This fixes part of PR45147.
Reviewers: spatel
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76133
As pointed out by @thakis, currently CallSiteSplitting bails out after
checking the first PHI node. We should check all PHI nodes, until we
find one where call site splitting is beneficial.
This patch also slightly simplifies the code using BasicBlock::phis().
Reviewers: davidxl, junbuml, thakis
Reviewed By: davidxl
Differential Revision: https://reviews.llvm.org/D77089
When compiling AMDGPUDisassembler.cpp in a stage 1 trunk build with
CMAKE_BUILD_TYPE=RelWithDebInfo LLVM_USE_SANITIZER=Address LiveDebugVariables
accounts for 21.5% wall clock time. This fix reduces that to 1.2% by switching
out a linked list lookup with a map lookup.
Note that the linked list is still used to group UserValues by vreg. The vreg
lookups don't cause any problems in this pathological case.
This is the same idea as D68816, which was reverted, except that it is a less
intrusive fix.
Reviewed By: vsk
Differential Revision: https://reviews.llvm.org/D77226
Different file formats have different naming style for the debug
sections. The method is implemented for ELF, COFF and Mach-O formats.
Differential Revision: https://reviews.llvm.org/D76276
This is a cleanup and normalization patch that also enables reuse with
Flang later on. A follow up will clean up and move the directive ->
clauses mapping.
Differential Revision: https://reviews.llvm.org/D77112
It was added by D76591 for migration purposes (not all
printBranchOperand users have migrated to the overload with `uint64_t Address`).
Now that all have been migrated, the parameter can go away.
There was a TODO in genericValueTraversal to provide the context
instruction and due to the lack of it users that wanted one just used
something available. Unfortunately, using a fixed instruction is wrong
in the presence of PHIs so we need to update the context instruction
properly.
Reviewed By: uenoku
Differential Revision: https://reviews.llvm.org/D76870
This cannot be triggered right now, as far as I know, but it doesn't
make sense to deduce a constant range on arguments of declarations.
Exposed during testing of AAValueSimplify extensions.
Use DL & ABI information for better alignment deduction, e.g., if a type
is accessed and the ABI specifies an alignment requirement for such an
access we can use it. This is based on a patch by @lebedev.ri and
inspired by getBaseAlign in Loads.cpp.
Depends on D76673.
Reviewed By: lebedev.ri
Differential Revision: https://reviews.llvm.org/D76674
If we have a must-tail call the callee and caller need to have matching
ABIs. Part of that is alignment which we might modify when we deduce
alignment of arguments of either. Since we would need to keep them in
sync, which is not as simple, we simply avoid deducing alignment for
arguments of the must-tail caller or callee.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D76673
Failure to export __cxa_atexit can lead to an attempt to import a definition
from the process itself (if __cxa_atexit is referenced from another JITDylib),
but the process definition will clash with the existing non-exported definition
to produce an unexpected DuplicateDefinitionError.
This patch fixes the immediate issue by exporting __cxa_atexit. It also fixes a
bug where atexit functions in other JITDylibs were not being run by adding a
copy of run_atexits_helper to every JITDylib.
A follow up patch will deal with the bug where definition generators are called
despite a non-exported definition being present.
We create a lot of AbstractAttributes and they live as long as
the Attributor does. It seems reasonable to allocate them via a
BumpPtrAllocator owned by the Attributor.
Reviewed By: baziotis
Differential Revision: https://reviews.llvm.org/D76589
This means the linker will be expect them be undefined at link time an
will generate imports from the `env` module rather than reporting
undefined externals.
Differential Revision: https://reviews.llvm.org/D77192
The legalizer has a tendency to lose DebugLoc's when expanding or
combining instructions. The verifier that detected these isn't ready
for upstreaming yet but this patch fixes the cases that came up when
applying it to our out-of-tree backend's CodeGen tests.
This pattern comes up a few more times in this file and probably in
the backends too but I'd prefer to fix the others separately (and
preferably when the lost-location verifier detects them).
The MemoryBuffer::getMemBuffer method's RequiresNullTerminator parameter
defaults to true, but object files are not null terminated so we need to
explicitly pass false here.
Negation is equivalent to bitwise-not + 1, so try to convert more
subtracts into adds using this relationship:
0 - (A ^ C) => ((A ^ C) ^ -1) + 1 => A ^ ~C + 1
I doubt this will recover the regression noted in rGf2fbdf76d8d0,
but seems like we're going to need to improve here and/or revive D68408?
Alive2 proofs:
http://volta.cs.utah.edu:8080/z/Re5tMUhttp://volta.cs.utah.edu:8080/z/An-uns
Differential Revision: https://reviews.llvm.org/D77230
find() was altering the UserChain, even in cases where it subsequently
discovered that the resulting constant was a 0. This confuses
rebuildWithoutConstOffset() when it attempts to walk the chain later, since it
is expected that the chain itself be a path down the use-def edges of an
expression.
Make the attributor pass aware of aligned_alloc for converting heap
allocations to stack ones.
Depends on D76971.
Differential Revision: https://reviews.llvm.org/D76974
Summary:
The previous code for determining the innermost region in CFGSort was
not correct. We determine subregion relationship by domination of their
headers, i.e., if region A's header dominates region B's header, B is a
subregion of A. Previously we assumed that if a BB belongs to both a
loop and an exception, the region with fewer number of BBs is the
innermost one. This may not be true, because while WebAssemblyException
contains BBs in all its subregions (loops or exceptions), MachineLoop
may not, because MachineLoop does not contain BBs that don't have a path
to its header even if they are dominated by its header.
Loop header <---|
| |
Exception header |
| \ |
A B |
| \ |
| C |
| |
Loop latch |
| |
-------------|
For example, in this CFG, the loop does not contain B and C, because
they don't have a path back to the loops header. But for CFGSort we
consider the exception here belongs to the loop and the exception should
be a subregion of the loop and scheduled together.
So here we should use `WE->contains(ML->getHeader())` (but not
`ML->contains(WE->getHeader())`, for the stated region above).
This also fixes some comments and deletes `Regions` vector in
`RegionInfo` class, which was not used anywere.
Reviewers: dschuff
Subscribers: sbc100, jgravelle-google, hiraditya, sunfish, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77181
Summary:
Currently, the comparison argument used for ATOMIC_CMP_XCHG is legalised
with GetPromotedInteger, which leaves the upper bits of the value
undefind. Since this is used for comparing in an LR/SC loop with a
full-width comparison, we must sign extend it. We introduce a new
getExtendForAtomicCmpSwapArg to complement getExtendForAtomicOps, since
many targets have compare-and-swap instructions (or pseudos) that
correctly handle an any-extend input, and the existing function
determines the extension of the result, whereas we are concerned with
the input.
This is related to https://reviews.llvm.org/D58829, which solved the
issue for ATOMIC_CMP_SWAP_WITH_SUCCESS, but not the simpler
ATOMIC_CMP_SWAP.
Reviewers: asb, lenary, efriedma
Reviewed By: asb
Subscribers: arichardson, hiraditya, rbar, johnrusso, simoncook, sabuasal, niosHD, kito-cheng, shiva0217, MaskRay, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, rkruppe, jfb, PkmX, jocewei, psnobl, benna, Jim, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, evandro, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D74453
First step toward making use of canLowerByDroppingEvenElements to match chains of PACKSS/PACKUS for compaction shuffles.
At the moment we still only match a single stage but the MatchPACK is now more general.
This allows the MVE VPT Block insertion pass to remove VPNOTs in
order to create more complex VPT blocks such as TE, TEET, TETE, etc.
Differential Revision: https://reviews.llvm.org/D75993
VPTMaskValue was using the "instruction" encoding to represent the masks
(= the same encoding as the one used by the instructions in an object file),
but it is only used to build MCOperands, so it should use the MCOperand
encoding of the masks, which is slightly different.
Differential Revision: https://reviews.llvm.org/D76139
Summary:
Aggregate types containing scalable vectors aren't supported and as far
as I can tell this pass is mostly concerned with optimisations on
aggregate types, so the majority of this pass isn't very useful for
scalable vectors.
This patch modifies SROA such that mem2reg is run on allocas with
scalable types that are promotable, but nothing else such as slicing is
done.
The use of TypeSize in this pass has also been updated to be explicitly
fixed size. When invoking the following methods in DataLayout:
* getTypeSizeInBits
* getTypeStoreSize
* getTypeStoreSizeInBits
* getTypeAllocSize
we now called getFixedSize on the resultant TypeSize. This is quite an
extensive change with around 50 calls to these functions, and also the
first change of this kind (being explicit about fixed vs scalable
size) as far as I'm aware, so feedback welcome.
A test is included containing IR with scalable vectors that this pass is
able to optimise.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D76720
PTEST/TESTP sets EFLAGS as:
TESTZ: ZF = (Op0 & Op1) == 0
TESTC: CF = (~Op0 & Op1) == 0
TESTNZC: ZF == 0 && CF == 0
If we are inverting the 0'th operand of a PTEST/TESTP instruction we can adjust the comparisons to correct handle the inversion implicitly.
Additionally, for "TESTZ" (ZF) cases, the allones case, PTEST(X,-1) can be simplified to PTEST(X,X).
We can expand this for the TESTZ(X,~Y) pattern and also handle KTEST/KORTEST in the future.
Differential Revision: https://reviews.llvm.org/D76984
Summary:
Make sure we do not assert on value types not being
simple in getFauxShuffleMask when analysing operations
such as "v8i16 = truncate v8i24".
Reviewers: RKSimon
Reviewed By: RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77136
Currently ConstantRange::binaryAnd/binaryOr results are too pessimistic
for single element constant ranges.
If both operands are single element ranges, we can use APInt's AND and
OR implementations directly.
Note that some other binary operations on constant ranges can cover the
single element cases naturally, but for OR and AND this unfortunately is
not the case.
Reviewers: nikic, spatel, lebedev.ri
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D76446
Currently, DAG combiner uses (fmul (rsqrt x) x) to estimate square
root of x. However, this method would return NaN if x is +Inf, which
is incorrect.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D76853
This patch adds checks to the verifier to ensure the dimension arguments
passed to the matrix intrinsics match the vector types for their
arugments/return values.
Reviewers: anemet, Gerolf, andrew.w.kaylor, LuoYuanke
Reviewed By: anemet
Differential Revision: https://reviews.llvm.org/D77129
getCallCost is only used within the different layers of TTI, with no
backend implementing it so fold the base implementation into
getUserCost. I think this is an NFC.
Differential Revision: https://reviews.llvm.org/D77050
Summary: These were templated due to SelectionDAG using int masks for shuffles and IR using unsigned masks for shuffles. But now that D72467 has landed we have an int mask version of IRBuilder::CreateShuffleVector. So just use int instead of a template
Reviewers: spatel, efriedma, RKSimon
Reviewed By: efriedma
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D77183
Also add lowerShuffleWithPACK call to lowerV32I16Shuffle - shuffle combining was catching it but we avoid a lot of temporary shuffle creations if we catch it at lowering first.
For the downstream RISCV maintenance, it would be easier to inherent
RISCVISelDAGToDAG by including header and only override the method that needs
to be customized for the provider non-standard ISA extension without touching
RISCVISelDAGToDAG.cpp which may cause conflict when upgrading the downstream
LLVM version.
Differential Revision: https://reviews.llvm.org/D77117
Summary:
In https://bugs.llvm.org/show_bug.cgi?id=45297, it fails selecting
instructions for `PPCISD::ST_VSR_SCAL_INT`. The reason it generate the
`PPCISD::ST_VSR_SCAL_INT` with `-power8-vector` in IR is PPC's
combiner checks `hasP8Altivec` rather than `hasP8Vector`. This patch
should resolve PR45297.
Differential Revision: https://reviews.llvm.org/D76773
Summary:
Select folding in JumpThreading can create a conditional branch on a
code patch that did not have one in the original program. This is not a
valid transformation in sanitize_memory functions.
Note that JumpThreading does select folding in 3 different places. Two
of them seem safe - they apply to a select instruction in a BB that ends
with an unconditional branch to another BB, which (in turn) ends with a
conditional branch or a switch with the same condition.
Fixes PR45220.
Reviewers: glider, dvyukov, efriedma
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76332
Summary:
When we encounter an XCOFF file, reflect that in the triple information.
In addition to knowing the object file format, we know that the
associated OS is AIX.
This means that we can expect that there is no output difference in the
processing of an XCOFF32 input file between cases where the triple is
left unspecified by the user and cases where the user specifies
`--triple powerpc-ibm-aix` explicitly.
Reviewers: jhenderson, sfertile, jasonliu, daltenty
Reviewed By: jasonliu
Subscribers: wuzish, nemanjai, hiraditya, MaskRay, rupprecht, steven.zhang, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77025
The UnwindHelp object is used during exception handling by runtime
code. It must be findable from a fixed offset from FP.
This change allocates the UnwindHelp object as a fixed object (as is
done for x86_64) to ensure that both the generated code and runtime
agree on the location of the object.
Fixes https://bugs.llvm.org/show_bug.cgi?id=45346
Differential Revision: https://reviews.llvm.org/D77016
The generated code for a funclet can have an add to sp in the epilogue
for which there is no corresponding sub in the prologue.
This patch removes the early return from emitPrologue that was
preventing the sub to sp, and instead conditionalizes the appropriate
parts of the rest of the function.
Fixes https://bugs.llvm.org/show_bug.cgi?id=45345
Differential Revision: https://reviews.llvm.org/D77015
This reverts commit 28518d9ae3.
There is a failure in MsgPackReader.cpp when built with clang. It
complains about "signext and zeroext" are incompatible. Investigating
offline if it is infact a UB in the MsgPackReader code.
Instead, represent the mask as out-of-line data in the instruction. This
should be more efficient in the places that currently use
getShuffleVector(), and paves the way for further changes to add new
shuffles for scalable vectors.
This doesn't change the syntax in textual IR. And I don't currently plan
to change the bitcode encoding in this patch, although we'll probably
need to do something once we extend shufflevector for scalable types.
I expect that once this is finished, we can then replace the raw "mask"
with something more appropriate for scalable vectors. Not sure exactly
what this looks like at the moment, but there are a few different ways
we could handle it. Maybe we could try to describe specific shuffles.
Or maybe we could define it in terms of a function to convert a fixed-length
array into an appropriate scalable vector, using a "step", or something
like that.
Differential Revision: https://reviews.llvm.org/D72467
The attached test case is simplified from tcmalloc. Both function calls should be optimized as tailcall. But llvm can only optimize the first call. The second call can't be optimized because function dupRetToEnableTailCallOpts failed to duplicate ret into block case2.
There 2 problems blocked the duplication:
1 Intrinsic call llvm.assume is not handled by dupRetToEnableTailCallOpts.
2 The control flow is more complex than expected, dupRetToEnableTailCallOpts can only duplicate ret into its predecessor, but here we have an intermediate block between call and ret.
The solutions:
1 Since CodeGenPrepare is already at the end of LLVM IR phase, we can simply delete the intrinsic call to llvm.assume.
2 A general solution to the complex control flow is hard, but for this case, after exit2 is duplicated into case1, exit2 is the only successor of exit1 and exit1 is the only predecessor of exit2, so they can be combined through eliminateFallThrough. But this function is called too late, there is no more dupRetToEnableTailCallOpts after it. We can add an earlier call to eliminateFallThrough to solve it.
Differential Revision: https://reviews.llvm.org/D76539
We have loads preserving low and high 16 bits of their
destinations. However, we always use a whole 32 bit register
for these. The same happens with 16 bit stores, we have to
use full 32 bit register so if high bits are clobbered the
register needs to be copied. One example of such code is
added to the load-hi16.ll.
The proper solution to the problem is to define 16 bit subregs
and use them in the operations which do not read another half
of a VGPR or preserve it if the VGPR is written.
This patch simply defines subregisters and register classes.
At the moment there should be no difference in code generation.
A lot more work is needed to actually use these new register
classes. Therefore, there are no new tests at this time.
Register weight calculation has changed with new subregs so
appropriate changes were made to keep all calculations just
as they are now, especially calculations of register pressure.
Differential Revision: https://reviews.llvm.org/D74873
Consider a callee function that has a call (C) within it which feeds
into the return. When we inline that callee into a callsite that has
return attributes, we can backward propagate those attributes to the
call (C) within that inlined callee body.
This is safe to do so only if we can guarantee transfer of execution to
successor in the window of instructions between return value (i.e. the
call C) and the return instruction.
See added test cases.
Reviewed-By: reames, jdoerfert
Differential Revision: https://reviews.llvm.org/D76140
Registers used in any address (as well as in a few other contexts)
have special semantics when a "zero" register is used, which is
why the back-end defines extra register classes ADDR32, ADDR64 etc
to be used to prevent the register allocator from using %r0 there.
However, when writing assembler code "by hand", you sometimes need
to trigger that special semantics. However, currently the AsmParser
will reject %r0 in those places. In some cases it may be possible
to write that instruction differently - but in others it is currently
not possible at all.
This check in AsmParser simply seems overly strict, so this patch
just removes the check completely. This brings the behaviour of
AsmParser in line with the GNU assembler as well.
Bugzilla: https://bugs.llvm.org/show_bug.cgi?id=45092
Make InstCombine aware of the aligned_alloc library function.
Signed-off-by: Uday Bondhugula <uday@polymagelabs.com>
Depends on D76970.
Differential Revision: https://reviews.llvm.org/D76971
Summary:
CGProfilePass is run by default in certain new pass manager optimization pipeline. Assemblers other than llvm as (such as gnu as) cannot recognize the .cgprofile entries generated and emitted from this pass, causing build time error.
This patch adds new options in clang CodeGenOpts and PassBuilder options so that we can turn cgprofile off when not using integrated assembler.
Reviewers: Bigcheese, xur, george.burgess.iv, chandlerc, manojgupta
Reviewed By: manojgupta
Subscribers: manojgupta, void, hiraditya, dexonsmith, llvm-commits, tcwang, llozano
Tags: #llvm, #clang
Differential Revision: https://reviews.llvm.org/D62627
Summary: this patch preserve information from various places in EarlyCSE into assume bundles.
Reviewers: jdoerfert
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76769
--no-threads is a name copied from gold.
gold has --no-thread, --thread-count and several other --thread-count-*.
There are needs to customize the number of threads (running several lld
processes concurrently or customizing the number of LTO threads).
Having a single --threads=N is a straightforward replacement of gold's
--no-threads + --thread-count.
--no-threads is used rarely. So just delete --no-threads instead of
keeping it for compatibility for a while.
If --threads= is specified (ELF,wasm; COFF /threads: is similar),
--thinlto-jobs= defaults to --threads=,
otherwise all available hardware threads are used.
There is currently no way to override a --threads={1,2,...}. It is still
a debate whether we should use --threads=all.
Reviewed By: rnk, aganea
Differential Revision: https://reviews.llvm.org/D76885
If canLowerByDroppingEvenElements indicates that the shuffle is a N:1 compaction pattern and the inputs are suitably sign/zero extended then we can use a chain of PACKSS/PACKUS to compact.
This helps avoid PSHUFB (and its mask load) for short shuffle chains, shuffle combining will still replace with a PSHUFB if we have enough shuffles as getFauxShuffleMask can recognise PACKSS/PACKUS chains.
This patch updates ValueLattice to distinguish between ranges that are
guaranteed to not include undef and ranges that may include undef.
A constant range guaranteed to not contain undef can be used to simplify
instructions to arbitrary values. A constant range that may contain
undef can only be used to simplify to a constant. If the value can be
undef, it might take a value outside the range. For example, consider
the snipped below
define i32 @f(i32 %a, i1 %c) {
br i1 %c, label %true, label %false
true:
%a.255 = and i32 %a, 255
br label %exit
false:
br label %exit
exit:
%p = phi i32 [ %a.255, %true ], [ undef, %false ]
%f.1 = icmp eq i32 %p, 300
call void @use(i1 %f.1)
%res = and i32 %p, 255
ret i32 %res
}
In the exit block, %p would be a constant range [0, 256) including undef as
%p could be undef. We can use the range information to replace %f.1 with
false because we remove the compare, effectively forcing the use of the
constant to be != 300. We cannot replace %res with %p however, because
if %a would be undef %cond may be true but the second use might not be
< 256.
Currently LazyValueInfo uses the new behavior just when simplifying AND
instructions and does not distinguish between constant ranges with and
without undef otherwise. I think we should address the remaining issues
in LVI incrementally.
Reviewers: efriedma, reames, aqjune, jdoerfert, sstefan1
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D76931
For the PHI node
%1 = phi [%A, %entry], [%X, %latch]
it is incorrect to use SCEV of backedge val %X as an exit value
of PHI unless %X is loop invariant.
This is because exit value of %1 is value of %X at one-before-last
iteration of the loop.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D73181
Canonicalize the case when a scalar extracted from a vector is
truncated. Transform such cases to bitcast-then-extractelement.
This will enable erasing the truncate operation.
This commit fixes PR45314.
reviewers: spatel
Differential revision: https://reviews.llvm.org/D76983
Summary:
In method SelectionDAGBuilder::LowerStatepoint, array SI.GCTransitionArgs
is initialized from wrong part of ImmutableStatepoint class.
We copy gc args instead of transitions args.
Reviewers: reames, skatkov
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77075
Add a new llvm.amdgcn.ballot intrinsic modeled on the ballot function
in GLSL and other shader languages. It returns a bitfield containing the
result of its boolean argument in all active lanes, and zero in all
inactive lanes.
This is intended to replace the existing llvm.amdgcn.icmp and
llvm.amdgcn.fcmp intrinsics after a suitable transition period.
Use the new intrinsic in the atomic optimizer pass.
Differential Revision: https://reviews.llvm.org/D65088
For casts with constant range operands, we can use
ConstantRange::castOp.
Reviewers: davide, efriedma, mssimpso
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D71938
Leverage ARM ELF build attribute section to create ELF attribute section
for RISC-V. Extract the common part of parsing logic for this section
into ELFAttributeParser.[cpp|h] and ELFAttributes.[cpp|h].
Differential Revision: https://reviews.llvm.org/D74023
Octeon branches (bbit0/bbit032/bbit1/bbit132) have an immediate operand,
so it is legal to have such replacement within
MipsBranchExpansion::replaceBranch().
According to the specification, a branch (e.g. bbit0 ) looks like:
bbit0 rs p offset // p is an immediate operand
if !rs<p> then branch
Without this patch, an assertion triggers in the method,
and the problem has been found in the real example.
Differential Revision: https://reviews.llvm.org/D76842
In the past, AVR functions were only lowered with interrupt-specific
machine code if the function was defined with the "avr-interrupt" or
"avr-signal" calling conventions.
This patch modifies the backend so that if the function does not have a
special calling convention, but does have an "interrupt" attribute,
that function is interpreted as a function with interrupts.
This also extracts the "is this function an interrupt" logic from
several disparate places in the backend into one AVRMachineFunctionInfo
attribute.
Bug found by Wilhelm Meier.
Compbinary format uses MD5 to represent strings in name table. That gives smaller profile without the need of compression/decompression when writing/reading the profile. The patch adds the support in extbinary format. It is off by default but user can choose to enable it.
Note the feature of using MD5 in name table can bring very small chance of name conflict leading to profile mismatch. Besides, profile using the feature won't have the profile remapping support.
Differential Revision: https://reviews.llvm.org/D76255
When we have
```
a = G_OR x, x
```
or
```
b = G_AND y, y
```
We can drop the G_OR/G_AND and just use x/y respectively.
Also update arm64-fallback.ll because there was an or in there which hits this
transformation.
Differential Revision: https://reviews.llvm.org/D77105
Implement identity combines for operations like the following:
```
%a = G_SUB %b, 0
```
This can just be replaced with %b.
Over CTMark, this gives some minor size improvements at -O3.
Differential Revision: https://reviews.llvm.org/D76640
This reverts commit b3297ef051.
This change is incorrect. The current semantic of null in the IR is a
pointer with the bitvalue 0. It is not a cast from an integer 0, so
this should preserve the pointer type.
Summary:
This code was throwing away the opcode for a boolean, which was then
reconstructing the opcode from that boolean. Just pass the opcode, and
forget the boolean.
Reviewers: srhines
Reviewed By: srhines
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77100
We currently don't have a way to map to the equivalent intrinsic
opcode, so track immediate 0s in place of the address for the
selection to know to change the final opcode.
Extend the FileCollector's API with addDirectory which adds a directory
and its contents to the VFS mapping.
Differential revision: https://reviews.llvm.org/D76671
InstCombine has a mess of logic that tries to preserve min/max patterns,
but AFAICT, this one is not necessary because we can always narrow the
corresponding select in this sequence to match the narrow compare.
The biggest danger for this patch is inducing infinite looping or
assert from exceeding max iterations. If any bots hit that in the
vicinity of this commit, this is the likely patch to blame.
Optimize the common case of splat vector constant. For large vector
going through all elements is expensive. For splatr/broadcast cases we
can skip going through all elements.
Differential Revision: https://reviews.llvm.org/D76664
Summary:
This change adds amdgcn.reloc.constant intrinsic to the amdgpu backend, which will compile into a relocation entry in the resulting elf.
The intrinsics takes a MetadataNode (String) as its only argument, which specifies the symbol name of the relocation entry.
`SelectionDAGBuilder::getValueImpl` is changed to allow metadata operands passed through to ISel.
Author: csyonghe <yonghe@google.com>
Reviewers: tpr, nhaehnle
Reviewed By: nhaehnle
Subscribers: arsenm, kzhuravl, jvesely, wdng, yaxunl, dstuttard, t-tye, hiraditya, kerbowa, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76440
For each natural loop with multiple exit blocks, this pass creates a
new block N such that all exiting blocks now branch to N, and then
control flow is redistributed to all the original exit blocks.
The bulk of the tranformation is a new function introduced in
BasicBlockUtils that an redirect control flow from a set of incoming
blocks to a set of outgoing blocks via a common "hub".
This is a useful workaround for a limitation in the structurizer which
incorrectly orders blocks when processing a nest of loops. This pass
bypasses that issue by ensuring that each natural loop is recognized
as a separate region. Since the structurizer is a region pass, it no
longer sees a nest of loops in a single region, and instead processes
each "level" in the nesting as a separate region.
The AMDGPU backend provides a new option to enable this pass before
the structurizer, which may eventually be enabled by default.
Reviewers: madhur13490, arsenm, nhaehnle
Reviewed By: nhaehnle
Differential Revision: https://reviews.llvm.org/D75865
In InnerLoopVectorizer::getOrCreateTripCount, when the backedge taken
count is a SCEV add expression, its type is defined by the type of the
last operand of the add expression.
In the test case from PR45259, this last operand happens to be a
pointer, which (according to llvm::Type) does not have a primitive size
in bits. In this case, LoopVectorize fails to truncate the SCEV and
crashes as a result.
Uing ScalarEvolution::getTypeSizeInBits makes the truncation work as expected.
https://bugs.llvm.org/show_bug.cgi?id=45259
Differential Revision: https://reviews.llvm.org/D76669
Summary:
Otherwise PostRA list scheduler may reorder instruction, such as
schedule this
'''
pushq $0x8
pop %rbx
lea 0x2a0(%rsp),%r15
'''
to
'''
pushq $0x8
lea 0x2a0(%rsp),%r15
pop %rbx
'''
by mistake. The patch is to prevent this to happen by making sure POP has
implicit use of SP.
Reviewers: craig.topper
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77031
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, hiraditya, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, sabuasal, niosHD, jrtc27, MaskRay, zzheng, edward-jones, atanasyan, rogfer01, MartinMosbeck, brucehoult, the_o, jfb, PkmX, jocewei, Jim, lenary, s.egerton, pzheng, sameer.abuasal, apazos, luismarques, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77059
I still think the call lowering type legalization logic split between
the generic code and target is too confusing, but largely induced by
the reliance on the DAG infrastructure.
If we are lowering to X86ISD::SHUF128 we are going to lose track of individual 128-bit lanes that are UNDEF, so if we can widen these to guarantee that they are sequential with their neighbour we should. This helps with later shuffle combines.
This fixes
warning: loop variable 'Def' of type 'const llvm::Register' creates a copy from type 'const llvm::Register' [-Wrange-loop-analysis]
llvm::Register just contains a single unsigned and should be copied.
Reviewers: rampitec
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D77011
Add a bit more logic into the 'FalseLaneZeros' tracking to enable
horizontal reductions and also make the VADDV variants
validForTailPredication.
Differential Revision: https://reviews.llvm.org/D76708
Summary:
Also deprecate getOriginalAlignment, getAlignment will take much more time as it is pervasive through the codebase (including TableGened files).
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: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D76933
In the original batch of MVE VMOVimm code generation VMOV.i64 was left
out due to the way it was done downstream. It turns out that it's fairly
simple though. This adds the codegen for it, similar to NEON.
Bigendian is technically incorrect in this version, which John is fixing
in a Neon patch.
In past, isGuaranteedToTransferExecutionToSuccessor contained some weird logic
for volatile loads/stores that was ultimately removed by patch D65375. It's time to
remove a piece of dependent logic that used to be a workaround for the code which
is now deleted.
Reviewed By: uenoku
Differential Revision: https://reviews.llvm.org/D76918
As we don't return the result of replaceInstUsesWith(), we are
responsible for erasing the instruction.
There is a small subtlety here in that we need to do this after
the other uses of Builder, which uses the original multiply as
the insertion point.
NFC apart from worklist order changes.
Aligned_alloc is a standard lib function and has been in glibc since
2.16 and in the C11 standard. It has semantics similar to malloc/calloc
for several analyses/transforms. This patch introduces aligned_alloc
in target library info and memory builtins. Subsequent ones will
make other passes aware and fix https://bugs.llvm.org/show_bug.cgi?id=44062
This change will also be useful to LLVM generators that need to allocate
buffers of vector elements larger than 16 bytes (for eg. 256-bit ones),
element boundary alignment for which is not typically provided by glibc malloc.
Signed-off-by: Uday Bondhugula <uday@polymagelabs.com>
Differential Revision: https://reviews.llvm.org/D76970
Rather than converting to a dummy select with equal true and false
ops, just directly return the resulting value.
As a side-effect, this fixes missing DCE of the previously replaced
operand.
BitVector::set_bits() returns an iterator range yielding unsinged
elements, which always will be copied while const & gives the impression
that there will be no copy. Newer version of clang complain:
warning: loop variable 'SetBitsIt' is always a copy because the range of type 'iterator_range<llvm::BitVector::const_set_bits_iterator>' (aka 'iterator_range<const_set_bits_iterator_impl<llvm::BitVector> >') does not return a reference [-Wrange-loop-analysis]
Reviewers: jdoerfert, rnk
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D77010
Usually when we replaceInstUsesWith() we also return the original
instruction, and InstCombine will take care of erasing it. Here
we don't do that, so we need to manually erase it.
NFC apart from worklist order changes.
As explained on PR40720, EXTRACTF128 is always as good/better than VPERM2F128, and we can use the implicit zeroing of the upper half.
I've added some extra tests to vector-shuffle-combining-avx2.ll to make sure we don't lose coverage.
This was an inner helper function for the real matchShuffleAsByteRotate function, but it is more generic and is used directly for VALIGN lowering which doesn't work at the byte level.
This patch changes VPWidenRecipe to only store a single original IR
instruction. This is the first required step towards modeling it's
operands as VPValues and also towards breaking it up into a
VPInstruction.
Discussed as part of D74695.
Reviewers: Ayal, gilr, rengolin
Reviewed By: gilr
Differential Revision: https://reviews.llvm.org/D76988
Simon Pilgrim