Alexey Bataev [Tue, 3 Apr 2018 16:40:33 +0000 (16:40 +0000)]
Recommit "[SLP] Fix issues with debug output in the SLP vectorizer."
The primary issue here is that using NDEBUG alone isn't enough to guard
debug printing -- instead the DEBUG() macro needs to be used so that the
specific pass debug logging check is employed. Without this, every
asserts-enabled build was printing out information when it hit this.
I also fixed another place where we had multiple statements in a DEBUG
macro to use {}s to be a bit cleaner. And I fixed a place that used
errs() rather than dbgs().
llvm-svn: 329082
Florian Hahn [Tue, 3 Apr 2018 16:37:58 +0000 (16:37 +0000)]
[LoopInterchange] Update tests so DA can handle access after D35430.
I have taken the opportunity to simplify some tests slightly and move
parts around.
It also brings back a few IR checks for interchangable loops.
Reviewers: karthikthecool, sebpop, grosser
Reviewed By: sebpop
Differential Revision: https://reviews.llvm.org/D45207
llvm-svn: 329081
Alexey Bataev [Tue, 3 Apr 2018 16:31:26 +0000 (16:31 +0000)]
[SLP] Added tests for checks of reordering of the repeated instructions,
NFC.
llvm-svn: 329080
Krzysztof Parzyszek [Tue, 3 Apr 2018 16:06:36 +0000 (16:06 +0000)]
[Hexagon] Remove -mhvx-double and the corresponding subtarget feature
Specifying the HVX vector length should be done via the -mhvx-length
option.
llvm-svn: 329079
Krzysztof Parzyszek [Tue, 3 Apr 2018 16:05:20 +0000 (16:05 +0000)]
[Hexagon] Remove unneeded attributes from lit test
llvm-svn: 329078
Krzysztof Parzyszek [Tue, 3 Apr 2018 15:59:10 +0000 (15:59 +0000)]
[Hexagon] Remove -mhvx-double and the corresponding subtarget feature
Specifying the HVX vector length should be done via the -mhvx-length
option.
llvm-svn: 329077
Puyan Lotfi [Tue, 3 Apr 2018 15:53:49 +0000 (15:53 +0000)]
Adding optional Name parameter to createVirtualRegister and createGenericVirtualRegister.
llvm-svn: 329076
Marshall Clow [Tue, 3 Apr 2018 15:48:24 +0000 (15:48 +0000)]
Implement P0754R2: The <version> header.
llvm-svn: 329075
Andrea Di Biagio [Tue, 3 Apr 2018 15:36:15 +0000 (15:36 +0000)]
[llvm-mca] Remove redundant include from BackendStatistics.h. NFC
Also use llvm::DenseMap for Histograms (instead of std::map).
llvm-svn: 329074
Haojian Wu [Tue, 3 Apr 2018 15:10:24 +0000 (15:10 +0000)]
[clang-tidy] Check for sizeof that call functions
Summary:
A common mistake that I have found in our codebase is calling a function to get an integer or enum that represents the type such as:
```
int numBytes = numElements * sizeof(x.GetType());
```
So this extends the `sizeof` check to check for these cases. There is also a `WarnOnSizeOfCall` option so it can be disabled.
Patch by Paul Fultz II!
Reviewers: hokein, alexfh, aaron.ballman, ilya-biryukov
Reviewed By: alexfh
Subscribers: lebedev.ri, xazax.hun, jkorous-apple, cfe-commits
Tags: #clang-tools-extra
Differential Revision: https://reviews.llvm.org/D44231
llvm-svn: 329073
Guansong Zhang [Tue, 3 Apr 2018 15:01:34 +0000 (15:01 +0000)]
[OpenMP] enable bc file compilation using the latest clang
Summary: adding cuda-rdc flag to allow extern global data
Reviewers: grokos
Reviewed By: grokos
Subscribers: gregrodgers, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D44992
llvm-svn: 329072
Benjamin Kramer [Tue, 3 Apr 2018 14:40:33 +0000 (14:40 +0000)]
Revert "[SLP] Fix PR36481: vectorize reassociated instructions."
This reverts commit r328980 and r329046. Makes the vectorizer crash.
llvm-svn: 329071
Ben Hamilton [Tue, 3 Apr 2018 14:07:11 +0000 (14:07 +0000)]
[clang-format/ObjC] Do not detect "[]" as ObjC method expression
Summary:
The following C++ code was being detected by
`guessLanguage()` as Objective-C:
#define FOO(...) auto bar = [] __VA_ARGS__;
This was because `[] __VA_ARGS__` is not currently detected as a C++
lambda expression (it has no parens or braces), so
`TokenAnnotator::parseSquare()` incorrectly treats the opening square
as an ObjC method expression.
We have two options to fix this:
1. Parse `[] __VA_ARGS__` explicitly as a C++ lambda
2. Make it so `[]` is never parsed as an Objective-C method expression
This diff implements option 2, which causes the `[` to be parsed
as `TT_ArraySubscriptLSquare` instead of `TT_ObjCMethodExpr`.
Note that when I fixed this, it caused one change in formatting
behavior, where the following was implicitly relying on the `[`
being parsed as `TT_ObjCMethodExpr`:
A<int * []> a;
becomes:
A<int *[]> a;
with `Style.PointerAlignment = Middle`.
I don't really know what the desired format is for this syntax; the
test was added by Janusz Sobczak and integrated by @djasper in
https://github.com/llvm-mirror/clang/commit/
b511fe9818829d7ece0cc0b2ce1fbe04a1f0739a
.
I went ahead and changed the test for now.
Test Plan: New tests added. Ran tests with:
% make -j12 FormatTests && ./tools/clang/unittests/Format/FormatTests
Fixes: https://bugs.llvm.org/show_bug.cgi?id=36248
Reviewers: djasper, jolesiak
Reviewed By: djasper
Subscribers: klimek, cfe-commits, djasper
Differential Revision: https://reviews.llvm.org/D45169
llvm-svn: 329070
Ben Hamilton [Tue, 3 Apr 2018 14:07:09 +0000 (14:07 +0000)]
[clang-format/ObjC] Do not insert space after opening brace of ObjC dict literal
Summary:
D44816 attempted to fix a few cases where `clang-format` incorrectly
inserted a space before the closing brace of an Objective-C dictionary
literal.
This revealed there were still a few cases where we inserted a space
after the opening brace of an Objective-C dictionary literal.
This fixes the formatting to be consistent and adds more tests.
Test Plan: New tests added. Confirmed tests failed before
diff and passed after diff.
Ran tests with:
% make -j12 FormatTests && ./tools/clang/unittests/Format/FormatTests
Reviewers: djasper, jolesiak, krasimir
Reviewed By: djasper
Subscribers: klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D45168
llvm-svn: 329069
Andrea Di Biagio [Tue, 3 Apr 2018 13:52:26 +0000 (13:52 +0000)]
[MC] Fix -Wmissing-field-initializer warning after r329067.
This should fix the problem reported by the lld buildbots:
- Builder lld-x86_64-darwin13, Build #19782
- Builder lld-perf-testsuite, Build #1419
llvm-svn: 329068
Andrea Di Biagio [Tue, 3 Apr 2018 13:36:24 +0000 (13:36 +0000)]
[MC][Tablegen] Allow the definition of processor register files in the scheduling model for llvm-mca
This patch allows the description of register files in processor scheduling
models. This addresses PR36662.
A new tablegen class named 'RegisterFile' has been added to TargetSchedule.td.
Targets can optionally describe register files for their processors using that
class. In particular, class RegisterFile allows to specify:
- The total number of physical registers.
- Which target registers are accessible through the register file.
- The cost of allocating a register at register renaming stage.
Example (from this patch - see file X86/X86ScheduleBtVer2.td)
def FpuPRF : RegisterFile<72, [VR64, VR128, VR256], [1, 1, 2]>
Here, FpuPRF describes a register file for MMX/XMM/YMM registers. On Jaguar
(btver2), a YMM register definition consumes 2 physical registers, while MMX/XMM
register definitions only cost 1 physical register.
The syntax allows to specify an empty set of register classes. An empty set of
register classes means: this register file models all the registers specified by
the Target. For each register class, users can specify an optional register
cost. By default, register costs default to 1. A value of 0 for the number of
physical registers means: "this register file has an unbounded number of
physical registers".
This patch is structured in two parts.
* Part 1 - MC/Tablegen *
A first part adds the tablegen definition of RegisterFile, and teaches the
SubtargetEmitter how to emit information related to register files.
Information about register files is accessible through an instance of
MCExtraProcessorInfo.
The idea behind this design is to logically partition the processor description
which is only used by external tools (like llvm-mca) from the processor
information used by the llvm machine schedulers.
I think that this design would make easier for targets to get rid of the extra
processor information if they don't want it.
* Part 2 - llvm-mca related *
The second part of this patch is related to changes to llvm-mca.
The main differences are:
1) class RegisterFile now needs to take into account the "cost of a register"
when allocating physical registers at register renaming stage.
2) Point 1. triggered a minor refactoring which lef to the removal of the
"maximum 32 register files" restriction.
3) The BackendStatistics view has been updated so that we can print out extra
details related to each register file implemented by the processor.
The effect of point 3. is also visible in tests register-files-[1..5].s.
Differential Revision: https://reviews.llvm.org/D44980
llvm-svn: 329067
Sanjay Patel [Tue, 3 Apr 2018 13:05:20 +0000 (13:05 +0000)]
[LangRef] fix description and examples of fptrunc
As noted in PR36966:
https://bugs.llvm.org/show_bug.cgi?id=36966
The old description doesn't match what we do in code,
so this just fixes the documentation to avoid confusion.
Differential Revision: https://reviews.llvm.org/D45190
llvm-svn: 329065
Hiroshi Inoue [Tue, 3 Apr 2018 12:49:42 +0000 (12:49 +0000)]
[PowerPC] reorder entries in P9InstrResources.td in alphabetical order; NFC
Reorder entries added in my previous commit (rL328969) to keep alphabetical order.
llvm-svn: 329064
George Rimar [Tue, 3 Apr 2018 12:39:28 +0000 (12:39 +0000)]
[ELF] - Check that output sections fit in address space.
Added checks to test that we do not produce
output where VA of sections overruns the address
space available.
Differential revision: https://reviews.llvm.org/D43820
llvm-svn: 329063
George Rimar [Tue, 3 Apr 2018 12:28:53 +0000 (12:28 +0000)]
[ELF] - Fix the comment. NFC.
llvm-svn: 329062
George Rimar [Tue, 3 Apr 2018 12:19:04 +0000 (12:19 +0000)]
[ELF] - Relax checks for R_386_8/R_386_16 relocations.
This fixes PR36927.
The issue is next. Imagine we have -Ttext 0x7c and code below.
.code16
.global _start
_start:
movb $_start+0x83,%ah
So we have R_386_8 relocation and _start at 0x7C.
Addend is 0x83 == 131. We will sign extend it to 0xffffffffffffff83.
Now, 0xffffffffffffff83 + 0x7c gives us 0xFFFFFFFFFFFFFFFF.
Techically 0x83 + 0x7c == 0xFF, we do not exceed 1 byte value, but
currently LLD errors out, because we use checkUInt<8>.
Let's try to use checkInt<8> now and the following code to see if it can help (no):
main.s:
.byte foo
input.s:
.globl foo
.hidden foo
foo = 0xff
Here, foo is 0xFF. And addend is 0x0. Final value is 0x00000000000000FF.
Again, it fits one byte well, but with checkInt<8>,
we would error out it, so we can't use it.
What we want to do is to check that the result fits 1 byte well.
Patch changes the check to checkIntUInt to fix the issue.
Differential revision: https://reviews.llvm.org/D45051
llvm-svn: 329061
George Rimar [Tue, 3 Apr 2018 12:06:29 +0000 (12:06 +0000)]
[ELF] - Simplify createFiles. NFCI.
Groups paired options together.
Differential revision: https://reviews.llvm.org/D45090
llvm-svn: 329060
George Rimar [Tue, 3 Apr 2018 11:58:23 +0000 (11:58 +0000)]
[ELF] - X86_64: don't allow 8/16 bit dynamic relocations.
Having 8/16 bits dynamic relocations is incorrect.
Both gold and bfd (built from latest sources) disallow
that too.
Differential revision: https://reviews.llvm.org/D45158
llvm-svn: 329059
George Rimar [Tue, 3 Apr 2018 10:40:39 +0000 (10:40 +0000)]
[ELF] - Remove dead declaration. NFC.
llvm-svn: 329058
Chandler Carruth [Tue, 3 Apr 2018 10:28:56 +0000 (10:28 +0000)]
[x86] Fix a pretty obvious think-o with my asm scrubbing. You have to in
fact use regular expression syntax to use regular expressions.
Should restore the bots. Sorry for the noise on this test.
Thanks to Philip for spotting the bug!
llvm-svn: 329057
Chandler Carruth [Tue, 3 Apr 2018 10:04:37 +0000 (10:04 +0000)]
[x86] Clean up and enhance a test around eflags copying.
This adds the basic test cases from all the EFLAGS bugs in more direct
forms. It also switches to generated check lines, and includes both
32-bit and 64-bit variations.
No functionality changing here, just setting things up to have a nice
clean asm diff in my EFLAGS patch.
llvm-svn: 329056
Chandler Carruth [Tue, 3 Apr 2018 09:57:05 +0000 (09:57 +0000)]
[x86] Extend my goofy SP offset scrubbing for llc test cases to actually
do explicit scrubbing of the offsets of stack spills and reloads.
You can always turn this off in order to test specific stack slot usage.
We were already hiding most of this, but the new logic hides it more
generically. Notably, we should effectively hide stack slot churn in
functions that have a frame pointer now, and should also hide it when
changing a function from stack pointer to frame pointer. That transition
already changes enough to be clearly noticed in the test case diff,
showing *every* spill and reload is really noisy without benefit. See
the test case I ran this on as a classic example.
llvm-svn: 329055
Alexander Potapenko [Tue, 3 Apr 2018 09:50:06 +0000 (09:50 +0000)]
MSan: introduce the conservative assembly handling mode.
The default assembly handling mode may introduce false positives in the
cases when MSan doesn't understand that the assembly call initializes
the memory pointed to by one of its arguments.
We introduce the conservative mode, which initializes the first
|sizeof(type)| bytes for every |type*| pointer passed into the
assembly statement.
llvm-svn: 329054
Simon Dardis [Tue, 3 Apr 2018 09:40:07 +0000 (09:40 +0000)]
[clangd][cmake] Provide libatomic when there is no native support for 64bit atomics
This addresses a persistent failure on clang-cmake-mips buildbot.
Reviewers: ioeric
Differential Revision: https://reviews.llvm.org/D44248
llvm-svn: 329053
Hans Wennborg [Tue, 3 Apr 2018 09:28:21 +0000 (09:28 +0000)]
UsersManual.rst: update text for /GX- to match r328708
llvm-svn: 329052
Serguei Katkov [Tue, 3 Apr 2018 07:29:00 +0000 (07:29 +0000)]
[SCEV] Fix PR36974.
The patch changes the usage of dominate to properlyDominate
to satisfy the condition !(a < a) while using std::max.
It is actually NFC due to set data structure is used to keep
the Loops and no two identical loops can be in collection.
So in reality there is no difference between usage of
dominate and properlyDominate in this particular case.
However it might be changed so it is better to fix it.
llvm-svn: 329051
Eric Christopher [Tue, 3 Apr 2018 07:01:33 +0000 (07:01 +0000)]
Add a wrapper around llvm-objdump to look for indirect calls/jmps in x86 assembly.
Useful when looking for indirect calls/jmps the need mitigation
via retpoline or other mitigations for Spectre v2.
Feedback, extension, additional patches welcome.
llvm-svn: 329050
Craig Topper [Tue, 3 Apr 2018 06:37:04 +0000 (06:37 +0000)]
[X86] Reduce number of OpPrefix bits in TSFlags to 2. NFCI
TSFlag doesn't need to disambiguate NoPrfx from PS. So shift the encodings so PS is NoPrfx|0x4.
llvm-svn: 329049
Craig Topper [Tue, 3 Apr 2018 06:37:01 +0000 (06:37 +0000)]
[X86][TableGen] Add a missing error check to make sure EVEX instructions use one PS/PD/XS/XD prefixes.
llvm-svn: 329048
Max Kazantsev [Tue, 3 Apr 2018 05:57:19 +0000 (05:57 +0000)]
[SCEV] Make computeExitLimit more simple and more powerful
Current implementation of `computeExitLimit` has a big piece of code
the only purpose of which is to prove that after the execution of this
block the latch will be executed. What it currently checks is actually a
subset of situations where the exiting block dominates latch.
This patch replaces all these checks for simple particular cases with
domination check over loop's latch which is the only necessary condition
of taking the exiting block into consideration. This change allows to
calculate exact loop taken count for simple loops like
for (int i = 0; i < 100; i++) {
if (cond) {...} else {...}
if (i > 50) break;
. . .
}
Differential Revision: https://reviews.llvm.org/D44677
Reviewed By: efriedma
llvm-svn: 329047
Chandler Carruth [Tue, 3 Apr 2018 05:27:28 +0000 (05:27 +0000)]
[SLP] Fix issues with debug output in the SLP vectorizer.
The primary issue here is that using NDEBUG alone isn't enough to guard
debug printing -- instead the DEBUG() macro needs to be used so that the
specific pass debug logging check is employed. Without this, every
asserts-enabled build was printing out information when it hit this.
I also fixed another place where we had multiple statements in a DEBUG
macro to use {}s to be a bit cleaner. And I fixed a place that used
`errs()` rather than `dbgs()`.
llvm-svn: 329046
Craig Topper [Tue, 3 Apr 2018 05:10:12 +0000 (05:10 +0000)]
[TableGen] Use llvm::cast instead of static_cast so that the cast will be checked. NFC
llvm-svn: 329045
Rafael Espindola [Tue, 3 Apr 2018 04:06:14 +0000 (04:06 +0000)]
Use OffsetMap in getSectionPiece.
OffsetMap maps to a SectionPiece index, but we were not taking
advantage of that in getSectionPiece.
With this patch both getOffset and getSectionPiece use OffsetMap and
the binary search is moved to findSectionPiece.
llvm-svn: 329044
Yonghong Song [Tue, 3 Apr 2018 03:56:37 +0000 (03:56 +0000)]
bpf: fix incorrect SELECT_CC lowering
Commit
37962a331c77 ("bpf: Improve expanding logic in LowerSELECT_CC")
intended to improve code quality for certain jmp conditions. The
commit, however, has a couple of issues:
(1). In code, just swap is not enough, ConditionalCode CC
should also be swapped, otherwise incorrect code will
be generated.
(2). The ConditionalCode swap should be subject to
getHasJmpExt(). If getHasJmpExt() is False, certain
conditional codes will not be supported and swap
may generate incorrect code.
The original goal for this patch is to optimize jmp operations
which does not have JmpExt turned on. If JmpExt is on,
better code could be generated. For example, the test
select_ri.ll is introduced to demonstrate the optimization.
The same result can be achieved with -mcpu=v2 flag.
Signed-off-by: Yonghong Song <yhs@fb.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
llvm-svn: 329043
Ikhlas Ajbar [Tue, 3 Apr 2018 03:39:43 +0000 (03:39 +0000)]
peel loops with runtime small trip counts
For Hexagon, peeling loops with small runtime trip count is beneficial for our
benchmarks. We set PeelCount in HexagonTargetInfo.cpp and we use PeelCount set
by the target for computing the desired peel count.
Differential Revision: https://reviews.llvm.org/D44880
llvm-svn: 329042
Nico Weber [Tue, 3 Apr 2018 02:37:49 +0000 (02:37 +0000)]
Remove utils/makellvm; it doesn't look like it works with cmake builds.
llvm-svn: 329041
Chandler Carruth [Tue, 3 Apr 2018 02:19:05 +0000 (02:19 +0000)]
[x86] Tidy up test case, generate check lines with script. NFC.
Just adds basic block labels and tidies up where comments go in the test
case and then generates fresh CHECK lines with the script. This way, the
check lines are much easier to maintain. They were already close to this
but not quite there.
llvm-svn: 329040
Rui Ueyama [Tue, 3 Apr 2018 02:06:57 +0000 (02:06 +0000)]
Merge LazyArchive::fetch() and ArchiveFile::getMember(). NFC.
They are to pull out an object file for a symbol, but for a historical
reason the code is written in two separate functions. This patch
merges them.
llvm-svn: 329039
Eric Fiselier [Tue, 3 Apr 2018 01:52:12 +0000 (01:52 +0000)]
Fix fs::proximate tests on platforms where /net exists.
The proximate tests depended on `/net` not being a valid path,
however, on OS X it is.
Correct the tests to handle this.
llvm-svn: 329038
Aaron Smith [Tue, 3 Apr 2018 00:22:12 +0000 (00:22 +0000)]
[lit] Prefer opening files with open (Python 2) rather than io.open which requires io.
Only rely on Python 3 (io.open) when necessary. This puts TestRunnyer.py closer to how it behaved
before the changes introduced in D43165 and silences a few Windows build bot failures.
Thanks to Stella Stamenova for the patch!
llvm-svn: 329037
Eugene Zelenko [Tue, 3 Apr 2018 00:11:50 +0000 (00:11 +0000)]
[AST] Fix some Clang-tidy modernize-use-auto warnings; other minor fixes (NFC).
llvm-svn: 329036
Haicheng Wu [Tue, 3 Apr 2018 00:05:10 +0000 (00:05 +0000)]
[SLP] Distinguish "demanded and shrinkable" from "demanded and not shrinkable" values when determining the minimum bitwidth
We use two approaches for determining the minimum bitwidth.
* Demanded bits
* Value tracking
If demanded bits doesn't result in a narrower type, we then try value tracking.
We need this if we want to root SLP trees with the indices of getelementptr
instructions since all the bits of the indices are demanded.
But there is a missing piece though. We need to be able to distinguish "demanded
and shrinkable" from "demanded and not shrinkable". For example, the bits of %i
in
%i = sext i32 %e1 to i64
%gep = getelementptr inbounds i64, i64* %p, i64 %i
are demanded, but we can shrink %i's type to i32 because it won't change the
result of the getelementptr. On the other hand, in
%tmp15 = sext i32 %tmp14 to i64
%tmp16 = insertvalue { i64, i64 } undef, i64 %tmp15, 0
it doesn't make sense to shrink %tmp15 and we can skip the value tracking.
Ideas are from Matthew Simpson!
Differential Revision: https://reviews.llvm.org/D44868
llvm-svn: 329035
Rui Ueyama [Mon, 2 Apr 2018 23:58:50 +0000 (23:58 +0000)]
Inline two trivial functions that are called only once. NFC.
llvm-svn: 329034
Brian Gesiak [Mon, 2 Apr 2018 23:39:40 +0000 (23:39 +0000)]
[Coroutines] Avoid assert splitting hidden coros
Summary:
When attempting to split a coroutine with 'hidden' visibility (for
example, a C++ coroutine that is inlined when compiled with the option
'-fvisibility-inlines-hidden'), LLVM would hit an assertion in
include/llvm/IR/GlobalValue.h:240: "local linkage requires default
visibility". The issue is that the visibility is copied from the source
of the function split in the `CloneFunctionInto` function, but the linkage
is not. To fix, create the new function first with external linkage,
then copy the linkage from the original function *after* `CloneFunctionInto`
is called.
Since `GlobalValue::setLinkage` in turn calls `maybeSetDsoLocal`, the
explicit call to `setDSOLocal` can be removed in CoroSplit.cpp.
Test Plan: check-llvm
Reviewers: GorNishanov, lewissbaker, EricWF, majnemer, rnk
Reviewed By: rnk
Subscribers: llvm-commits, eric_niebler
Differential Revision: https://reviews.llvm.org/D44185
llvm-svn: 329033
Petr Hosek [Mon, 2 Apr 2018 23:36:14 +0000 (23:36 +0000)]
[Driver] Wire up the -f[no-]rtlib-add-rpath flag and tests
D30700 added the -f[no-]rtlib-add-rpath flag, but that flag was never
wired up in the driver and tests were updated to check whether it
actually does anything. This patch wires up the flag and updates test.
Differential Revision: https://reviews.llvm.org/D45145
llvm-svn: 329032
Eric Fiselier [Mon, 2 Apr 2018 23:35:24 +0000 (23:35 +0000)]
Implement P0430R2 - File system library on non-POSIX systems.
This patch implements P0430R2, who's largest change is adding the path::format
enumeration for supporting path format conversions in path constructors.
However, since libc++'s filesystem only really supports POSIX like systems,
there are no real changes needed. This patch simply adds the format enum
and then ignores it when it's passed to constructors.
llvm-svn: 329031
Rafael Espindola [Mon, 2 Apr 2018 23:20:30 +0000 (23:20 +0000)]
Align stubs for external and common global variables to pointer size.
This patch fixes PR36885: clang++ generates unaligned stub symbol
holding a pointer.
Patch by Rahul Chaudhry!
llvm-svn: 329030
Eric Christopher [Mon, 2 Apr 2018 23:17:55 +0000 (23:17 +0000)]
Remove llvm-mcmarkup.
It was never used and I've checked with the original authors.
llvm-svn: 329029
Eric Fiselier [Mon, 2 Apr 2018 23:03:41 +0000 (23:03 +0000)]
Implement filesystem NB comments, relative paths, and related issues.
This is a fairly large patch that implements all of the filesystem NB comments
and the relative paths changes (ex. adding weakly_canonical). These issues
and papers are all interrelated so their implementation couldn't be split up
nicely.
This patch upgrades <experimental/filesystem> to match the C++17 spec and not
the published experimental TS spec. Some of the changes in this patch are both
API and ABI breaking, however libc++ makes no guarantee about stability for
experimental implementations.
The major changes in this patch are:
* Implement NB comments for filesystem (P0492R2), including:
* Implement `perm_options` enum as part of NB comments, and update the
`permissions` function to match.
* Implement changes to `remove_filename` and `replace_filename`
* Implement changes to `path::stem()` and `path::extension()` which support
splitting examples like `.profile`.
* Change path iteration to return an empty path instead of '.' for trailing
separators.
* Change `operator/=` to handle absolute paths on the RHS.
* Change `absolute` to no longer accept a current path argument.
* Implement relative paths according to NB comments (P0219r1)
* Combine `path.cpp` and `operations.cpp` since some path functions require
access to the operations internals, and some fs operations require access
to the path parser.
llvm-svn: 329028
Reid Kleckner [Mon, 2 Apr 2018 22:49:44 +0000 (22:49 +0000)]
[InstCombine] Don't strip function type casts from musttail calls
Summary:
The cast simplifications that instcombine does here do not make any
attempt to obey the verifier rules for musttail calls. Therefore we have
to disable them.
Reviewers: efriedma, majnemer, pcc
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D45186
llvm-svn: 329027
Aaron Smith [Mon, 2 Apr 2018 22:34:35 +0000 (22:34 +0000)]
[lit] One more try at fixing TestRunner.py for D43165
llvm-svn: 329026
Saleem Abdulrasool [Mon, 2 Apr 2018 22:25:50 +0000 (22:25 +0000)]
CodeGenCXX: support PreserveMostCC in MS ABI
Microsoft has reserved 'U' for the PreserveMostCC which is used in the
swift runtime. Add support for this. This allows the swift runtime to
be built for Windows again.
llvm-svn: 329025
Reid Kleckner [Mon, 2 Apr 2018 22:19:42 +0000 (22:19 +0000)]
[lit] Attempt to fix builtin diff code for Python 2
llvm-svn: 329024
Volodymyr Sapsai [Mon, 2 Apr 2018 22:09:57 +0000 (22:09 +0000)]
[libcxx] Disable testing with system lib for 2 tests verifying debug mode.
llvm-svn: 329023
Aaron Smith [Mon, 2 Apr 2018 22:08:56 +0000 (22:08 +0000)]
[lit] Fix problem in how Python versions open files with different encodings
Reapply D43165 which was reverted because of different versions of python failing.
The one line fix for the different python versions was commited at the same time
that D43165 was reverted. If this change is giving you issues then get in touch
with your python version and we will fix it.
llvm-svn: 329022
Harlan Haskins [Mon, 2 Apr 2018 21:45:35 +0000 (21:45 +0000)]
Fix Go IR test for changes in DIBuilder API
llvm-svn: 329021
Aaron Smith [Mon, 2 Apr 2018 21:44:51 +0000 (21:44 +0000)]
[lit] Use io.open to compare two files since it supports different encodings while older versions of open do not
llvm-svn: 329020
Lang Hames [Mon, 2 Apr 2018 21:39:39 +0000 (21:39 +0000)]
[ORC] Add a debugging check to catch dangling references to SymbolStringPools.
llvm-svn: 329019
Lang Hames [Mon, 2 Apr 2018 21:34:48 +0000 (21:34 +0000)]
[ORC] Fix a think-o: ExecutionSession should create a SymbolStringPool if none
is passed in.
llvm-svn: 329018
Reid Kleckner [Mon, 2 Apr 2018 21:33:46 +0000 (21:33 +0000)]
Revert r329012 "[lit] Fix problem in how Python versions open files with different encodings"
This doesn't work with Python 2. See this build-in-progress:
http://lab.llvm.org:8011/builders/clang-x86-windows-msvc2015/builds/11105
llvm-svn: 329017
Rong Xu [Mon, 2 Apr 2018 21:24:07 +0000 (21:24 +0000)]
[profile] Move newly added test to the supported platform
Move test/profile/instrprof-value-merge.c from r328987 to Linux directory.
llvm-svn: 329016
Reid Kleckner [Mon, 2 Apr 2018 21:23:16 +0000 (21:23 +0000)]
Treat inlining a notail call as a regular, non-tail call
Otherwise, we end up inlining a musttail call into a non-tail position,
which breaks verifier invariants.
Fixes PR31014
llvm-svn: 329015
Rafael Espindola [Mon, 2 Apr 2018 21:11:13 +0000 (21:11 +0000)]
Define TrapInst for ppc64.
This is nice for testing since it is the first TrapInst whose bytes
are not all the same.
llvm-svn: 329014
Lang Hames [Mon, 2 Apr 2018 20:57:56 +0000 (20:57 +0000)]
[ORC] Create a new SymbolStringPool by default in ExecutionSession constructor.
This makes the common case of constructing an ExecutionSession tidier.
llvm-svn: 329013
Aaron Smith [Mon, 2 Apr 2018 20:57:06 +0000 (20:57 +0000)]
[lit] Fix problem in how Python versions open files with different encodings
Summary:
This issue was found when running the clang unit test on Windows. Python 3.x cannot open some of the files that the tests are using with a simple open because of their encoding. Python 2.7+ and Python 3.x both support io.open which allows for an encoding to be specified.
This change will determine whether two files being compared should be opened (and then compared) as text or binary and whether to use utf-8 or the default encoding before proceeding with a line-by-line comparison.
Patch by Stella Stamenova!
Reviewers: zturner, llvm-commits, rnk, MaggieYi
Reviewed By: zturner
Subscribers: asmith, MatzeB, stella.stamenova, delcypher, llvm-commits
Differential Revision: https://reviews.llvm.org/D43165
llvm-svn: 329012
Sanjay Patel [Mon, 2 Apr 2018 20:37:40 +0000 (20:37 +0000)]
[InstCombine] add folds for icmp + sub (PR36969)
(A - B) >u A --> A <u B
C <u (C - D) --> C <u D
https://rise4fun.com/Alive/e7j
Name: ugt
%sub = sub i8 %x, %y
%cmp = icmp ugt i8 %sub, %x
=>
%cmp = icmp ult i8 %x, %y
Name: ult
%sub = sub i8 %x, %y
%cmp = icmp ult i8 %x, %sub
=>
%cmp = icmp ult i8 %x, %y
This should fix:
https://bugs.llvm.org/show_bug.cgi?id=36969
llvm-svn: 329011
Sanjay Patel [Mon, 2 Apr 2018 20:23:54 +0000 (20:23 +0000)]
[InstCombine] add tests for icmp (sub x, y), x (PR36969); NFC
llvm-svn: 329010
Reid Kleckner [Mon, 2 Apr 2018 20:20:33 +0000 (20:20 +0000)]
[MS] Emit vftable thunks for functions with incomplete prototypes
Summary:
The following class hierarchy requires that we be able to emit a
this-adjusting thunk for B::foo in C's vftable:
struct Incomplete;
struct A {
virtual A* foo(Incomplete p) = 0;
};
struct B : virtual A {
void foo(Incomplete p) override;
};
struct C : B { int c; };
This TU is valid, but lacks a definition of 'Incomplete', which makes it
hard to build a thunk for the final overrider, B::foo.
Before this change, Clang gives up attempting to emit the thunk, because
it assumes that if the parameter types are incomplete, it must be
emitting the thunk for optimization purposes. This is untrue for the MS
ABI, where the implementation of B::foo has no idea what thunks C's
vftable may require. Clang needs to emit the thunk without necessarily
having access to the complete prototype of foo.
This change makes Clang emit a musttail variadic call when it needs such
a thunk. I call these "unprototyped" thunks, because they only prototype
the "this" parameter, which must always come first in the MS C++ ABI.
These thunks work, but they create ugly LLVM IR. If the call to the
thunk is devirtualized, it will be a call to a bitcast of a function
pointer. Today, LLVM cannot inline through such a call, but I want to
address that soon, because we also use this pattern for virtual member
pointer thunks.
This change also implements an old FIXME in the code about reusing the
thunk's computed CGFunctionInfo as much as possible. Now we don't end up
computing the thunk's mangled name and arranging it's prototype up to
around three times.
Fixes PR25641
Reviewers: rjmccall, rsmith, hans
Subscribers: Prazek, cfe-commits
Differential Revision: https://reviews.llvm.org/D45112
llvm-svn: 329009
Douglas Yung [Mon, 2 Apr 2018 20:05:05 +0000 (20:05 +0000)]
Another attempt to fix papertrail-warnings.test on Windows bots by making expected message less case sensitive.
llvm-svn: 329008
Reid Kleckner [Mon, 2 Apr 2018 20:00:39 +0000 (20:00 +0000)]
Fix some DenseMap use-after-rehash bugs and hoist MethodVFTableLocation
This re-lands r328845 with fixes for crbug.com/827810.
The initial motiviation was to hoist MethodVFTableLocation to global
scope so it could be forward declared.
In this patch, I noticed that MicrosoftVTableContext uses some risky
patterns. It has methods that return references to data stored in
DenseMaps. I've made some of them return by value for trivial structs
and I've moved some things into separate allocations.
llvm-svn: 329007
Sean Fertile [Mon, 2 Apr 2018 19:47:21 +0000 (19:47 +0000)]
[PPC64] Minor changes for Plt relocations.
The Plt relative relocations are R_PPC64_JMP_SLOT in the V2 abi, and we only
reserve 2 double words instead of 3 at the start of the array of PLT entries for
lazy linking.
Differential Revision: https://reviews.llvm.org/D44951
llvm-svn: 329006
Joel E. Denny [Mon, 2 Apr 2018 19:43:34 +0000 (19:43 +0000)]
[Attr] [NFC] Revert accidental change from r327405
llvm-svn: 329005
Sean Fertile [Mon, 2 Apr 2018 19:17:12 +0000 (19:17 +0000)]
[PPC64] Write plt stubs for ElfV2 abi
Add the default version of a plt stub for the V2 Elf abi.
Differential Revision: https://reviews.llvm.org/D44850
llvm-svn: 329004
Harlan Haskins [Mon, 2 Apr 2018 19:11:44 +0000 (19:11 +0000)]
Fix header mismatch in DIBuilder Type APIs
Some of the headers changed slightly, and the accompanying
implementation didn't change. This caused a silent failure.
llvm-svn: 329003
Zachary Turner [Mon, 2 Apr 2018 18:35:21 +0000 (18:35 +0000)]
[llvm-pdbutil] Add an export subcommand.
This command can dump the binary contents of a stream to a file.
This is useful when you want to do side-by-side comparisons of
a specific stream from two PDBs to examine the differences between
them. You can export both of them to a file, then open them up
side by side in a hex editor (for example), so as to eliminate any
differences that might arise from the contents being on different
blocks in the PDB.
In subsequent patches I plan to improve the "explain" subcommand
so that you can explain the contents of a binary file that isn't
necessarily a full PDB, but one of these dumped streams, by telling
the subcommand how to interpret the contents.
llvm-svn: 329002
Eric Christopher [Mon, 2 Apr 2018 18:33:47 +0000 (18:33 +0000)]
Temporarily revert r328404:
commit
519b97132a4c960e8dedbfe4290d86970d92e995
Author: Richard Trieu <rtrieu@google.com>
Date: Sat Mar 24 00:52:44 2018 +0000
[ODRHash] Support pointer and reference types.
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@328404
91177308-0d34-0410-b5e6-
96231b3b80d8
As it's breaking some tests. I've communicated with Richard offline about testcases.
llvm-svn: 329001
Richard Smith [Mon, 2 Apr 2018 18:29:44 +0000 (18:29 +0000)]
Add -fclang-abi-compat=6 flag for upcoming ABI changes.
llvm-svn: 329000
Richard Smith [Mon, 2 Apr 2018 18:29:43 +0000 (18:29 +0000)]
Add helper to determine if a field is a zero-length bitfield.
llvm-svn: 328999
Nico Weber [Mon, 2 Apr 2018 18:13:05 +0000 (18:13 +0000)]
Remove unused PACKAGE_TARNAME; looks like a remnant of the old autoconf build.
llvm-svn: 328998
Mike Edwards [Mon, 2 Apr 2018 18:02:36 +0000 (18:02 +0000)]
Revert r328932 as it caused Windows and MacOS bot failures.
http://green.lab.llvm.org/green/job/clang-stage1-configure-RA/43991/
llvm-svn: 328997
Nico Weber [Mon, 2 Apr 2018 17:52:19 +0000 (17:52 +0000)]
Remove some unused cmake standard library includes.
llvm-svn: 328996
Nico Weber [Mon, 2 Apr 2018 17:45:47 +0000 (17:45 +0000)]
Remove more feature test macros that became unused after r328989.
llvm-svn: 328995
Yaxun Liu [Mon, 2 Apr 2018 17:38:24 +0000 (17:38 +0000)]
[CUDA] Let device-side shared variables be initialized with undef
CUDA shared variable should be initialized with undef.
Patch by Greg Rodgers.
Revised and lit test added by Yaxun Liu.
Differential Revision: https://reviews.llvm.org/D44985
llvm-svn: 328994
Gor Nishanov [Mon, 2 Apr 2018 17:35:37 +0000 (17:35 +0000)]
[coroutines] Add __builtin_coro_noop => llvm.coro.noop
A recent addition to Coroutines TS (https://wg21.link/p0913) adds a pre-defined
coroutine noop_coroutine that does nothing. To implement this feature, we implemented
an llvm.coro.noop intrinsic that returns a coroutine handle to a coroutine that
does nothing when resumed or destroyed.
This patch adds a builtin __builtin_coro_noop() that maps to llvm.coro.noop intrinsic.
Related llvm change: https://reviews.llvm.org/D45114
llvm-svn: 328993
Nico Weber [Mon, 2 Apr 2018 17:32:48 +0000 (17:32 +0000)]
Remove HAVE_LIBPSAPI, HAVE_SHELL32.
These used to be set in the old autoconf build, but the cmake build has had a
"TODO: actually check for these" comment since it was checked in, and they
were set to 1 on mingw unconditionally. It seems safe to say that they always
exist under mingw, so just remove them and assume they're set exactly when on
mingw (with msvc, we use `pragma comment` instead of linking these via flags).
llvm-svn: 328992
Rong Xu [Mon, 2 Apr 2018 17:27:38 +0000 (17:27 +0000)]
[DeadArgumentElim] Clone function level metadatas
Some Function level metadatas, such as function entry count, are not cloned in
DeadArgumentElim. This happens a lot in lto/thinlto because of DeadArgumentElim
after internalization.
This patch clones the metadatas in the original function to the new function.
Differential Revision: https://reviews.llvm.org/D44127
llvm-svn: 328991
Greg Clayton [Mon, 2 Apr 2018 17:20:21 +0000 (17:20 +0000)]
Added a BSD archive tool.
This is a combination stand alone BSD archive tool that can dump BSD archives:
% bsd.py /path/to/foo.a
Search archives for an object file:
% bsd.py --object foo.o bar.a
Dump the symbol definitions found in the __.SYMDEF objects:
% bsd.py --symdef bar.a
Find symbols by name that are listed in the __.SYMDEF objects:
% bsd.py --symbol _Z123 bar.a
Extract objects from BSD archives:
% bsd.py --object foo.o bar.a --extract
% bsd.py --object foo.o bar.a --extract --outfile /tmp/foo.o
% bsd.py --object foo.o bar.a --extract --mtime 0x1234556
It also has installs a new LLDB command line command when imported into LLDB:
(lldb) command script import ~/Dropbox/bin/bsd.py
The "verify-debug-map-objects" command has been installed, type "help verify-debug-map-objects" for detailed help.
(lldb) verify-debug-map-objects a.out
This will iterate through all object files and verify the modification times match for any .o files, it will verify any .o files from BSD archives are found and have matching modification times and print out errors if any are found.
llvm-svn: 328990
Nico Weber [Mon, 2 Apr 2018 17:17:29 +0000 (17:17 +0000)]
Remove HAVE_DIRENT_H.
The autoconf manual: "This macro is obsolescent, as all current systems with
directory libraries have <dirent.h>. New programs need not use this macro."
llvm-svn: 328989
Dmitry Preobrazhensky [Mon, 2 Apr 2018 17:09:20 +0000 (17:09 +0000)]
[AMDGPU][MC][GFX9] Added instructions v_cvt_norm_*16_f16, v_sat_pk_u8_i16
See bug 36847: https://bugs.llvm.org/show_bug.cgi?id=36847
Differential Revision: https://reviews.llvm.org/D45097
Reviewers: artem.tamazov, arsenm, timcorringham
llvm-svn: 328988
Rong Xu [Mon, 2 Apr 2018 16:57:00 +0000 (16:57 +0000)]
[profile] Fix value profile runtime merging issues
This patch fixes the following issues:
(1) The strong definition of the merge hook function was not working which
breaks the online value profile merging. This patch removes the weak
attribute of VPMergeHook and assigns the value dynamically.
(2) Truncate the proifle file so that we don't have garbage data at the end of
the file.
(3) Add new __llvm_profile_instrument_target_value() interface to do the value
profile update in batch. This is needed as the original incremental by 1
in __llvm_profile_instrument_target() is too slow for online merge.
Differential Revision: https://reviews.llvm.org/D44847
llvm-svn: 328987
Gor Nishanov [Mon, 2 Apr 2018 16:55:12 +0000 (16:55 +0000)]
[coroutines] Add support for llvm.coro.noop intrinsics
Summary:
A recent addition to Coroutines TS (https://wg21.link/p0913) adds a pre-defined coroutine noop_coroutine that does nothing.
To implement this feature, we implemented an llvm.coro.noop intrinsic that returns a coroutine handle to a coroutine that does nothing when resumed or destroyed.
Reviewers: EricWF, modocache, rnk, lewissbaker
Reviewed By: modocache
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45114
llvm-svn: 328986
Davide Italiano [Mon, 2 Apr 2018 16:50:54 +0000 (16:50 +0000)]
[Core] Grab-bag of improvements for Scalar.
Remove Scalar::Cast.
It was noted on the list that this method is unused. So, this patch
removes it.
Fix Scalar::Promote for most integer types
This fixes promotion of most integer types (128- and 256-bit types are
handled in a subsequent patch) to floating-point types. Previously
promotion was done bitwise, where value preservation is correct.
Fix Scalar::Promote for 128- and 256-bit integer types
This patch fixes the behavior of Scalar::Promote when trying to
perform a binary operation involving a 128- or 256-bit integer type
and a floating-point type. Now, the integer is cast to the floating
point type for the operation.
Patch by Tom Tromey!
Differential Revision: https://reviews.llvm.org/D44907
llvm-svn: 328985
Frederic Riss [Mon, 2 Apr 2018 16:18:32 +0000 (16:18 +0000)]
Support template template parameters
Summary:
We would fail to resolve (and thus display the value of) any
templated type which contained a template template argument even
though we don't really use template arguments.
This patch adds minimal support for template template arguments,
but I doubt we need any more than that.
Reviewers: clayborg, jingham
Subscribers: JDevlieghere, lldb-commits
Differential Revision: https://reviews.llvm.org/D44613
llvm-svn: 328984
Dmitry Preobrazhensky [Mon, 2 Apr 2018 16:10:25 +0000 (16:10 +0000)]
[AMDGPU][MC][GFX9] Added s_atomic_* and s_buffer_atomic_* instructions
Fixed a bug which caused Tablegen crash.
See bug 36837: https://bugs.llvm.org/show_bug.cgi?id=36837
Differential Revision: https://reviews.llvm.org/D45085
Reviewers: artem.tamazov, arsenm, timcorringham
llvm-svn: 328983
Sean Fertile [Mon, 2 Apr 2018 15:42:07 +0000 (15:42 +0000)]
[PPC] Add a test for toc-relative access on ppc64le.
Adds a simple test for accessing a local global variable in the ElfV2 abi.
Checks that the toc base used is the expected offset from the .TOC. symbol,
and that the offsets for the global are calculated relative to the toc base.
llvm-svn: 328982