David Tenty [Fri, 17 Apr 2020 17:06:08 +0000 (13:06 -0400)]
[AIX] Return the correct set of callee saved regs
Summary:
r13 isn't reserved on 32-bit AIX, which is reflected in our calling
convention but not callee saved regs.
Reviewers: sfertile, ZarkoCA, cebowleratibm, jasonliu
Reviewed By: sfertile
Subscribers: lei, wuzish, nemanjai, hiraditya, kbarton, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77101
Alexander Belyaev [Mon, 20 Apr 2020 13:16:52 +0000 (15:16 +0200)]
[MLIR] Add AtomicRMWRegionOp.
https://llvm.discourse.group/t/rfc-add-std-atomic-rmw-op/489
Differensial revision: https://reviews.llvm.org/D78352
Petre-Ionut Tudor [Mon, 20 Apr 2020 14:43:56 +0000 (15:43 +0100)]
Revert "[ARM] Fix conditions for lowering to S[LR]I"
This reverts commit
cabfcf840a9d15d92466c6774953d3aa399cde92.
The patch introduced another bug in the optimization.
Simon Pilgrim [Mon, 20 Apr 2020 14:59:24 +0000 (15:59 +0100)]
BranchFolding.h - cleanup includes and forward declarations. NFC.
Push MBFIWrapper.h include down to BranchFolding.cpp/IfConversion.cpp
Simon Pilgrim [Mon, 20 Apr 2020 13:33:21 +0000 (14:33 +0100)]
MIRVRegNamerUtils.h - remove unnecessary includes. NFC.
Replace with forward declarations or push down to MIRVRegNamerUtils.cpp where necessary.
Fangrui Song [Sun, 5 Apr 2020 04:31:36 +0000 (21:31 -0700)]
[ELF] Add --warn-backrefs-exclude=<glob>
D77522 changed --warn-backrefs to not warn for linking sandwich
problems (-ldef1 -lref -ldef2). This removed lots of false positives.
However, glibc still has some problems. libc.a defines some symbols
which are normally in libm.a and libpthread.a, e.g. __isnanl/raise.
For a linking order `-lm -lpthread -lc`, I have seen:
```
// different resolutions: GNU ld/gold select libc.a(s_isnan.o) as the definition
backward reference detected: __isnanl in libc.a(printf_fp.o) refers to libm.a(m_isnanl.o)
// different resolutions: GNU ld/gold select libc.a(raise.o) as the definition
backward reference detected: raise in libc.a(abort.o) refers to libpthread.a(pt-raise.o)
```
To facilitate deployment of --warn-backrefs, add --warn-backrefs-exclude= so that
certain known issues (which may be impractical to fix) can be whitelisted.
Deliberate choices:
* Not a comma-separated list (`--warn-backrefs-exclude=liba.a,libb.a`).
-Wl, splits the argument at commas, so we cannot use commas.
--export-dynamic-symbol is similar.
* Not in the style of `--warn-backrefs='*' --warn-backrefs=-liba.a`.
We just need exclusion, not inclusion. For easier build system
integration, we should avoid order dependency. With the current
scheme, we enable --warn-backrefs, and indivial libraries can add
--warn-backrefs-exclude=<glob> to their LDFLAGS.
Reviewed By: psmith
Differential Revision: https://reviews.llvm.org/D77512
Tom Weaver [Mon, 20 Apr 2020 14:46:55 +0000 (15:46 +0100)]
[Dexter][NFC] Add Debugger Controller To Dexter
Add DebuggerControllerBase and DefaultController to Dexter
implements a new architecture that supports new and novel ways of running
a debugger under dexter.
Current implementation adds the original default behaviour via the new
architecture via the DefaultController, this should have NFC.
Reviewers: Orlando
Differential Revision: https://reviews.llvm.org/D76926
Louis Dionne [Mon, 20 Apr 2020 09:22:20 +0000 (05:22 -0400)]
[libc++] Fix the no-exceptions build of libc++ on Apple
We previously tried re-exporting symbols that didn't exist when
exceptions were disabled. Note that building libc++abi without
exceptions still doesn't work when linking against the default-provided
libSystem.dylib, because it transitively depends on libobjc.dylib,
and that requires __gxx_personality_v0. But building libc++abi
with exceptions and libc++ without exceptions does work.
Emre Kultursay [Mon, 20 Apr 2020 14:17:36 +0000 (16:17 +0200)]
[lldb] Remove m_last_file_sp from SourceManager
Summary:
...and replace it with m_last_file_spec instead.
When Source Cache is enabled, the value stored in m_last_file_sp is
already in the Source Cache, and caching it again in SourceManager
brings no extra benefit. All we need is to "remember" the last used
file, and FileSpec can serve the same purpose.
When Source Cache is disabled, the user explicitly requested no caching
of source files, and therefore, m_last_file_sp should NOT be used.
Bug: llvm.org/PR45310
Depends on D76805.
Reviewers: labath, jingham
Reviewed By: jingham
Subscribers: labath, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D76806
Emre Kultursay [Mon, 20 Apr 2020 14:17:27 +0000 (16:17 +0200)]
[lldb] Fix SourceManager::SourceFileCache insertion
Summary:
Lookup and subsequent insert was done using uninitialized
FileSpec object, which caused the cache to be a no-op.
Bug: llvm.org/PR45310
Depends on D76804.
Reviewers: labath, JDevlieghere
Reviewed By: labath
Subscribers: mgorny, jingham, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D76805
Emre Kultursay [Mon, 20 Apr 2020 14:17:17 +0000 (16:17 +0200)]
[lldb] Add new LLDB setting: use-source-cache
Summary:
LLDB memory-maps large source files, and at the same time, caches
all source files in the Source Cache.
On Windows, memory-mapped source files are not writeable, causing
bad user experience in IDEs (such as errors when saving edited files).
IDEs should have the ability to disable the Source Cache at LLDB
startup, so that users can edit source files while debugging.
Bug: llvm.org/PR45310
Reviewers: labath, JDevlieghere, jingham
Reviewed By: labath
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D76804
Pavel Labath [Mon, 20 Apr 2020 14:07:16 +0000 (16:07 +0200)]
[lldb] Fix gcc warnings in TypeCategory.cpp
The cleanup in
3e3701f8a0bf left these variable unused.
Yuri Per [Mon, 20 Apr 2020 14:03:42 +0000 (16:03 +0200)]
[lldb] Prefer executable files from sysroot over files from local filesystem
Summary:
In D49685 sysroot behaviour was partially fixed. But files from local filesystem with same path still has priority over files from sysroot.
This patch fixes it by removing fallback to local filesystem from RemoteAwarePlatform::GetModuleSpec(). It is not actually required because higher level code do such fallback itself. See, for example, resolver in Platform::GetSharedModule().
Reviewers: labath, clayborg, EugeneBi
Reviewed By: labath
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D77529
Sander de Smalen [Mon, 20 Apr 2020 13:57:13 +0000 (14:57 +0100)]
[SveEmitter] Add immediate checks for lanes and complex imms
Adds another bunch of of intrinsics that take immediates with
varying ranges based, some being a complex rotation immediate
which are a set of allowed immediates rather than a range.
svmla_lane: lane immediate ranging 0..(128/(1*sizeinbits(elt)) - 1)
svcmla_lane: lane immediate ranging 0..(128/(2*sizeinbits(elt)) - 1)
svdot_lane: lane immediate ranging 0..(128/(4*sizeinbits(elt)) - 1)
svcadd: complex rotate immediate [90, 270]
svcmla:
svcmla_lane: complex rotate immediate [0, 90, 180, 270]
Reviewers: efriedma, SjoerdMeijer, rovka
Reviewed By: efriedma
Tags: #clang
Differential Revision: https://reviews.llvm.org/D76680
Bjorn Pettersson [Sun, 19 Apr 2020 10:56:16 +0000 (12:56 +0200)]
[Scalarizer] Fix a non-deterministic scatter order problem
Summary:
The indexing operator in Scatterer may result in building new
instructions. When using multiple such operators in a function
argument list the order in which we build instructions depend on
argument evaluation order (which is undefined in C++).
This patch avoid such problems by expanding the components using
the [] operator prior to the function call.
Problem was seen when comparing output, while builing LLVM with
different compilers (clang vs gcc).
Reviewers: foad, cameron.mcinally, uabelho
Reviewed By: foad
Subscribers: hiraditya, mgrang, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78455
Sander de Smalen [Mon, 20 Apr 2020 13:41:58 +0000 (14:41 +0100)]
[SveEmitter] Add more immediate operand checks.
This patch adds a number of intrinsics that take immediates with
varying ranges based on the element size one of the operands.
svext: immediate ranging 0 to (2048/sizeinbits(elt) - 1)
svasrd: immediate ranging 1..sizeinbits(elt)
svqshlu: immediate ranging 1..sizeinbits(elt)/2
ftmad: immediate ranging 0..(sizeinbits(elt) - 1)
Reviewers: efriedma, SjoerdMeijer, rovka, rengolin
Reviewed By: SjoerdMeijer
Tags: #clang
Differential Revision: https://reviews.llvm.org/D76679
Simon Pilgrim [Mon, 20 Apr 2020 12:55:46 +0000 (13:55 +0100)]
X86MacroFusion.cpp - ensure X86MacroFusion.h module header is included first. NFC.
Michael Forster [Mon, 20 Apr 2020 13:16:11 +0000 (15:16 +0200)]
Explicitly move from llvm::json Array/Object to Value
Summary: The implicit conversion fails under Clang 3.8.
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: jkorous, arphaman, kadircet, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D78487
LLVM GN Syncbot [Mon, 20 Apr 2020 13:10:45 +0000 (13:10 +0000)]
[gn build] Port
12030494fce
Konstantin Schwarz [Wed, 8 Apr 2020 12:40:43 +0000 (14:40 +0200)]
[GlobalISel] Introduce InlineAsmLowering class
Summary:
Similar to the CallLowering class used for lowering LLVM IR calls to MIR calls,
we introduce a separate class for lowering LLVM IR inline asm to MIR INLINEASM.
There is no functional change yet, all existing tests should pass.
Reviewers: arsenm, dsanders, aemerson, volkan, t.p.northover, paquette
Reviewed By: aemerson
Subscribers: gargaroff, wdng, mgorny, rovka, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78316
Peter Smith [Mon, 20 Apr 2020 12:55:01 +0000 (13:55 +0100)]
[LLD][ELF][ARM] Convert ADR/LDR to .inst .reloc
After D78301 MC no longer emits a relocation for this case. Change to use
.inst and .reloc to synthesize the same instruction and relocation.
serge-sans-paille [Mon, 20 Apr 2020 12:43:21 +0000 (14:43 +0200)]
[nfc] Cleanup extension header generation
Artem Dergachev [Mon, 20 Apr 2020 12:30:30 +0000 (15:30 +0300)]
[analyzer] GenericTaint: Don't expect CallEvent to always have a Decl.
This isn't the case when the callee is completely unknown,
eg. when it is a symbolic function pointer.
Mark Murray [Mon, 20 Apr 2020 12:01:36 +0000 (13:01 +0100)]
Revert
3ce0ad1b336e67a76d78ae7ff7d66fe127586620 Die to breakage in check-lld.
Requested-by: Nico Weber
Ayke van Laethem [Wed, 15 Apr 2020 15:30:47 +0000 (17:30 +0200)]
[AVR] Do not place functions in .progmem.data
Previously, the AVR backend would put functions in .progmem.data. This
is probably a regression from when functions still lived in address
space 0. With this change, only global constants are placed in
.progmem.data.
This is not complete: avr-gcc additionally respects -fdata-sections for
progmem global constants, which LLVM doesn't yet do. But fixing that is
a bit more complicated (and I believe other backends such as RISC-V
might also have similar issues).
Differential Revision: https://reviews.llvm.org/D78212
Ayke van Laethem [Sat, 18 Apr 2020 21:42:48 +0000 (23:42 +0200)]
[AVR] Do not use divmod calls for bigger integers
The avr-libc provides *divmodqi4, *divmodhi4, and *divmodsi4 functions,
but does not provide a *divmoddi4. Instead it provides regular *divdi3
and *moddi3 functions.
Note that avr-libc doesn't support *divti3 or *modti3 for 128-bit
integer manipulation.
Source:
https://github.com/gcc-mirror/gcc/blob/releases/gcc-5.4.0/libgcc/config/avr/lib1funcs.S
Differential Revision: https://reviews.llvm.org/D78437
Georgii Rymar [Wed, 15 Apr 2020 12:30:21 +0000 (15:30 +0300)]
[FileCheck] - Refactor the code related to string arrays. NFCI.
There are few `std::vector<std::string>` members in
`FileCheckRequest`. This patch changes these arrays to `std::vector<StringRef>`
and refactors the code related to cleanup/improve/simplify it.
Differential revision: https://reviews.llvm.org/D78202
Georgii Rymar [Tue, 7 Apr 2020 15:20:51 +0000 (18:20 +0300)]
[obj2yaml] - Teach obj2yaml to dump SHT_NOBITS sections when dumping program headers.
SHT_NOBITS are a bit special because occupy no physical space.
This patch adds support for them.
Differential revision: https://reviews.llvm.org/D77805
Georgii Rymar [Wed, 1 Apr 2020 13:40:25 +0000 (16:40 +0300)]
[llvm-readobj] - Fix crashes and misbehaviors when reading strings from broken string tables.
There are cases when we either might print garbage or crash when
reading strings for dumping dynamic tags.
For example when a string table is not null-terminated or goes past the EOF.
This patch fixes issues mentioned.
Differential revision: https://reviews.llvm.org/D77216
Simon Pilgrim [Mon, 20 Apr 2020 11:10:25 +0000 (12:10 +0100)]
X86Subtarget.h - remove unused includes. NFC.
Replace with forward declarations.
Dmitry Preobrazhensky [Mon, 20 Apr 2020 10:53:15 +0000 (13:53 +0300)]
[MC][DWARF] Corrected handling of is_stmt flag in .loc directives
According to DWARF standard, is_stmt is a global flag; when set or cleared it should affect subsequent .loc directives.
However llvm assembler handled is_stmt differently: it forced all locations to have is_stmt=1 unless is_stmt was specified explicitly as 0.
The fix utilizes current DWARF state flags to compute correct is_stmt values.
See https://bugs.llvm.org/show_bug.cgi?id=45529 for a detailed issue description.
Reviewers: arsenm, probinson, enderby
Differential Revision: https://reviews.llvm.org/D78102
Simon Pilgrim [Mon, 20 Apr 2020 10:54:04 +0000 (11:54 +0100)]
X86Subtarget.cpp - sort includes. NFC
Ensure X86Subtarget.h module header is at the top, and sort the remaining includes.
Simon Pilgrim [Mon, 20 Apr 2020 10:39:21 +0000 (11:39 +0100)]
X86MCTargetDesc.h - remove unnecessary MCStreamer.h include. NFC.
We don't need all of MCStreamer.h, just FormattedStream.h. The rest can be replaced with forward declarations.
X86WinAllocaExpander.cpp had an implicit dependency on MapVector.h which I've added locally.
Simon Pilgrim [Mon, 20 Apr 2020 10:22:26 +0000 (11:22 +0100)]
X86MCAsmInfo.h - remove unnecessary MCAsmInfo.h include. NFC.
We only use the COFF/Darwin/ELF classes directly.
Simon Pilgrim [Mon, 20 Apr 2020 10:20:22 +0000 (11:20 +0100)]
X86InstrFoldTables.h - remove unnecessary include. NFC.
We don't need the limits defines, just the sized integer types so use cstdint system header directly.
Kerry McLaughlin [Mon, 20 Apr 2020 10:07:48 +0000 (11:07 +0100)]
[AArch64][SVE] Remove LD1/ST1 dependency on llvm.masked.load/store
Summary:
The SVE masked load and store intrinsics introduced in D76688 rely on
common llvm.masked.load/store nodes. This patch creates new ISD nodes
for LD1(S) & ST1 to remove this dependency.
Additionally, this adds support for sign & zero extending
loads and truncating stores.
Reviewers: sdesmalen, efriedma, cameron.mcinally, c-rhodes, rengolin
Reviewed By: efriedma
Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl, danielkiss, andwar, cfe-commits, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78204
Valeriy Savchenko [Mon, 20 Apr 2020 08:08:57 +0000 (11:08 +0300)]
[analyzer] Do not report CFError null dereference for nonnull params.
We want to trust user type annotations and stop assuming pointers declared
as nonnull still can be null. This functionality is implemented as part
of NonNullParamChecker because it already checks parameter attributes.
Whenever we start analyzing a new function, we assume that all parameters
with 'nonnull' attribute are indeed non-null.
Patch by Valeriy Savchenko!
Differential Revision: https://reviews.llvm.org/D77806
Valeriy Savchenko [Wed, 8 Apr 2020 15:58:40 +0000 (18:58 +0300)]
[analyzer] Do not report NSError null dereference for _Nonnull params.
We want to trust user type annotations and stop assuming pointers declared
as _Nonnull still can be null. This functionality is implemented as part
of NullabilityChecker as it already tracks non-null types.
Patch by Valeriy Savchenko!
Differential Revision: https://reviews.llvm.org/D77722
Florian Hahn [Mon, 20 Apr 2020 08:42:02 +0000 (09:42 +0100)]
[VPlan] Clean up tryToCreate(Widen)Recipe. (NFC)
This patch includes some clean-ups to tryToCreateRecipe, suggested in
D77973.
It includes:
* Renaming tryToCreateRecipe to tryToCreateWidenRecipe.
* Move VPBB insertion logic to caller of tryToCreateWidenRecipe.
* Hoists instruction checks to tryToCreateWidenRecipe, making it
clearer which instructions are handled by which recipe, simplifying
the checks by using early exits.
* Split up handling of induction PHIs and truncates using inductions.
Reviewers: gilr, rengolin, Ayal, hsaito
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D78287
Sam Parker [Wed, 15 Apr 2020 10:10:09 +0000 (11:10 +0100)]
[ARM][MVE] Add patterns for VRHADD
Add patterns which use standard add nodes along with arm vshr imm
nodes.
Differential Revision: https://reviews.llvm.org/D77069
Mark Murray [Thu, 16 Apr 2020 14:17:34 +0000 (15:17 +0100)]
[ARM][MC][Thumb] Revert relocation for some pc-relative fixups.
Summary:
This commit reverts https://reviews.llvm.org/D75039. Concensus appears to
be in favour of assembly-time resolution of these ADR and LDR relocations,
in line with GNU.
Reviewers: psmith
Subscribers: kristof.beyls, hiraditya, danielkiss, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78301
Florian Hahn [Mon, 20 Apr 2020 08:11:32 +0000 (09:11 +0100)]
[PredicateInfo] Use new Instruction::comesBefore instead of OI (NFC).
The recently added Instruction::comesBefore can be used instead of
OrderedInstructions.
Reviewers: rnk, nikic, efriedma
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D78452
Alexander Belyaev [Mon, 20 Apr 2020 08:16:52 +0000 (10:16 +0200)]
[MLIR] Add #include "llvm/ADT/SmallPtrSet.h" back to LoopUtils.h.
Sam Parker [Fri, 17 Apr 2020 12:29:31 +0000 (13:29 +0100)]
[NFC][TTI] Explicit use of VectorType
The API for shuffles and reductions uses generic Type parameters,
instead of VectorType, and so assertions and casts are used a lot.
This patch makes those types explicit, which means that the clients
can't be lazy, but results in less ambiguity, and that can only be a
good thing.
Bugzilla: https://bugs.llvm.org/show_bug.cgi?id=45562
Differential Revision: https://reviews.llvm.org/D78357
Kang Zhang [Mon, 20 Apr 2020 07:27:41 +0000 (07:27 +0000)]
[CodeGen] Support freeze expand for ppc_fp128
Summary:
The patch D29014 has added the new ISD::FREEZE and can deal with the
integer.
The patch D76980 has added SoftenFloatRes_FREEZE for float point.
But we still lack of expand for ppc_fp128, this will cause assertion for
some cases.
This patch is to support freeze expand for ppc_fp128.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D78278
Alexander Belyaev [Mon, 20 Apr 2020 07:17:45 +0000 (09:17 +0200)]
[MLIR] Move `replaceAllUsesExcept` from LoopUtil.h to Value.h.
Differential Revision: https://reviews.llvm.org/D78426
Craig Topper [Mon, 20 Apr 2020 07:05:18 +0000 (00:05 -0700)]
[CallSite removal][SCCP] Use CallBase instead of CallSite. NFC
Differential Revision: https://reviews.llvm.org/D78470
Kadir Cetinkaya [Mon, 20 Apr 2020 06:27:15 +0000 (08:27 +0200)]
[clangd][test] Make sed git bash compliant
Craig Topper [Mon, 20 Apr 2020 05:39:24 +0000 (22:39 -0700)]
[CallSite removal][Analysis] Use CallBase instead of CallSite in SparsePropagation unit test. NFC
Craig Topper [Mon, 20 Apr 2020 04:54:38 +0000 (21:54 -0700)]
[CallSite removal][CalledValuePropagation] Use CallBase instead of CallSite. NFC
Differential Revision: https://reviews.llvm.org/D78467
David Blaikie [Mon, 20 Apr 2020 04:32:33 +0000 (21:32 -0700)]
llvm-dwarfdump: Fix UB (unsequenced writes) introduced in
e0fd87cc64d5
Unsequenced write due to "x &= f()" where 'f()' modifies 'x'.
Detected by the llvm-clang-x86_64-expensive-checks-win buildbot.
Investigated/identified by Galina - thanks!
Max Kazantsev [Mon, 20 Apr 2020 03:26:15 +0000 (10:26 +0700)]
[Test] Fix test failure: platform-dependent printout
Xiang1 Zhang [Mon, 20 Apr 2020 02:17:34 +0000 (10:17 +0800)]
Handle CET for -exception-model sjlj
Summary:
In SjLj exception mode, the old landingpad BB will create a new landingpad BB and use indirect branch jump to the old landingpad BB in lowering.
So we should add 2 endbr for this exception model.
Reviewers: hjl.tools, craig.topper, annita.zhang, LuoYuanke, pengfei, efriedma
Reviewed By: LuoYuanke
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77124
Bill Wendling [Mon, 20 Apr 2020 03:08:30 +0000 (20:08 -0700)]
[Object] Use BFD name for little-endian PowerPC64
Summary:
Little-endian PowerPC object files should report "elf64-powerpcle" instead of
"elf64-powerpc".
Reviewers: jhenderson, MaskRay, espindola, alexshap, rupprecht, #powerpc
Reviewed By: MaskRay
Subscribers: wuzish, emaste, nemanjai, shchenz, steven.zhang, llvm-commits
Tags: #llvm, #powerpc
Differential Revision: https://reviews.llvm.org/D78344
Max Kazantsev [Mon, 20 Apr 2020 02:37:30 +0000 (09:37 +0700)]
[Test] Add a test showing how CFG analyses are invalidated after LV
It demonstrates that, even if LV does no actual vectorization and only
forms LCSSA, CFG analyses get dropped.
Shengchen Kan [Sat, 18 Apr 2020 08:52:50 +0000 (16:52 +0800)]
[X86][MC][NFC] Reduce the parameters of functions in X86MCCodeEmitter(Part III)
Summary:
When we encode an instruction, we need to know the number of bytes being
emitted to determine the fixups in `X86MCCodeEmitter::emitImmediate`.
There are only two callers for `emitImmediate`: `emitMemModRMByte` and
`encodeInstruction`.
Before this patch, we kept track of the current byte being emitted
by passing a reference parameter `CurByte` across all the `emit*`
funtions, which is ugly and unnecessary. For example, we don't have any
fixups when emitting prefixes, so we don't need to track this value.
In this patch, we use `StartByte` to record the initial status of the
streamer, and use `OS.tell()` to get the current status of the streamer
when we need to know the number of bytes being emitted. On one hand,
this eliminates the parameter `CurByte` for most `emit*` functions, on
the other hand, this make things clear: Only pass the parameter when we
really need it.
Reviewers: craig.topper, pengfei, MaskRay
Reviewed By: craig.topper, MaskRay
Subscribers: hiraditya, llvm-commits, annita.zhang
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78419
Craig Topper [Mon, 20 Apr 2020 01:32:20 +0000 (18:32 -0700)]
[CallSite removal][WebAssembly] Replace CallSite with CallBase in WebAssemblyOptimizeReturned.
Differential Revision: https://reviews.llvm.org/D78451
Craig Topper [Mon, 20 Apr 2020 01:32:03 +0000 (18:32 -0700)]
[CallSite removal][Analysis] Replace CallSite with CallBase in MemoryBuiltins. NFC
Differential Revision: https://reviews.llvm.org/D78449
Craig Topper [Mon, 20 Apr 2020 01:31:42 +0000 (18:31 -0700)]
[CallSite removal][Lint] Replace visitCallSite with visitCallBase. NFC
Drop visitCallInst/visitInvokeInst since the generic implementation
will delegate to visitCallBase. They appear to be from a time
before visitCallSite existed in the InstVisitor system.
Differential Revision: https://reviews.llvm.org/D78448
Craig Topper [Sun, 19 Apr 2020 23:13:22 +0000 (16:13 -0700)]
[X86] Add X86ISD nodes for PDEP and PEXT.
This will allow use to add DAG combines for these instructions.
Craig Topper [Sun, 19 Apr 2020 07:41:43 +0000 (00:41 -0700)]
[CallSite removal][AMDGPU] Use CallBase instead of CallSite in AMDGPUFixFunctionBitcasts. NFC
Simon Pilgrim [Sun, 19 Apr 2020 20:24:25 +0000 (21:24 +0100)]
X86CallLowering.h - remove unnecessary ArrayRef.h include. NFC.
Simon Pilgrim [Sat, 18 Apr 2020 17:52:07 +0000 (18:52 +0100)]
SectionMemoryManager.h - remove unnecessary StringRef.h include. NFC.
Nikita Popov [Thu, 9 Apr 2020 19:13:03 +0000 (21:13 +0200)]
[LVI] Use Optional instead of out parameter (NFC)
As suggested on D76788, this switches the LVI implementation to
return Optional<ValueLatticeElement> from various methods, instead
of passing in a ValueLatticeElement reference and returning a boolean.
Differential Revision: https://reviews.llvm.org/D78383
Florian Hahn [Sun, 19 Apr 2020 19:02:45 +0000 (20:02 +0100)]
[TTI] Clean up includes (NFC).
Remove some unnecessary includes, replace some with forward
declarations.
This also exposed a few places that were missing some includes.
Florian Hahn [Sun, 19 Apr 2020 18:39:55 +0000 (19:39 +0100)]
[IVDescriptors] Clean up includes.
Some includes are not required and forward declarations can be used
instead. This also exposed a few places that were not directly including
required files.
Florian Hahn [Sun, 19 Apr 2020 17:17:20 +0000 (18:17 +0100)]
[LoopUtils] Clean up includes, use forward decls if appropriate (NFC).
Most of the includes in LoopUtils.h are not required in the header and
they can be replaced by forward declarations.
Unfortunately includes of TargetTransformInfo.h and IVDescriptors.h pull
in a bunch of additional things, but there is no easy way to get rid of
them at the moment I think.
Mehdi Amini [Sun, 19 Apr 2020 17:03:01 +0000 (17:03 +0000)]
Fix one more link for a Rationale doc moved under Rationale/
Simon Pilgrim [Sun, 19 Apr 2020 16:52:35 +0000 (17:52 +0100)]
DependenceGraphBuilder.h - remove unused includes. NFC.
Replace with forward declarations.
Mehdi Amini [Fri, 17 Apr 2020 19:23:57 +0000 (19:23 +0000)]
Fix documentation link to MlirSpirvAbi
Differential Revision: https://reviews.llvm.org/D78394
Mehdi Amini [Sun, 19 Apr 2020 16:52:05 +0000 (16:52 +0000)]
Fix one more doc links after moving the document under Tutorials
Mehdi Amini [Sun, 19 Apr 2020 16:49:30 +0000 (16:49 +0000)]
Fix more broken doc links after some moved under the Rationale category
Simon Pilgrim [Sun, 19 Apr 2020 15:52:26 +0000 (16:52 +0100)]
X86MachineFunctionInfo.h - remove unused include. NFC.
Simon Pilgrim [Sun, 19 Apr 2020 13:44:19 +0000 (14:44 +0100)]
X86InstrInfo.h - remove unused forward declarations. NFC.
Simon Pilgrim [Sun, 19 Apr 2020 13:41:25 +0000 (14:41 +0100)]
X86DisassemblerDecoder.h - remove unused forward declaration. NFC.
Fangrui Song [Sun, 19 Apr 2020 15:46:47 +0000 (08:46 -0700)]
[CMake] Delete HAVE_SCHED_GETAFFINITY and HAVE_CPU_COUNT
sched_getaffinity (Linux specific) has been available
* in glibc since 2002-08-08 (commit
972e719e8154eec5f543b027e2a08dfa285d55d5)
* in musl since the initial check-in.
Sanjay Patel [Sun, 19 Apr 2020 14:42:08 +0000 (10:42 -0400)]
[InstSimplify] add tests for logic+icmp folds for nullptr; NFC
See discussion in D78430.
Sanjay Patel [Sun, 19 Apr 2020 14:19:37 +0000 (10:19 -0400)]
[x86] avoid build warning for enum mismatch; NFC
gcc may warn here because X86ISD::NodeType is specified as "unsigned",
but ISD::NodeType is a naked C enum (although passed as an "unsigned"
throughout SDAG).
Florian Hahn [Sun, 19 Apr 2020 14:11:25 +0000 (15:11 +0100)]
[LAA] Remove unnecessary includes (NFC).
Simon Pilgrim [Sun, 19 Apr 2020 13:29:35 +0000 (14:29 +0100)]
X86MCTargetDesc.h - remove unnecessary includes and forward declarations. NFC.
Simon Pilgrim [Sun, 19 Apr 2020 13:28:52 +0000 (14:28 +0100)]
X86.h - remove unused forward declarations. NFC.
Simon Pilgrim [Sun, 19 Apr 2020 13:20:53 +0000 (14:20 +0100)]
X86SelectionDAGInfo.h - remove unnecessary includes and forward declarations. NFC.
Simon Pilgrim [Sun, 19 Apr 2020 13:03:29 +0000 (14:03 +0100)]
X86TargetTransformInfo.h - remove unnecessary includes. NFC.
David Green [Sun, 19 Apr 2020 12:21:01 +0000 (13:21 +0100)]
[ARM] Regenerate tests. NFC
Simon Pilgrim [Sun, 19 Apr 2020 12:34:58 +0000 (13:34 +0100)]
[X86][SSE] getFauxShuffle - don't combine shuffles with small truncated scalars (PR45604)
getFauxShuffle attempts to combine INSERT_VECTOR_ELT(TRUNCATE/EXTEND(EXTRACT_VECTOR_ELT(x))) patterns into a target shuffle chain.
PR45604 identified an issue where the scalar was truncated to a size smaller than the destination vector element and then zero extended back, which requires the upper bits to be zero'd which we don't currently do.
To avoid the bug I've added an early out in these truncation cases, a future commit should allow us to handle this by inserting the necessary SM_SentinelZero padding.
Sam McCall [Sun, 19 Apr 2020 12:33:00 +0000 (14:33 +0200)]
[clangd] Add index export to dexp
Summary: Add a command to dexp that exports index data in chosen format (e.g. YAML).
Reviewers: sammccall
Subscribers: kbobyrev, mgorny, ilya-biryukov, MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits
Tags: #clang-tools-extra, #clang
Differential Revision: https://reviews.llvm.org/D77385
Sanjay Patel [Sun, 19 Apr 2020 12:32:02 +0000 (08:32 -0400)]
[x86] use vector instructions to lower more FP->int->FP casts
This is an enhancement to D77895 to avoid another
round-trip from XMM->GPR->XMM. This time we handle
the case of starting/ending with an f64 and casting
to signed i32 as the intermediate value.
It's a bit more involved than I initially assumed
because we need to use target-specific opcodes to
represent the non-standard cast ops.
Differential Revision: https://reviews.llvm.org/D78362
Sam McCall [Sun, 19 Apr 2020 12:22:23 +0000 (14:22 +0200)]
[clangd] Extend YAML Serialization
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: mgorny, ilya-biryukov, MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits
Tags: #clang-tools-extra, #clang
Differential Revision: https://reviews.llvm.org/D77938
Sanjay Patel [Sun, 19 Apr 2020 12:06:17 +0000 (08:06 -0400)]
[VectorCombine] transform bitcasted shuffle to wider elements
bitcast (shuf V, MaskC) --> shuf (bitcast V), MaskC'
This is the widen shuffle elements enhancement to D76727.
It builds on the analysis and simplifications in
D77881 and rG6a7e958a423e.
The phase ordering tests show that we can simplify inverse
shuffles across a binop in both directions (widen/narrow or
narrow/widen) now.
There's another potential transform visible in some of the
remaining TODOs - move a bitcasted operand of a shuffle
after the shuffle.
Differential Revision: https://reviews.llvm.org/D78371
Sanjay Patel [Fri, 17 Apr 2020 18:05:04 +0000 (14:05 -0400)]
[InstSimplify] add tests for logic-of-icmp with min/max constant; NFC
See PR45510:
https://bugs.llvm.org/show_bug.cgi?id=45510
We had partial coverage for some of these patterns, so removing duplicate tests
with the complete set in the new test file.
Simon Pilgrim [Sun, 19 Apr 2020 12:13:54 +0000 (13:13 +0100)]
[X86][SSE] Add test case for PR45604
Simon Pilgrim [Sun, 19 Apr 2020 11:38:41 +0000 (12:38 +0100)]
SelectionDAGBuilder.h - remove unused includes + forward declarations. NFC.
Replace SelectionDAG.h include with SelectionDAG forward declaration.
Simon Pilgrim [Sun, 19 Apr 2020 11:17:56 +0000 (12:17 +0100)]
X86InstrFMA3Info.h - remove unnecessary includes. NFC.
There were a number of cpp files explicitly relying on X86InstrFMA3Info.h to include the X86.h header - so I've had to add it locally.
Simon Pilgrim [Sun, 19 Apr 2020 10:52:31 +0000 (11:52 +0100)]
InstrEmitter.h - reduce SelectionDAG.h include to SelectionDAGNodes.h include.
Add SDDbgLabel/TargetLowering forward declarations.
Add the full SelectionDAG.h include to InstrEmitter.cpp.
Simon Pilgrim [Sun, 19 Apr 2020 10:38:50 +0000 (11:38 +0100)]
X86AsmPrinter.h - cleanup includes and forward declarations. NFC.
Reduce X86Subtarget.h/MCCodeEmitter.h/TargetMachine.h includes to forward declarations
Add explicit X86Subtarget.h/TargetMachine.h includes to X86AsmPrinter.cpp/X86MCInstLower.cpp
Remove unused MCSymbol forward declaration
Simon Pilgrim [Sun, 19 Apr 2020 10:13:26 +0000 (11:13 +0100)]
DebugHandlerBase.h - reduce MachineInstr.h include to DebugLoc.h include.
We were only including MachineInstr.h for DebugLoc.h. This exposes an implicit include dependency in BTFDebug.h where I've had to add the MachineInstr.h include.
Simon Pilgrim [Sun, 19 Apr 2020 10:00:30 +0000 (11:00 +0100)]
BuildLibCalls.h - remove unnecessary TargetLibraryInfo forward declaration. NFC
We already have to include the TargetLibraryInfo.h header.
Simon Pilgrim [Sun, 19 Apr 2020 09:55:23 +0000 (10:55 +0100)]
TypeBasedAliasAnalysis.h - replace InstrTypes.h include with forward declaration. NFC.
Benjamin Kramer [Sun, 19 Apr 2020 09:48:28 +0000 (11:48 +0200)]
Remove remaining callers of CreateShuffleVector with unsigned indices and mark it as deprecated
No functionality change intended.
Simon Pilgrim [Sun, 19 Apr 2020 09:29:38 +0000 (10:29 +0100)]
OMPConstants.h - replace StringRef.h include with forward declaration. NFC.