Philip Reames [Wed, 12 Jun 2019 19:04:51 +0000 (19:04 +0000)]
[Tests] Add tests to highlight sibling loop optimization order issue for exit rewriting
The issue addressed in r363180 is more broadly relevant. For the moment, we don't actually get any of these cases because we a) restrict SCEV formation due to SCEExpander needing to preserve LCSSA, and b) don't iterate between loops.
Jordan Rupprecht [Wed, 12 Jun 2019 18:41:27 +0000 (18:41 +0000)]
[llvm-ar][test] Relax lit directory assumptions in thin-archive.test
Summary: thin-archive.test assumes the Output/<testname> structure that lit creates. Rewrite the test in a way that still tests the same thing (creating via relative path and adding via absolute path) but doesn't assume this specific lit structure, making it possible to run in a lit emulator.
Philip Reames [Wed, 12 Jun 2019 17:21:47 +0000 (17:21 +0000)]
[SCEV] Teach computeSCEVAtScope benefit from one-input Phi. PR39673
SCEV does not propagate arguments through one-input Phis so as to make it easy for the SCEV expander (and related code) to preserve LCSSA. It's not entirely clear this restriction is neccessary, but for the moment it exists. For this reason, we don't analyze single-entry phi inputs. However it is possible that when an this input leaves the loop through LCSSA Phi, it is a provable constant. Missing that results in an order of optimization issue in loop exit value rewriting where we miss some oppurtunities based on order in which we visit sibling loops.
This patch teaches computeSCEVAtScope about this case. We can generalize it later, but so far we can only replace LCSSA Phis with their constant loop-exiting values. We should probably also add similiar logic directly in the SCEV construction path itself.
Simon Pilgrim [Wed, 12 Jun 2019 17:14:03 +0000 (17:14 +0000)]
[TargetLowering] Add MachineMemOperand::Flags to allowsMemoryAccess tests (PR42123)
As discussed on D62910, we need to check whether particular types of memory access are allowed, not just their alignment/address-space.
This NFC patch adds a MachineMemOperand::Flags argument to allowsMemoryAccess and allowsMisalignedMemoryAccesses, and wires up calls to pass the relevant flags to them.
If people are happy with this approach I can then update X86TargetLowering::allowsMisalignedMemoryAccesses to handle misaligned NT load/stores.
Matt Arsenault [Wed, 12 Jun 2019 14:23:33 +0000 (14:23 +0000)]
StackProtector: Use PointerMayBeCaptured
This was using its own, outdated list of possible captures. This was
at minimum not catching cmpxchg and addrspacecast captures.
One change is now any volatile access is treated as capturing. The
test coverage for this pass is quite inadequate, but this required
removing volatile in the lifetime capture test.
Also fixes some infrastructure issues to allow running just the IR
pass.
Mikael Holmen [Wed, 12 Jun 2019 14:19:22 +0000 (14:19 +0000)]
[ARM] Fix compiler warning
Without this fix clang 3.6 complains with:
../lib/Target/ARM/ARMAsmPrinter.cpp:1473:18: error: variable 'BranchTarget' is used uninitialized whenever 'if' condition is false [-Werror,-Wsometimes-uninitialized]
} else if (MI->getOperand(1).isSymbol()) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~
../lib/Target/ARM/ARMAsmPrinter.cpp:1479:22: note: uninitialized use occurs here
MCInst.addExpr(BranchTarget);
^~~~~~~~~~~~
../lib/Target/ARM/ARMAsmPrinter.cpp:1473:14: note: remove the 'if' if its condition is always true
} else if (MI->getOperand(1).isSymbol()) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../lib/Target/ARM/ARMAsmPrinter.cpp:1465:33: note: initialize the variable 'BranchTarget' to silence this warning
const MCExpr *BranchTarget;
^
= nullptr
1 error generated.
Matt Arsenault [Wed, 12 Jun 2019 14:05:58 +0000 (14:05 +0000)]
LoopVersioning: Respect convergent
This changes the standalone pass only. Arguably the utility class
itself should assert there are no convergent calls. However, a target
pass with additional context may still be able to version a loop if
all of the dynamic conditions are sufficiently uniform.
Matt Arsenault [Wed, 12 Jun 2019 13:34:19 +0000 (13:34 +0000)]
LoopDistribute/LAA: Respect convergent
This case is slightly tricky, because loop distribution should be
allowed in some cases, and not others. As long as runtime dependency
checks don't need to be introduced, this should be OK. This is further
complicated by the fact that LoopDistribute partially ignores if LAA
says that vectorization is safe, and then does its own runtime pointer
legality checks.
Note this pass still does not handle noduplicate correctly, as this
should always be forbidden with it. I'm not going to bother trying to
fix it, as it would require more effort and I think noduplicate should
be removed.
Nico Weber [Wed, 12 Jun 2019 12:46:46 +0000 (12:46 +0000)]
Fix a Wunused-lambda-capture warning.
The capture was added in the first commit of https://reviews.llvm.org/D61934
when it was used. In the reland, the use was removed but the capture
wasn't removed.
Sam Parker [Wed, 12 Jun 2019 12:00:42 +0000 (12:00 +0000)]
[ARM] Implement TTI::isHardwareLoopProfitable
Implement the backend target hook to drive the HardwareLoops pass.
The low-overhead branch extension for Arm M-class cores is flexible
enough that we don't have to ensure correctness at this point, except
checking that the loop counter variable can be stored in LR - a
32-bit register. For it to be profitable, we want to avoid loops that
contain function calls, or any other instruction that alters the PC.
This implementation uses TargetLoweringInfo, to query type and
operation actions, looks at intrinsic calls and also performs some
manual checks for remainder/division and FP operations.
I think this should be a good base to start and extra details can be
filled out later.
Sam Parker [Wed, 12 Jun 2019 11:53:55 +0000 (11:53 +0000)]
[NFC][SCEV] Add NoWrapFlag argument to InsertBinOp
'Use wrap flags in InsertBinop' (rL362687) was reverted due to
miscompiles. This patch introduces the previous change to pass
no-wrap flags but now only FlagAnyWrap is passed.
Nico Weber [Wed, 12 Jun 2019 11:32:43 +0000 (11:32 +0000)]
Share /machine: handling code with llvm-cvtres too
r363016 let lld-link and llvm-lib share the /machine: parsing code.
This lets llvm-cvtres share it as well.
Making llvm-cvtres depend on llvm-lib seemed a bit strange (it doesn't
need llvm-lib's dependencies on BinaryFormat and BitReader) and I
couldn't find a good place to put this code. Since it's just a few
lines, put it in lib/Object for now.
Xing GUO [Wed, 12 Jun 2019 11:24:22 +0000 (11:24 +0000)]
[DOC] Fix `load` instructions' syntax, function definition.
Summary: In this patch, I updated `load` instruction syntax and fixed function definition. Besides, I re-named some variables to make them obey SSA rule.
Simon Pilgrim [Wed, 12 Jun 2019 11:08:29 +0000 (11:08 +0000)]
[XCore] CombineSTORE - Use allowsMemoryAccess wrapper. NFCI.
Noticed in D63075 - there was a allowsMisalignedMemoryAccesses call to check for unaligned loads and a check for aligned legal type loads - which is exactly what allowsMemoryAccess does.
Ben Dunbobbin [Wed, 12 Jun 2019 11:07:56 +0000 (11:07 +0000)]
[ThinLTO]LTO]Legacy] Fix dependent libraries support by adding querying of the IRSymtab
Dependent libraries support for the legacy api was committed in a
broken state (see: https://reviews.llvm.org/D60274). This was missed
due to the painful nature of having to integrate the changes into a
linker in order to test. This change implements support for dependent
libraries in the legacy LTO api:
- I have removed the current api function, which returns a single
string, and added functions to access each dependent library
specifier individually.
- To reduce the testing pain, I have made the api functions as thin as
possible to maximize coverage from llvm-lto.
- When doing ThinLTO the system linker will load the modules lazily
when scanning the input files. Unfortunately, when modules are
lazily loaded there is no access to module level named metadata. To
fix this I have added api functions that allow querying the IRSymtab
for the dependent libraries. I hope to expand the api in the future
so that, eventually, all the information needed by a client linker
during scan can be retrieved from the IRSymtab.
Simon Pilgrim [Wed, 12 Jun 2019 10:46:50 +0000 (10:46 +0000)]
[XCore] LowerLOAD/LowerSTORE - Use allowsMemoryAccess wrapper. NFCI.
Noticed in D63075 - there was a allowsMisalignedMemoryAccesses call to check for unaligned loads and a check for aligned legal type loads - which is exactly what allowsMemoryAccess does.
James Henderson [Wed, 12 Jun 2019 10:44:41 +0000 (10:44 +0000)]
[llvm-nm] Fix docs and help text for --print-size
The --print-size help text and documentation claimed that the size was
printed instead of the address, but this is incorrect. It is printed as
well as the address. This patch fixes this issue.
Dylan McKay [Wed, 12 Jun 2019 08:31:07 +0000 (08:31 +0000)]
[AVR] Fix the 'avr-tiny.ll' and 'avr25.ll' subtarget feature tests
When these tests were originally written, the middle end would introduce
an unnecessary copy from r24:r23->GPR16->r24:r23, and these tests
mistakenly relied on it.
The most optimal codegen for the functions in the test cases before this patch
would be NOPs. This is because the first i16 argument always gets the same register
allocation as an i16 return value in the AVR calling convention.
These tests broke in r362963 when the codegen was improved and the
redundant copy was eliminated. After this, the test functions
were lowered to their optimal form - a 'ret' and nothing else.
This patch prepends an extra i16 operand to each of the test functions
so that a 16-bit copy must be inserted for the program to be correct.
Sjoerd Meijer [Wed, 12 Jun 2019 08:28:35 +0000 (08:28 +0000)]
[AArch64] Merge globals when optimising for size
Extern global merging is good for code-size. There's definitely potential for
performance too, but there's one regression in a benchmark that needs
investigating, so that's why we enable it only when we optimise for size for
now.
Craig Topper [Wed, 12 Jun 2019 06:29:53 +0000 (06:29 +0000)]
[X86] Add VCMPSSZrr_Intk and VCMPSDZrr_Intk to isNonFoldablePartialRegisterLoad.
The non-masked versions are already in there. I'm having some
trouble coming up with a way to test this right now. Most load
folding should happen during isel so I'm not sure how to get
peephole pass to do it.
Philip Reames [Tue, 11 Jun 2019 23:21:24 +0000 (23:21 +0000)]
Fix a bug in getSCEVAtScope w.r.t. non-canonical loops
The issue is that if we have a loop with multiple predecessors outside the loop, the code was expecting to merge them and only return if equal, but instead returned the first one seen.
I have no idea if this actually tripped anywhere. I noticed it by accident when reading the code and have no idea how to go about constructing a test case.
Philip Reames [Tue, 11 Jun 2019 22:43:25 +0000 (22:43 +0000)]
Generalize icmp matching in IndVars' eliminateTrunc
We were only matching RHS being a loop invariant value, not the inverse. Since there's nothing which appears to canonicalize loop invariant values to RHS, this means we missed cases.
Sanjay Patel [Tue, 11 Jun 2019 22:25:18 +0000 (22:25 +0000)]
[Analysis] add isSplatValue() for vectors in IR
We have the related getSplatValue() already in IR (see code just above the proposed addition).
But sometimes we only need to know that the value is a splat rather than capture the splatted
scalar value. Also, we have an isSplatValue() function already in SDAG.
Motivation - recent bugs that would potentially benefit from improved splat analysis in IR:
https://bugs.llvm.org/show_bug.cgi?id=37428
https://bugs.llvm.org/show_bug.cgi?id=42174
Jinsong Ji [Tue, 11 Jun 2019 22:09:33 +0000 (22:09 +0000)]
[PowerPC][NFC]Remove sms-simple.ll test temporarily.
Looks like a MachinePipeliner algorithm problem found by
sanitizer-x86_64-linux-fast.
I will backout this test first while investigating the problem to
unblock buildbot.
==49637==ERROR: AddressSanitizer: heap-buffer-overflow on address
0x614000002e08 at pc 0x000004364350 bp 0x7ffe228a3bd0 sp 0x7ffe228a3bc8
READ of size 4 at 0x614000002e08 thread T0
#0 0x436434f in
llvm::SwingSchedulerDAG::checkValidNodeOrder(llvm::SmallVector<llvm::NodeSet,
8u> const&) const
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/CodeGen/MachinePipeliner.cpp:3736:11
#1 0x4342cd0 in llvm::SwingSchedulerDAG::schedule()
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/CodeGen/MachinePipeliner.cpp:486:3
#2 0x434042d in
llvm::MachinePipeliner::swingModuloScheduler(llvm::MachineLoop&)
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/CodeGen/MachinePipeliner.cpp:385:7
#3 0x433eb90 in
llvm::MachinePipeliner::runOnMachineFunction(llvm::MachineFunction&)
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/CodeGen/MachinePipeliner.cpp:207:5
#4 0x428b7ea in
llvm::MachineFunctionPass::runOnFunction(llvm::Function&)
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/CodeGen/MachineFunctionPass.cpp:73:13
#5 0x4d1a913 in llvm::FPPassManager::runOnFunction(llvm::Function&)
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/IR/LegacyPassManager.cpp:1648:27
#6 0x4d1b192 in llvm::FPPassManager::runOnModule(llvm::Module&)
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/IR/LegacyPassManager.cpp:1685:16
#7 0x4d1c06d in runOnModule
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/IR/LegacyPassManager.cpp:1752:27
#8 0x4d1c06d in llvm::legacy::PassManagerImpl::run(llvm::Module&)
/b/sanitizer-x86_64-linux-fast/build/llvm/lib/IR/LegacyPassManager.cpp:1865
#9 0xa48ca3 in compileModule(char**, llvm::LLVMContext&)
/b/sanitizer-x86_64-linux-fast/build/llvm/tools/llc/llc.cpp:611:8
#10 0xa4270f in main
/b/sanitizer-x86_64-linux-fast/build/llvm/tools/llc/llc.cpp:365:22
#11 0x7fec902572e0 in __libc_start_main
(/lib/x86_64-linux-gnu/libc.so.6+0x202e0)
#12 0x971b69 in _start
(/b/sanitizer-x86_64-linux-fast/build/llvm_build_asan/bin/llc+0x971b69)
Jordan Rupprecht [Tue, 11 Jun 2019 21:13:01 +0000 (21:13 +0000)]
[docs] Add "GNU binutils Replacements" section to command guide
Summary:
This splits out a section in the command guide for llvm tools that can be used as replacements for GNU tools. For pages that didn't exist, I added stub pages that can be individually filled in by followup patches.
Tested by running `ninja docs-llvm-html` and inspecting locally.
Alina Sbirlea [Tue, 11 Jun 2019 18:27:49 +0000 (18:27 +0000)]
Only passes that preserve MemorySSA must mark it as preserved.
Summary:
The method `getLoopPassPreservedAnalyses` should not mark MemorySSA as
preserved, because it's being called in a lot of passes that do not
preserve MemorySSA.
Instead, mark the MemorySSA analysis as preserved by each pass that does
preserve it.
These changes only affect the new pass mananger.
Nico Weber [Tue, 11 Jun 2019 15:22:28 +0000 (15:22 +0000)]
lld-link: Reject more than one resource .obj file
Users are exepcted to pass all .res files to the linker, which then
merges all the resource in all .res files into a tree structure and then
converts the final tree structure to a .obj file with .rsrc$01 and
.rsrc$02 sections and then links that.
If the user instead passes several .obj files containing such resources,
the correct thing to do would be to have custom code to merge the trees
in the resource sections instead of doing normal section merging -- but
link.exe rejects if multiple resource obj files are passed in with
LNK4078, so let lld-link do that too instead of silently writing broken
.rsrc sections in that case.
The only real way to run into this is if users manually convert .res
files to .obj files by running cvtres and then handing the resulting
.obj files to lld-link instead, which in practice likely never happens.
(lld-link is slightly stricter than link.exe now: If link.exe is passed
one .obj file created by cvtres, and a .res file, for some reason it
just emits a warning instead of an error and outputs strange looking
data. lld-link now errors out on mixed input like this.)
One way users could accidentally run into this is the following
scenario: If a .res file is passed to lib.exe, then lib.exe calls
cvtres.exe on the .res file before putting it in the output .lib.
(llvm-lib currently doesn't do this.)
link.exe's /wholearchive seems to only add obj files referenced from the
static library index, but lld-link current really adds all files in the
archive. So if lld-link /wholearchive is used with .lib files produced
by lib.exe and .res files were among the files handed to lib.exe, we
previously silently produced invalid output, but now we error out.
link.exe's /wholearchive semantics on the other hand mean that it
wouldn't load the resource object files from the .lib file at all.
Since this scenario is probably still an unlikely corner case,
the difference in behavior here seems fine -- and lld-link might have to
change to use link.exe's /wholearchive semantics in the future anyways.
James Henderson [Tue, 11 Jun 2019 13:51:18 +0000 (13:51 +0000)]
[llvm-dwarfdump] Simplify --ignore-case help text and documentation
There was a typo in the --ignore-case help text that was copied into the
llvm-dwarfdump command-guide. Additionally, this patch simplifies the
wording, since it was unnecessarily verbose: the switch applies for
searching in general and doesn't need explicitly stating different
search modes (which might go out-of-date as options are added or
removed).
Lewis Revill [Tue, 11 Jun 2019 12:57:47 +0000 (12:57 +0000)]
[RISCV] Add lowering of addressing sequences for PIC
This patch allows lowering of PIC addresses by using PC-relative
addressing for DSO-local symbols and accessing the address through the
global offset table for non-DSO-local symbols.
Mikhail Maltsev [Tue, 11 Jun 2019 12:04:32 +0000 (12:04 +0000)]
[ARM] First MVE instructions: scalar shifts.
This introduces a new decoding table for MVE instructions, and starts
by adding the family of scalar shift instructions that are part of the
MVE architecture extension: saturating shifts within a single GPR, and
long shifts across a pair of GPRs (both saturating and normal).
Some of these shift instructions have only 3-bit register fields in
the encoding, with the low bit fixed. So they can only address an odd
or even numbered GPR (depending on the operand), and therefore I add
two new register classes, GPREven and GPROdd.
Nico Weber [Tue, 11 Jun 2019 11:26:50 +0000 (11:26 +0000)]
Let writeWindowsResourceCOFF() take a TimeStamp parameter
For lld, pass in Config->Timestamp (which is set based on lld's
/timestamp: and /Brepro flags). Since the writeWindowsResourceCOFF()
data is only used in-memory by LLD and the obj's timestamp isn't used
for anything in the output, this doesn't change behavior.
For llvm-cvtres, add an optional /timestamp: parameter, and use the
current behavior of calling time() if the parameter is not passed in.
This doesn't really change observable behavior (unless someone passes
/timestamp: to llvm-cvtres, which wasn't possible before), but it
removes the last unqualified call to time() from llvm/lib, which seems
like a good thing.
As suggested by @arsenm on D63075 - this adds a TargetLowering::allowsMemoryAccess wrapper that takes a Load/Store node's MachineMemOperand to handle the AddressSpace/Alignment arguments and will also implicitly handle the MachineMemOperand::Flags change in D63075.
The bug reports that a vectorized loop is stepped through 4 times and each step through the loop seemed to show a different path. I found two problems here:
A) An incorrect line number on a preheader block (for.body.preheader) instruction causes a step into the loop before it begins.
B) Instructions in the middle block have different line numbers which give the impression of another iteration.
In this patch I give all of the middle block instructions the line number of the scalar loop latch terminator branch. This seems to provide the smoothest debugging experience because the vectorized loops will always end on this line before dropping into the scalar loop. To solve problem A I have altered llvm::SplitBlockPredecessors to accommodate loop header blocks.
I have set up a separate review D61933 for a fix which is required for this patch.
James Henderson [Tue, 11 Jun 2019 10:20:07 +0000 (10:20 +0000)]
[llvm-dwarfdump] Add -o to help text and remove --out-file from doc
-o is in the documentation, but not in the llvm-dwarfdump help text.
This patch adds it by inverting the -o and --out-file aliasing. It also
removes --out-file from the documentation, since we don't really want
people to be using this switch in practice.
Simon Tatham [Tue, 11 Jun 2019 10:09:12 +0000 (10:09 +0000)]
[ARM] Fix unused-variable warning in rL363039.
The variable `OffsetMask` is currently only used in an assertion, so
if assertions are compiled out and -Werror is enabled, it becomes a
build failure.
Simon Tatham [Tue, 11 Jun 2019 09:29:18 +0000 (09:29 +0000)]
[ARM] Add the non-MVE instructions in Arm v8.1-M.
This adds support for the new family of conditional selection /
increment / negation instructions; the low-overhead branch
instructions (e.g. BF, WLS, DLS); the CLRM instruction to zero a whole
list of registers at once; the new VMRS/VMSR and VLDR/VSTR
instructions to get data in and out of 8.1-M system registers,
particularly including the new VPR register used by MVE vector
predication.
To support this, we also add a register name 'zr' (used by the CSEL
family to force one of the inputs to the constant 0), and operand
types for lists of registers that are also allowed to include APSR or
VPR (used by CLRM). The VLDR/VSTR instructions also need a new
addressing mode.
The low-overhead branch instructions exist in their own separate
architecture extension, which we treat as enabled by default, but you
can say -mattr=-lob or equivalent to turn it off.
Sander de Smalen [Tue, 11 Jun 2019 08:22:10 +0000 (08:22 +0000)]
Change semantics of fadd/fmul vector reductions.
This patch changes how LLVM handles the accumulator/start value
in the reduction, by never ignoring it regardless of the presence of
fast-math flags on callsites. This change introduces the following
new intrinsics to replace the existing ones:
ld: warning: direct access in function 'llvm::initializeEvexToVexInstPassPass(llvm::PassRegistry&)'
from file '../../lib/libLLVMX86CodeGen.a(X86EvexToVex.cpp.o)' to global weak symbol
'void std::__1::__call_once_proxy<std::__1::tuple<void* (&)(llvm::PassRegistry&),
std::__1::reference_wrapper<llvm::PassRegistry>&&> >(void*)' from file '../../lib/libLLVMCore.a(Verifier.cpp.o)'
means the weak symbol cannot be overridden at runtime. This was likely caused by different translation
units being compiled with different visibility settings.
Symbolize: Make DWPName a symbolizer option instead of an argument to symbolize{,Inlined}Code.
This makes the interface simpler and more consistent with the interface for
.dSYM files and fixes a bug where llvm-symbolizer would not read the dwp if
it was asked to symbolize data before symbolizing code.
projects / llvm / log