Craig Topper [Fri, 13 Oct 2017 06:07:10 +0000 (06:07 +0000)]
[X86] Remove patterns that select unmasked vbroadcastf2x32/vbroadcasti2x32. Prefer vbroadcastsd/vpbroadcastq instead.
There's no advantage to using these instructions when they aren't masked. This enables some additional execution domain switching without needing to update the table.
Sanjoy Das [Fri, 13 Oct 2017 05:50:52 +0000 (05:50 +0000)]
[SCEV] Maintain loop use lists, and use them in forgetLoop
Summary:
Currently we do not correctly invalidate memoized results for add recurrences
that were created directly (i.e. they were not created from a `Value`). This
change fixes this by keeping loop use lists and using the loop use lists to
determine which SCEV expressions to invalidate.
Here are some statistics on the number of uses of in the use lists of all loops
on a clang bootstrap (config: release, no asserts):
Keith Wyss [Fri, 13 Oct 2017 00:06:35 +0000 (00:06 +0000)]
Removing default case statement from covered switch.
Previous patch did not count on the llvm command line parser to restrict the
inputs, but it is safe to do so.
Fix forward for patch with details:
-- https://reviews.llvm.org/D38650 and
-- https://llvm.org/svn/llvm-project/llvm/trunk@315635 91177308-0d34-0410-b5e6-96231b3b80d8
Adrian Prantl [Thu, 12 Oct 2017 22:54:36 +0000 (22:54 +0000)]
Deprecate DwarfUnit::addBlockByrefAddress().
The clang frontend already creates a DIExpression that replicates the
logic in addBlockByrefAddress() exactly, thus making this function
effectively unreachable. To guard against human error I'm hereby
marking the function with an assertion and let it hit the bots before
eventually removing it.
Keith Wyss [Thu, 12 Oct 2017 22:47:42 +0000 (22:47 +0000)]
[XRay][tools] Updated stacks tool with flamegraph output.
Summary:
As the first step to allow analysis and visualization of xray collected data,
allow using the llvm-xray stacks tool to emit a complete listing of stacks in
the format consumable by a flamegraph tool.
Possible follow up formats include chrome trace viewer format and sql load
files.
As a POC, I'm able to generate flamegraphs of an xray instrumented llc compiling
hello world.
Anna Thomas [Thu, 12 Oct 2017 22:39:52 +0000 (22:39 +0000)]
[CVP] Process binary operations even when def is local
Summary:
This patch adds processing of binary operations when the def of operands are in
the same block (i.e. local processing).
Earlier we bailed out in such cases (the bail out was introduced in rL252032)
because LVI at that time was more precise about context at the end of basic
blocks, which implied local def and use analysis didn't benefit CVP.
Since then we've added support for LVI in presence of assumes and guards. The
test cases added show how local def processing in CVP helps adding more
information to the ashr, sdiv, srem and add operators.
Note: processCmp which suffers from the same problem will
be handled in a later patch.
Matthias Braun [Thu, 12 Oct 2017 22:28:54 +0000 (22:28 +0000)]
TargetMachine: Merge TargetMachine and LLVMTargetMachine
Merge LLVMTargetMachine into TargetMachine.
- There is no in-tree target anymore that just implements TargetMachine
but not LLVMTargetMachine.
- It should still be possible to stub out all the various functions in
case a target does not want to use lib/CodeGen
- This simplifies the code and avoids methods ending up in the wrong
interface.
Zachary Turner [Thu, 12 Oct 2017 21:56:05 +0000 (21:56 +0000)]
[lit] Raise the logic for enabling clang & lld substitutions to llvm.
This paves the way for other projects which might /use/ clang or
lld but not necessarily need to the full set of functionality
available to clang and lld tests to be able to have a basic set
of substitutions that allow a project to run the clang or lld
executables.
Eli Friedman [Thu, 12 Oct 2017 20:54:08 +0000 (20:54 +0000)]
[DWARF] Fix bad comparator in sortGlobalExprs.
The comparator passed to std::sort must provide a strict weak ordering;
otherwise, the behavior is undefined.
Fixes an assertion failure generating debug info for globals
split by GlobalOpt. I have a testcase, but not sure how to reduce it,
so not included here. (Someone else came up with a testcase, but I
can't reproduce the crash with it, presumably because my version of LLVM
ends up sorting the array differently.)
This isn't really a complete fix (see the FIXME in the patch), but at
least it doesn't have undefined behavior.
Justin Bogner [Thu, 12 Oct 2017 18:10:22 +0000 (18:10 +0000)]
llvm-isel-fuzzer: Work around BUILD_SHARED_LIBS testing issues
Building with BUILD_SHARED_LIBS makes it tricky to copy around
executables at will, since they won't be able to find the LLVM
libraries any more. This makes testing a feature that's based on the
executable name problematic, so we'll just disable these two tests in
that configuration.
We could potentially fix this by symlinking the lib directory into the
test directory, but that wouldn't work on windows, and losing testing
on windows would be far worse than losing testing on a configuration
that's barely even supported.
Hans Wennborg [Thu, 12 Oct 2017 17:38:22 +0000 (17:38 +0000)]
Work around lack of Wine support for SetFileInformationByHandle harder
In r315079 I added a check for the ERROR_CALL_NOT_IMPLEMENTED error
code, but it turns out earlier versions of Wine just returned false
without setting any error code.
Reintroduce "[SCCP] Propagate integer range info for parameters in IPSCCP."
This is r315288 & r315294, which were reverted due to stage2 bot
failures.
Summary:
This updates the SCCP solver to use of the ValueElement lattice for
parameters, which provides integer range information. The range
information is used to remove unneeded icmp instructions.
For the following function, f() can be optimized to `ret i32 2` with
this change
Lei Huang [Thu, 12 Oct 2017 16:43:33 +0000 (16:43 +0000)]
[PowerPC] Add profitablilty check for conversion to mtctr loops
Add profitability checks for modifying counted loops to use the mtctr instruction.
The latency of mtctr is only justified if there are more than 4 comparisons that
will be removed as a result. Usually counted loops are formed relatively early
and before unrolling, so most low trip count loops often don't survive. However
we want to ensure that if they do, we do not mistakenly update them to mtctr loops.
Use CodeMetrics to ensure we are only doing this for small loops with small trip counts.
Tim Renouf [Thu, 12 Oct 2017 16:16:41 +0000 (16:16 +0000)]
[AMDGPU] For amdpal, widen interpolation mode workaround
Summary:
The interpolation mode workaround ensures that at least one
interpolation mode is enabled in PSInputAddr. It does not also check
PSInputEna on the basis that the user might enable bits in that
depending on run-time state.
However, for amdpal os type, the user does not enable some bits after
compilation based on run-time states; the register values being
generated here are the final ones set in the hardware. Therefore, apply
the workaround to PSInputAddr and PSInputEnable together. (The case
where a bit is set in PSInputAddr but not in PSInputEnable is where the
frontend set up an input arg for a particular interpolation mode, but
nothing uses that input arg. Really we should have an earlier pass that
removes such an arg.)
Diana Picus [Thu, 12 Oct 2017 13:59:51 +0000 (13:59 +0000)]
MachineInstr: Make isEqual agree with getHashValue in MachineInstrExpressionTrait
MachineInstr::isIdenticalTo has a lot of logic for dealing with register
Defs (i.e. deciding whether to take them into account or ignore them).
This logic gets things wrong in some obscure cases, for instance if an
operand is not a Def for both the current MI and the one we are
comparing to.
I'm not sure if it's possible for this to happen for regular register
operands, but it may happen in the ARM backend for special operands
which use sentinel values for the register (i.e. 0, which is neither a
physical register nor a virtual one).
This causes MachineInstrExpressionTrait::isEqual (which uses
MachineInstr::isIdenticalTo) to return true for the following
instructions, which are the same except for the fact that one sets the
flags and the other one doesn't:
%1114 = ADDrsi %1113, %216, 17, 14, _, def _
%1115 = ADDrsi %1113, %216, 17, 14, _, _
OTOH, MachineInstrExpressionTrait::getHashValue returns different values
for the 2 instructions due to the different isDef on the last operand.
In practice this means that when trying to add those instructions to a
DenseMap, they will be considered different because of their different
hash values, but when growing the map we might get an assertion while
copying from the old buckets to the new buckets because isEqual
misleadingly returns true.
This patch makes sure that isEqual and getHashValue agree, by improving
the checks in MachineInstr::isIdenticalTo when we are ignoring virtual
register definitions (which is what the Trait uses). Firstly, instead of
checking isPhysicalRegister, we use !isVirtualRegister, so that we cover
both physical registers and sentinel values. Secondly, instead of
checking MachineOperand::isReg, we use MachineOperand::isIdenticalTo,
which checks isReg, isSubReg and isDef, which are the same values that
the hash function uses to compute the hash.
Note that the function is symmetric with this change, since if the
current operand is not a Def, we check MachineOperand::isIdenticalTo,
which returns false if the operands have different isDef's.
Daniel Jasper [Thu, 12 Oct 2017 13:25:05 +0000 (13:25 +0000)]
Reinstantiate old/bad deduplication logic that was removed in r315279.
While this shouldn't be necessary anymore, we have cases where we run
into the assertion below, i.e. cases with two non-fragment entries for the
same variable at different frame indices.
This should be fixed, but for now, we should revert to a version that
does not trigger asserts.
Hiroshi Inoue [Thu, 12 Oct 2017 06:26:04 +0000 (06:26 +0000)]
[ScheduleDAGInstrs] fix behavior of getUnderlyingObjectsForCodeGen when no identifiable object found
This patch fixes the bug introduced in https://reviews.llvm.org/D35907; the bug is reported by http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20171002/491452.html.
Before D35907, when GetUnderlyingObjects fails to find an identifiable object, allMMOsOkay lambda in getUnderlyingObjectsForInstr returns false and Objects vector is cleared. This behavior is unintentionally changed by D35907.
This patch makes the behavior for such case same as the previous behavior.
Since D35907 introduced a wrapper function getUnderlyingObjectsForCodeGen around GetUnderlyingObjects, getUnderlyingObjectsForCodeGen is modified to return a boolean value to ask the caller to clear the Objects vector.
Justin Bogner [Thu, 12 Oct 2017 04:35:32 +0000 (04:35 +0000)]
Re-commit "llvm-isel-fuzzer: Handle a subset of backend flags in the exec name"
Here we add a secondary option parser to llvm-isel-fuzzer (and provide
it for use with other fuzzers). With this, you can copy the fuzzer to
a name like llvm-isel-fuzzer=aarch64-gisel for a fuzzer that fuzzer
AArch64 with GlobalISel enabled, or fuzzer=x86_64 to fuzz x86, with no
flags required. This should be useful for running these in OSS-Fuzz.
Note that this handrolls a subset of cl::opts to recognize, rather
than embedding a complete command parser for argv[0]. If we find we
really need the flexibility of handling arbitrary options at some
point we can rethink this.
This re-applies 315545 using "=" instead of ":" as a separator for
arguments.
> llvm-isel-fuzzer: Handle a subset of backend flags in the executable name
>
> Here we add a secondary option parser to llvm-isel-fuzzer (and provide
> it for use with other fuzzers). With this, you can copy the fuzzer to
> a name like llvm-isel-fuzzer:aarch64-gisel for a fuzzer that fuzzer
> AArch64 with GlobalISel enabled, or fuzzer:x86_64 to fuzz x86, with no
> flags required. This should be useful for running these in OSS-Fuzz.
>
> Note that this handrolls a subset of cl::opts to recognize, rather
> than embedding a complete command parser for argv[0]. If we find we
> really need the flexibility of handling arbitrary options at some
> point we can rethink this.
Justin Bogner [Thu, 12 Oct 2017 01:57:49 +0000 (01:57 +0000)]
llvm-isel-fuzzer: Handle a subset of backend flags in the executable name
Here we add a secondary option parser to llvm-isel-fuzzer (and provide
it for use with other fuzzers). With this, you can copy the fuzzer to
a name like llvm-isel-fuzzer:aarch64-gisel for a fuzzer that fuzzer
AArch64 with GlobalISel enabled, or fuzzer:x86_64 to fuzz x86, with no
flags required. This should be useful for running these in OSS-Fuzz.
Note that this handrolls a subset of cl::opts to recognize, rather
than embedding a complete command parser for argv[0]. If we find we
really need the flexibility of handling arbitrary options at some
point we can rethink this.
This is failing due to some code that isn't built on MSVC
so I didn't catch. Not immediately obvious how to fix this
at first glance, so I'm reverting for now.
Lang Hames [Wed, 11 Oct 2017 23:34:47 +0000 (23:34 +0000)]
[MC] Have MCObjectStreamer take its MCAsmBackend argument via unique_ptr.
MCObjectStreamer owns its MCCodeEmitter -- this fixes the types to reflect that,
and allows us to remove the last instance of MCObjectStreamer's weird "holding
ownership via someone else's reference" trick.
Zachary Turner [Wed, 11 Oct 2017 23:33:06 +0000 (23:33 +0000)]
[ADT] Make Twine's copy constructor private.
There's a lot of misuse of Twine scattered around LLVM. This
ranges in severity from benign (returning a Twine from a function
by value that is just a string literal) to pretty sketchy (storing
a Twine by value in a class). While there are some uses for
copying Twines, most of the very compelling ones are confined
to the Twine class implementation itself, and other uses are
either dubious or easily worked around.
This patch makes Twine's copy constructor private, and fixes up
all callsites.
AMDGPU/NFC: Rename code object metadata as HSA metadata
- Rename AMDGPUCodeObjectMetadata to AMDGPUMetadata (PAL metadata will be included in this file in the follow up change)
- Rename AMDGPUCodeObjectMetadataStreamer to AMDGPUHSAMetadataStreamer
- Introduce HSAMD namespace
- Other minor name changes in function and test names
Hans Wennborg [Wed, 11 Oct 2017 22:04:14 +0000 (22:04 +0000)]
Support: Work around missing SetFileInformationByHandle on Wine
In r315079, fs::rename was reimplemented in terms of CreateFile and
SetFileInformationByHandle. Unfortunately, the latter isn't supported by
Wine. This adds a fallback to MoveFileEx for that case.
Reid Kleckner [Wed, 11 Oct 2017 21:24:33 +0000 (21:24 +0000)]
[codeview] Implement FPO data assembler directives
Summary:
This adds a set of new directives that describe 32-bit x86 prologues.
The directives are limited and do not expose the full complexity of
codeview FPO data. They are merely a convenience for the compiler to
generate more readable assembly so we don't need to generate tons of
labels in CodeGen. If our prologue emission changes in the future, we
can change the set of available directives to suit our needs. These are
modelled after the .seh_ directives, which use a different format that
interacts with exception handling.
I tried to follow the implementation of ARM EHABI CFI directives by
sinking most directives out of MCStreamer and into X86TargetStreamer.
This helps avoid polluting non-X86 code with WinCOFF specific logic.
I used cdb to confirm that this can show locals in parent CSRs in a few
cases, most importantly the one where we use ESI as a frame pointer,
i.e. the one in http://crbug.com/756153#c28
Once we have cdb integration in debuginfo-tests, we can add integration
tests there.
Lei Huang [Wed, 11 Oct 2017 20:20:58 +0000 (20:20 +0000)]
[PowerPC] Utilize DQ-Form instructions for spill/restore and fix FrameIndex elimination to only use `lis/addi` if necessary.
Currently we produce a bunch of unnecessary code when emitting the
prologue/epilogue for spills/restores. Namely, if the load from stack
slot/store to stack slot instruction is an X-Form instruction, we will
always produce an LIS/ORI sequence for the stack offset.
Furthermore, we have not exploited the P9 vector D-Form loads/stores for this
purpose.
This patch address both issues.
Specifying the D-Form load as the instruction to use for stack spills/reloads
should be safe because:
1. The stack should be aligned according to the ABI
2. If the stack isn't aligned, PPCRegisterInfo::eliminateFrameIndex() will
check for the offset being a multiple of 16 and will convert it to an
X-Form instruction if it isn't.
Zachary Turner [Wed, 11 Oct 2017 20:12:09 +0000 (20:12 +0000)]
[llvm-rc] Use proper search algorithm for finding resources.
Previously we would only look in the current directory for a
resource, which might not be the same as the directory of the
rc file. Furthermore, MSVC rc supports a /I option, and can
also look in the system environment. This patch adds support
for this search algorithm.
Daniel Neilson [Wed, 11 Oct 2017 19:05:14 +0000 (19:05 +0000)]
[SCEV] Properly handle the case of a non-constant start with a zero accum in ScalarEvolution::createAddRecFromPHIWithCastsImpl
Summary:
This patch fixes an error in the patch to ScalarEvolution::createAddRecFromPHIWithCastsImpl
made in D37265. In that patch we handle the cases where the either the start or accum values can be
zero after truncation. But, we assume that the start value must be a constant if the accum is
zero. This is clearly an erroneous assumption. This change removes that assumption.
Jake Ehrlich [Wed, 11 Oct 2017 18:09:18 +0000 (18:09 +0000)]
Reland "[llvm-objcopy] Add support for --strip-sections to remove all section headers leaving only program headers and loadable segment data"
ubsan caught an issue I made where I was converting a null pointer to a
reference.
elf utils implements a particularly extreme form of stripping that I'd
like to support. eu-strip has an option called "strip-sections" that
removes all section headers and leaves only program headers and the
segment data. I have implemented this option partly as a test but mainly
because in Fuchsia we would like to use this option to minimize the size
of our executables. The other strip options that are on my list include
--strip-all and --strip-debug. This is a preliminary implementation that
I'd like to start using in Fuchsia builds if possible. This change
implements such a stripping option for llvm-objcopy
projects / llvm / log