OSDN Git Service
Simon Pilgrim [Thu, 14 Jun 2018 09:35:00 +0000 (09:35 +0000)]
[CostModel] Recognise REVERSE shuffle mask if the elements come from the second src
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334698
91177308-0d34-0410-b5e6-
96231b3b80d8
Clement Courbet [Thu, 14 Jun 2018 06:57:52 +0000 (06:57 +0000)]
[llvm-exegesis] Use BenchmarkResult::Instructions instead of OpcodeName
Summary:
Get rid of OpcodeName.
To remove the opcode name from an old file:
```
cat old_file | sed '/opcode_name.*/d'
```
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48121
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334691
91177308-0d34-0410-b5e6-
96231b3b80d8
Hiroshi Inoue [Thu, 14 Jun 2018 05:41:49 +0000 (05:41 +0000)]
[NFC] fix trivial typos in comments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334687
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 14 Jun 2018 03:16:58 +0000 (03:16 +0000)]
[x86] fix mappings of cvttp2si/cvttp2ui x86 intrinsics to x86-specific nodes and isel patterns (PR37551)
Summary:
The tests in:
https://bugs.llvm.org/show_bug.cgi?id=37751
...show miscompiles because we wrongly mapped and folded x86-specific intrinsics into generic DAG nodes.
This patch corrects the mappings in X86IntrinsicsInfo.h and adds isel matching corresponding to the new patterns. The complete tests for the failure cases should be in avx-cvttp2si.ll and sse-cvttp2si.ll and avx512-cvttp2i.ll
Reviewers: RKSimon, gbedwell, spatel
Reviewed By: spatel
Subscribers: mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D47993
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334685
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Davis [Thu, 14 Jun 2018 01:20:18 +0000 (01:20 +0000)]
[llvm-mca] Introduce the ExecuteStage (was originally the Scheduler class).
Summary: This patch transforms the Scheduler class into the ExecuteStage. Most of the logic remains.
Reviewers: andreadb, RKSimon, courbet
Reviewed By: andreadb
Subscribers: mgorny, javed.absar, tschuett, gbedwell, llvm-commits
Differential Revision: https://reviews.llvm.org/D47246
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334679
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Wed, 13 Jun 2018 22:30:47 +0000 (22:30 +0000)]
AMDGPU/GlobalISel: Implement select() for 32-bit G_FADD and G_FMUL
Reviewers: arsenm, nhaehnle
Reviewed By: arsenm
Subscribers: kzhuravl, wdng, yaxunl, rovka, kristof.beyls, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D46171
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334665
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Wed, 13 Jun 2018 21:24:19 +0000 (21:24 +0000)]
Revert "Enable ThreadPool to queue tasks that return values."
This is failing to compile when LLVM_ENABLE_THREADS is false,
and the fix is not immediately obvious, so reverting while I look
into it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334658
91177308-0d34-0410-b5e6-
96231b3b80d8
Francis Visoiu Mistrih [Wed, 13 Jun 2018 21:03:56 +0000 (21:03 +0000)]
Reland: [Timers] Use the pass argument name for JSON keys in time-passes
When using clang --save-stats -mllvm -time-passes, both timers and stats
end up in the same json file.
We could end up with things like:
{
"asm-printer.EmittedInsts": 1,
"time.pass.Virtual Register Map.wall": 2.
9015541076660156e-04,
"time.pass.Virtual Register Map.user": 2.
0500000000000379e-04,
"time.pass.Virtual Register Map.sys": 8.
5000000000001741e-05,
}
This patch makes use of the pass argument name (if available) in the
JSON key to end up with things like:
{
"asm-printer.EmittedInsts": 1,
"time.pass.virtregmap.wall": 2.
9015541076660156e-04,
"time.pass.virtregmap.user": 2.
0500000000000379e-04,
"time.pass.virtregmap.sys": 8.
5000000000001741e-05,
}
This also helps avoiding to write another JSON printer to handle all the
cases that we could have in our pass names.
Fixed test instead of adding a new one originally from r334649.
Differential Revision: https://reviews.llvm.org/D48109
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334657
91177308-0d34-0410-b5e6-
96231b3b80d8
Florian Hahn [Wed, 13 Jun 2018 20:59:53 +0000 (20:59 +0000)]
[TableGen] Make getOnlyTree return a const ref (NFC)
This avoids some unnecessary copies of shared_ptrs.
Those changes were suggested post-commit for D47463.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334656
91177308-0d34-0410-b5e6-
96231b3b80d8
George Karpenkov [Wed, 13 Jun 2018 20:48:53 +0000 (20:48 +0000)]
Update comments of CheckedArithmetic API based on Philip Reames feedback.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334655
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Wed, 13 Jun 2018 20:47:21 +0000 (20:47 +0000)]
[WinASan] Don't instrument globals in sections containing '$'
Such globals are very likely to be part of a sorted section array, such
the .CRT sections used for dynamic initialization. The uses its own
sorted sections called ATL$__a, ATL$__m, and ATL$__z. Instead of special
casing them, just look for the dollar sign, which is what invokes linker
section sorting for COFF.
Avoids issues with ASan and the ATL uncovered after we started
instrumenting comdat globals on COFF.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334653
91177308-0d34-0410-b5e6-
96231b3b80d8
Francis Visoiu Mistrih [Wed, 13 Jun 2018 20:44:02 +0000 (20:44 +0000)]
Revert r334649 "[Timers] Use the pass argument name for JSON keys in time-passes"
This reverts commit r334649.
This breaks a test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334651
91177308-0d34-0410-b5e6-
96231b3b80d8
Francis Visoiu Mistrih [Wed, 13 Jun 2018 20:09:59 +0000 (20:09 +0000)]
[Timers] Use the pass argument name for JSON keys in time-passes
When using clang --save-stats -mllvm -time-passes, both timers and stats
end up in the same json file.
We could end up with things like:
{
"asm-printer.EmittedInsts": 1,
"time.pass.Virtual Register Map.wall": 2.
9015541076660156e-04,
"time.pass.Virtual Register Map.user": 2.
0500000000000379e-04,
"time.pass.Virtual Register Map.sys": 8.
5000000000001741e-05,
}
This patch makes use of the pass argument name (if available) in the
JSON key to end up with things like:
{
"asm-printer.EmittedInsts": 1,
"time.pass.virtregmap.wall": 2.
9015541076660156e-04,
"time.pass.virtregmap.user": 2.
0500000000000379e-04,
"time.pass.virtregmap.sys": 8.
5000000000001741e-05,
}
This also helps avoiding to write another JSON printer to handle all the
cases that we could have in our pass names.
Differential Revision: https://reviews.llvm.org/D48109
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334649
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 13 Jun 2018 20:03:42 +0000 (20:03 +0000)]
[X86] Move RCPSSr_Int, RSQRTSSr_Int, SQRTSDr_Int, SQRTSSr_Int to the correct load folding table.
They were in the operand 1 folding table, but their foldable operand is operand 2.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334648
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Wed, 13 Jun 2018 19:37:41 +0000 (19:37 +0000)]
Add missing #include.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334644
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Wed, 13 Jun 2018 19:29:16 +0000 (19:29 +0000)]
Enable ThreadPool to support tasks that return values.
Previously ThreadPool could only queue async "jobs", i.e. work
that was done for its side effects and not for its result. It's
useful occasionally to queue async work that returns a value.
From an API perspective, this is very intuitive. The previous
API just returned a shared_future<void>, so all we need to do is
make it return a shared_future<T>, where T is the type of value
that the operation returns.
Making this work required a little magic, but ultimately it's not
too bad. Instead of keeping a shared queue<packaged_task<void()>>
we just keep a shared queue<unique_ptr<TaskBase>>, where TaskBase
is a class with a pure virtual execute() method, then have a
templated derived class that stores a packaged_task<T()>. Everything
else works out pretty cleanly.
Differential Revision: https://reviews.llvm.org/D48115
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334643
91177308-0d34-0410-b5e6-
96231b3b80d8
Stanislav Mekhanoshin [Wed, 13 Jun 2018 18:52:54 +0000 (18:52 +0000)]
[AMDGPU] Corrected computeKnownBits for V_PERM_B32
Differential Revision: https://reviews.llvm.org/D48133
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334640
91177308-0d34-0410-b5e6-
96231b3b80d8
George Karpenkov [Wed, 13 Jun 2018 18:32:02 +0000 (18:32 +0000)]
Add checkMulAdd helper function to CheckedArithmetic
Multiplication followed by addition
(https://en.wikipedia.org/wiki/Multiply–accumulate_operation) is a
sufficiently common use-case to warrant a separate helper.
Differential Revision: https://reviews.llvm.org/D48138
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334635
91177308-0d34-0410-b5e6-
96231b3b80d8
George Karpenkov [Wed, 13 Jun 2018 18:31:43 +0000 (18:31 +0000)]
Change checked arithmetic functions API to return Optional
Returning optional is much safer.
The previous API had potential to cause use of undefined variables, if
the value passed by pointer was accidentally read afterwards.
Differential Revision: https://reviews.llvm.org/D48137
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334634
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrea Di Biagio [Wed, 13 Jun 2018 18:30:14 +0000 (18:30 +0000)]
[llvm-mca] Fixed a bug in the logic that checks if a memory operation is ready to execute.
Fixes PR37790.
In some (very rare) cases, the LSUnit (Load/Store unit) was wrongly marking a
load (or store) as "ready to execute" effectively bypassing older memory barrier
instructions.
To reproduce this bug, the memory barrier must be the first instruction in the
input assembly sequence, and it doesn't have to perform any register writes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334633
91177308-0d34-0410-b5e6-
96231b3b80d8
Jordan Rose [Wed, 13 Jun 2018 18:21:47 +0000 (18:21 +0000)]
[CMake] Handle 'libtool' being at a path with spaces in it.
This can happen on macOS if the user's Xcode is at a path with spaces in it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334632
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Wed, 13 Jun 2018 18:03:14 +0000 (18:03 +0000)]
LTO: Keep file handles open for memory mapped files.
On Windows we've observed that if you open a file, write to it, map it into
memory and close the file handle, the contents of the memory mapping can
sometimes be incorrect. That was what we did when adding an entry to the
ThinLTO cache using the TempFile and MemoryBuffer classes, and it was causing
intermittent build failures on Chromium's ThinLTO bots on Windows. More
details are in the associated Chromium bug (crbug.com/786127).
We can prevent this from happening by keeping a handle to the file open while
the mapping is active. So this patch changes the mapped_file_region class to
duplicate the file handle when mapping the file and close it upon unmapping it.
One gotcha is that the file handle that we keep open must not have been
created with FILE_FLAG_DELETE_ON_CLOSE, as otherwise the operating system
will prevent other processes from opening the file. We can achieve this
by avoiding the use of FILE_FLAG_DELETE_ON_CLOSE altogether. Instead,
we use SetFileInformationByHandle with FileDispositionInfo to manage the
delete-on-close bit. This lets us remove the hack that we used to use to
clear the delete-on-close bit on a file opened with FILE_FLAG_DELETE_ON_CLOSE.
A downside of using SetFileInformationByHandle/FileDispositionInfo as
opposed to FILE_FLAG_DELETE_ON_CLOSE is that it prevents us from using
CreateFile to open the file while the flag is set, even within the same
process. This doesn't seem to matter for almost every client of TempFile,
except for LockFileManager, which calls sys::fs::create_link to create a
hard link from the lock file, and in the process of doing so tries to open
the file. To prevent this change from breaking LockFileManager I changed it
to stop using TempFile by effectively reverting r318550.
Differential Revision: https://reviews.llvm.org/D48051
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334630
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 13 Jun 2018 17:55:13 +0000 (17:55 +0000)]
[X86] Add one more intrinsic and test cases to avx512-cvttp2i.ll.
spatel noticed it was missing in D47993.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334629
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Wed, 13 Jun 2018 17:51:27 +0000 (17:51 +0000)]
IR: fix documentation markup
Use `\brief` instead of `\Brief`. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334627
91177308-0d34-0410-b5e6-
96231b3b80d8
Yaxun Liu [Wed, 13 Jun 2018 17:31:51 +0000 (17:31 +0000)]
[AMDGPU] Change enqueue kernel handle type
Currently the handle type is a global pointer which holds 8 bytes.
We need a larger type which hold 16 bytes, therefore change it
to [i64 x 2].
Differential Revision: https://reviews.llvm.org/D48094
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334625
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 13 Jun 2018 17:12:11 +0000 (17:12 +0000)]
[CostModel][X86] Test showing failure to recognise REVERSE shuffle mask if the elements come from the second src
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334623
91177308-0d34-0410-b5e6-
96231b3b80d8
Dmitry Preobrazhensky [Wed, 13 Jun 2018 17:02:03 +0000 (17:02 +0000)]
[AMDGPU][MC] Enabled parsing of relocations on VALU instructions
See bug 37566: https://bugs.llvm.org/show_bug.cgi?id=37566
Reviewers: artem.tamazov, arsenm, nhaehnle
Differential Revision: https://reviews.llvm.org/D47884
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334622
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 13 Jun 2018 16:52:02 +0000 (16:52 +0000)]
[CostModel] Recognise BROADCAST shuffle mask if the elements come from the second src
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334620
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrea Di Biagio [Wed, 13 Jun 2018 16:33:52 +0000 (16:33 +0000)]
Revert: [llvm-mca] Flush the output stream before we start the analysis of a new code region. NFC
Not sure why, but it breaks buildbot clang-cmake-armv8-full.
It causes a failure in TEST 'Xray-armhf-linux :: TestCases/Posix/profiling-single-threaded.cc'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334617
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 13 Jun 2018 16:33:42 +0000 (16:33 +0000)]
[CostModel][X86] Test showing failure to recognise BROADCAST shuffle mask if the elements come from the second src
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334616
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrea Di Biagio [Wed, 13 Jun 2018 15:43:56 +0000 (15:43 +0000)]
[llvm-mca] Flush the output stream before we start the analysis of a new code region. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334610
91177308-0d34-0410-b5e6-
96231b3b80d8
Dmitry Preobrazhensky [Wed, 13 Jun 2018 15:32:46 +0000 (15:32 +0000)]
[AMDGPU][MC][GFX8][GFX9] Allow LDS direct reads for BUFFER_LOAD_DWORDX2/X3/X4
See bug 37653: https://bugs.llvm.org/show_bug.cgi?id=37653
Reviewers: artem.tamazov, arsenm
Differential Revision: https://reviews.llvm.org/D47885
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334609
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 13 Jun 2018 15:22:48 +0000 (15:22 +0000)]
[DAGCombiner] remove hasOneUse() check from fadd constants transform
We're constant folding here, so we shouldn't check uses. This matches
the IR optimizer behavior.
The x86 test shows the expected win. The AArch64 test shows something
else. This only seems to happen if the "generic" AArch64 CPU model is
used by MachineCombiner, so I'll file a bug report to follow-up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334608
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Wed, 13 Jun 2018 15:06:37 +0000 (15:06 +0000)]
AMDGPU: Move isSDNodeSourceOfDivergence() implementation to SITargetLowering
Summary:
The code that handles ISD:Register and ISD::CopyFromReg assumes
the target is amdgcn, so this is broken on r600. We don't
need this analysis on r600 anyway so we can safely move
it to SITargetLowering.
Reviewers: alex-t, arsenm, nhaehnle
Reviewed By: arsenm
Subscribers: msearles, kzhuravl, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D46298
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334607
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 13 Jun 2018 15:01:07 +0000 (15:01 +0000)]
[x86] add test for fadd with more than one use; NFC
The equivalent AArch64 test added at rL334556 isn't showing
the expected output from the DAGCombiner code change that
would fix this example. That's a machine combiner bug from
what I see.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334605
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron McInally [Wed, 13 Jun 2018 14:32:12 +0000 (14:32 +0000)]
[FPEnv] Expand constrained FP operations
Add a helper function to expand constrained FP operations as needed.
Note that the Strict POWI operation is not handled in this patch since
the format is slightly different from the others.
Differential Revision: https://reviews.llvm.org/D47491
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334603
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Wed, 13 Jun 2018 14:29:26 +0000 (14:29 +0000)]
Do not enforce absolute path argv0 in windows
Even if we support no-canonical-prefix on
clang-cl(https://reviews.llvm.org/D47480), argv0 becomes absolute path
in clang-cl and that embeds absolute path in /showIncludes.
This patch removes such full path normalization from InitLLVM on
windows, and that removes absolute path from clang-cl output
(obj/stdout/stderr) when debug flag is disabled.
Patch by Takuto Ikuta!
Differential Revision https://reviews.llvm.org/D47578
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334602
91177308-0d34-0410-b5e6-
96231b3b80d8
Guillaume Chatelet [Wed, 13 Jun 2018 14:07:36 +0000 (14:07 +0000)]
[llvm-exegesis] Fix buildbot - power was using native target for X86.
Reviewers: courbet
Reviewed By: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48125
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334601
91177308-0d34-0410-b5e6-
96231b3b80d8
Guillaume Chatelet [Wed, 13 Jun 2018 13:53:56 +0000 (13:53 +0000)]
[llvm-exegesis] Fix failing assert when creating Snippet for LAHF.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48123
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334599
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Wed, 13 Jun 2018 13:49:06 +0000 (13:49 +0000)]
Revert "Improve handling of COPY instructions with identical value numbers"
This reverts r334594, it breaks buildbots and fails with expensive checks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334598
91177308-0d34-0410-b5e6-
96231b3b80d8
Guillaume Chatelet [Wed, 13 Jun 2018 13:24:41 +0000 (13:24 +0000)]
[llvm-exegesis] Cleaner design without mutable data.
Summary: Previous design was relying on the 'mutate' keyword and was quite confusing. This version separate mutable from immutable data and makes it clearer what changes and what doesn't.
Reviewers: courbet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48020
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334596
91177308-0d34-0410-b5e6-
96231b3b80d8
Zoran Jovanovic [Wed, 13 Jun 2018 12:51:37 +0000 (12:51 +0000)]
[mips][microMIPS] Extending size reduction pass with LWP and SWP
Author: milena.vujosevic.janicic
Reviewers: sdardis
The patch extends size reduction pass for MicroMIPS.
It introduces reduction of two instructions into one instruction:
Two SW instructions are transformed into one SWP instrucition.
Two LW instructions are transformed into one LWP instrucition.
Differential Revision: https://reviews.llvm.org/D39115
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334595
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Wed, 13 Jun 2018 12:47:17 +0000 (12:47 +0000)]
Improve handling of COPY instructions with identical value numbers
Differential Revision: https://reviews.llvm.org/D48102
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334594
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 13 Jun 2018 12:28:32 +0000 (12:28 +0000)]
[x86] eliminate even more sign-bit tests with vector select
This shortcoming was noted in D47330, and the test diffs show we already
had other examples where we failed to fold to a SHRUNKBLEND:
/// Dynamic (non-constant condition) vector blend where only the sign bits
/// of the condition elements are used. This is used to enforce that the
/// condition mask is not valid for generic VSELECT optimizations.
This patch implements an idea from D48043 and would obsolete that patch
because it catches more cases (notable the AVX1 case that was missed there).
All we're doing is allowing the existing transform to fire more often by
removing the post-legalize constraint. All of the relevant feature checks
and other predicates are left as-is.
Differential Revision: https://reviews.llvm.org/D48078
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334592
91177308-0d34-0410-b5e6-
96231b3b80d8
Alex Bradbury [Wed, 13 Jun 2018 12:04:51 +0000 (12:04 +0000)]
[RISCV] Add codegen support for atomic load/stores with RV32A
Fences are inserted according to table A.6 in the current draft of version 2.3
of the RISC-V Instruction Set Manual, which incorporates the memory model
changes and definitions contributed by the RISC-V Memory Consistency Model
task group.
Instruction selection failures will now occur for 8/16/32-bit atomicrmw and
cmpxchg operations when targeting RV32IA until lowering for these operations
is added in a follow-on patch.
Differential Revision: https://reviews.llvm.org/D47589
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334591
91177308-0d34-0410-b5e6-
96231b3b80d8
Alex Bradbury [Wed, 13 Jun 2018 11:58:46 +0000 (11:58 +0000)]
[RISCV] Codegen support for atomic operations on RV32I
This patch adds lowering for atomic fences and relies on AtomicExpandPass to
lower atomic loads/stores, atomic rmw, and cmpxchg to __atomic_* libcalls.
test/CodeGen/RISCV/atomic-* are modelled on the exhaustive
test/CodeGen/PPC/atomics-regression.ll, and will prove more useful once RV32A
codegen support is introduced.
Fence mappings are taken from table A.6 in the current draft of version 2.3 of
the RISC-V Instruction Set Manual, which incorporates the memory model changes
and definitions contributed by the RISC-V Memory Consistency Model task group.
Differential Revision: https://reviews.llvm.org/D47587
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334590
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 13 Jun 2018 10:49:24 +0000 (10:49 +0000)]
[SLPVectorizer] getSameOpcode - remove useless cast [NFC]
There's no need to cast the base Value to an Instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334588
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 13 Jun 2018 10:14:27 +0000 (10:14 +0000)]
[SLPVectorizer] getSameOpcode - remove unusued alternate code [NFC]
We early-out for the case where we don't use alternate opcodes, so no need to check for it later.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334587
91177308-0d34-0410-b5e6-
96231b3b80d8
Clement Courbet [Wed, 13 Jun 2018 09:41:49 +0000 (09:41 +0000)]
[TableGen] Emit a fatal error on inconsistencies in resource units vs cycles.
Summary:
For targets I'm not familiar with, I've automatically made the "default to 1 for each resource" behaviour explicit in the td files.
For more obvious cases, I've ventured a fix.
Some notes:
- Exynos is especially fishy.
- AArch64SchedThunderX2T99.td had some truncated entries. If I understand correctly, the person who wrote that interpreted the ResourceCycle as a range. I made the decision to use the upper/lower bound for consistency with the 'Latency' value. I'm sure there is a better choice.
- The change to X86ScheduleBtVer2.td is an NFC, it just makes values more explicit.
Also see PR37310.
Reviewers: RKSimon, craig.topper, javed.absar
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D46356
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334586
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Wed, 13 Jun 2018 09:11:10 +0000 (09:11 +0000)]
GettingStarted.rst: Fix 'If you you' typo (PR37787)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334584
91177308-0d34-0410-b5e6-
96231b3b80d8
Hiroshi Inoue [Wed, 13 Jun 2018 08:54:13 +0000 (08:54 +0000)]
[PowerPC] fix trivial typos in comment, NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334583
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Wed, 13 Jun 2018 08:43:03 +0000 (08:43 +0000)]
Fix -DLLVM_ENABLE_THREADS=OFF build after r334537
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334582
91177308-0d34-0410-b5e6-
96231b3b80d8
Pavel Labath [Wed, 13 Jun 2018 08:29:19 +0000 (08:29 +0000)]
Fix "Optional" is ambiguous error on some bots
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334580
91177308-0d34-0410-b5e6-
96231b3b80d8
Hiroshi Inoue [Wed, 13 Jun 2018 08:25:14 +0000 (08:25 +0000)]
[PowerPC] avoid verification failure due to PowerPC VSX Swap Removal pass
This patch fixes a failure in lnt tests with -verify-machineinstrs option.
When VSX Swap Removal pass swaps two register operands, it did not maintain kill flags associated with operands. This patch swaps flags as well as register number to avoid inconsistent kill flags information.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334579
91177308-0d34-0410-b5e6-
96231b3b80d8
Pavel Labath [Wed, 13 Jun 2018 08:14:27 +0000 (08:14 +0000)]
[DWARF/AccelTable] Remove getDIESectionOffset for DWARF v5 entries
Summary:
This method was not correct for entries in DWO files as it assumed it
could just add up the CU and DIE offsets to get the absolute DIE offset.
This is not correct for the DWO files, as here the CU offset will
reference the skeleton unit, whereas the DIE offset will be the offset
in the full unit in the DWO file.
Unfortunately, this means that we are not able to determine the absolute
DIE offset using the information in the .debug_names section alone,
which means we have to offload some of this work to the users of this
class.
To demonstrate how this can be done, I've added/fixed the ability to
lookup entries using accelerator tables in DWO files in llvm-dwarfdump.
To make this happen, I've needed to make two extra changes in other
classes:
- made the DWARFContext method to lookup a CU based on the section
offset public. I've needed this functionality to lookup a CU, and this
seems like a useful thing in general.
- made DWARFUnit::getDWOId call extractDIEsIfNeeded. Before this, the
DWOId was filled in only if the root DIE happened to be parsed
before we called the accessor. Since the lazy parsing is supposed to
happen under the hood, calling extractDIEsIfNeeded seems appropriate.
Reviewers: JDevlieghere, aprantl, dblaikie
Subscribers: mgrang, llvm-commits
Differential Revision: https://reviews.llvm.org/D48009
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334578
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 13 Jun 2018 07:19:21 +0000 (07:19 +0000)]
[X86] Remove masking from avx512vbmi2 concat and shift by immediate intrinsics. Use select in IR instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334576
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Wed, 13 Jun 2018 05:53:59 +0000 (05:53 +0000)]
Set the code model when specified.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334571
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Wed, 13 Jun 2018 05:14:10 +0000 (05:14 +0000)]
utils/release: Add merge-git.sh
Summary:
This script allows you to use git to backport a commit to a stable
branch while generating the exact same commit message (ignoring
whitespace) that you would get from using the merge.sh script with svn.
Reviewers: hansw
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47760
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334568
91177308-0d34-0410-b5e6-
96231b3b80d8
Max Kazantsev [Wed, 13 Jun 2018 02:25:32 +0000 (02:25 +0000)]
[SimplifyIndVars] Ignore dead users
IndVarSimplify sometimes makes transforms basing on users that are trivially dead. In particular,
if DCE wasn't run before it, there may be a dead `sext/zext` in loop that will trigger widening
transforms, however it makes no sense to do it.
This patch teaches IndVarsSimplify ignore the mist trivial cases of that.
Differential Revision: https://reviews.llvm.org/D47974
Reviewed By: sanjoy
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334567
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 13 Jun 2018 02:16:01 +0000 (02:16 +0000)]
Revert "Fix how LLVMOPTIONALCOMPONENTS is passed to llvm-build"
This reverts commit r334543.
My understanding is, that commit is intended to make the llvm-build
invocation have a correct "--enable-optional-components" value, but:
- it already has a value: it's quoted in the command line a few lines
below, and, if I hack llvm-build to print sys.argv, it does look correct:
-- llvm-build output: ['.../utils/llvm-build/llvm-build',
'--native-target', 'X86', '--enable-targets', 'X86;ARM;AArch64',
'--enable-optional-components', '',
'--write-library-table',
'.../build/tools/llvm-config/LibraryDependencies.inc',
'--write-cmake-fragment', '.../build/LLVMBuild.cmake']
- the " " string seems to evaluate to TRUE in CMake (*sigh*), so this
basically force-enables LLVM_USE_INTEL_JITEVENTS, regardless of the
value of the option.
On Darwin, JITEvents is not supported, so this bypasses that OS check
but is guaranteed to fail later.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334566
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 13 Jun 2018 00:04:08 +0000 (00:04 +0000)]
[X86] Mark all instructions that have masked store semantics with NotMemoryFoldable. Remove dependency on SchedRW from memory table autogenerator.
Previously we were whitelisting in instructions based on their SchedRW value. With the masked store instructions explicitly removed via NotMemoryFoldable, we don't seem to need this check anymore.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334563
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 13 Jun 2018 00:04:04 +0000 (00:04 +0000)]
[X86] Remove VPCOMPRESSB/W from the autogenerated load folding table.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334562
91177308-0d34-0410-b5e6-
96231b3b80d8
Stanislav Mekhanoshin [Tue, 12 Jun 2018 23:50:37 +0000 (23:50 +0000)]
[AMDGPU] DAG combine to produce V_PERM_B32
Differential Revision: https://reviews.llvm.org/D48099
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334559
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 12 Jun 2018 22:50:37 +0000 (22:50 +0000)]
[AArch64] add tests for fadd with more than one use; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334556
91177308-0d34-0410-b5e6-
96231b3b80d8
Raphael Isemann [Tue, 12 Jun 2018 22:40:20 +0000 (22:40 +0000)]
Added missing include to AMDHSAKernelDescriptor.h
We use size_t in this header, so we also need to include
cstddef to make it compile. Fixes the module builds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334555
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Tue, 12 Jun 2018 21:51:49 +0000 (21:51 +0000)]
[DAGCombiner] Recognize more patterns for ABS
Differential Revision: https://reviews.llvm.org/D47831
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334553
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 12 Jun 2018 21:42:42 +0000 (21:42 +0000)]
[X86] add avx512 tests for potentially miscompiling cvttp2si/cvttp2ui (PR37751).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334551
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Tue, 12 Jun 2018 21:15:27 +0000 (21:15 +0000)]
Remove malloc.h include from Intel JIT events code
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334547
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Tue, 12 Jun 2018 20:54:11 +0000 (20:54 +0000)]
Add null check to Intel JIT event listener
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334544
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Tue, 12 Jun 2018 20:53:46 +0000 (20:53 +0000)]
Fix how LLVMOPTIONALCOMPONENTS is passed to llvm-build
Patch by Force.Charlie-I
If LLVM_USE_INTEL_JITEVENTS and LLVM_USE_OPROFILE not set,
"${LLVMOPTIONALCOMPONENTS}" is empty, but
**--enable-optional-components** need arg, Cause
**--write-library-table** to be skipped parsed.
Differential Revision: https://reviews.llvm.org/D47982
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334543
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Tue, 12 Jun 2018 20:43:18 +0000 (20:43 +0000)]
[ORC] Add a fallback definition generator for VSOs.
If a VSO has a fallback definition generator attached it will be called during
lookup (and lookupFlags) for any unresolved symbols. The definition generator
can add new definitions to the VSO for any unresolved symbol. This allows VSOs
to generate new definitions on demand.
The immediate use case for this code is supporting VSOs that can import
definitions found via dlsym on demand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334538
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Tue, 12 Jun 2018 20:43:17 +0000 (20:43 +0000)]
[ORC] Refactor blocking lookup logic into the blockingLookup function, and
implement existing blocking lookups (the lookup function) and
JITSymbolResolverAdapter on top of that.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334537
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Tue, 12 Jun 2018 20:43:17 +0000 (20:43 +0000)]
[RuntimeDyld] Add an assert to catch misbehaving symbol resolvers.
Resolvers are required to find results for all requested symbols or return an
error, but if a resolver fails to adhere to this contract (by returning results
for only a subset of the requested symbols) then this code will infinite loop.
This assertion catches resolvers that fail to adhere to the contract.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334536
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Tue, 12 Jun 2018 20:43:15 +0000 (20:43 +0000)]
[MCJIT] Call materializeAll on modules before compiling them in MCJIT.
This only affects modules with lazy GVMaterializers attached (usually modules
read off disk using the lazy bitcode reader). For such modules, materializing
before compiling prevents crashes due to missing function bodies /
initializers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334535
91177308-0d34-0410-b5e6-
96231b3b80d8
Petr Hosek [Tue, 12 Jun 2018 20:00:50 +0000 (20:00 +0000)]
[AArch64] Support reserving x20 register
Register x20 is a callee-saved register which may be used for other
purposes in certain contexts, for example to hold special variables
within the kernel. This change adds support for reserving this register
both to frontend and backend to make this register usable for these
purposes.
Differential Revision: https://reviews.llvm.org/D46552
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334531
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 12 Jun 2018 19:59:10 +0000 (19:59 +0000)]
[X86] Remove unnecessary include from one of the tblgen emitters.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334530
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 12 Jun 2018 19:59:08 +0000 (19:59 +0000)]
[X86] Remove mayLoad flag from AVX512 truncating store instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334529
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Tue, 12 Jun 2018 18:56:05 +0000 (18:56 +0000)]
[MS][ARM64] Hoist __ImageBase handling into TargetLoweringObjectFileCOFF
All COFF targets should use @IMGREL32 relocations for symbol differences
against __ImageBase. Do the same for getSectionForConstant, so that
immediates lowered to globals get merged across TUs.
Patch by Chris January
Differential Revision: https://reviews.llvm.org/D47783
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334523
91177308-0d34-0410-b5e6-
96231b3b80d8
Konstantin Zhuravlyov [Tue, 12 Jun 2018 18:33:51 +0000 (18:33 +0000)]
AMDHSA/NFC: Code object v3 updates (additional):
- Move section selection and alignment to AMDGPUAsmPrinter
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334521
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Tereshin [Tue, 12 Jun 2018 18:30:37 +0000 (18:30 +0000)]
[MIR][MachineCSE] Implementing proper MachineInstr::getNumExplicitDefs()
Apparently, MachineInstr class definition as well as pretty much all of
the machine passes assume that the only kind of MachineInstr's operands
that is variadic for variadic opcodes is explicit non-definitions.
In particular, this assumption is made by MachineInstr::defs(), uses(),
and explicit_uses() methods, as well as by MachineCSE pass.
The assumption is incorrect judging from at least TableGen backend
implementation, that recognizes variable_ops in OutOperandList, and the
very existence of G_UNMERGE_VALUES generic opcode, or ARM load multiple
instructions, all of which have variadic defs.
In particular, MachineCSE pass breaks MIR with CSE'able G_UNMERGE_VALUES
instructions in it.
This commit implements MachineInstr::getNumExplicitDefs() similar to
pre-existing MachineInstr::getNumExplicitOperands(), fixes
MachineInstr::defs(), uses(), and explicit_uses(), and fixes MachineCSE
pass.
As the issue addressed seems to affect only machine passes that could be
ran mid-GlobalISel pipeline at the moment, the other passes aren't fixed
by this commit, like MachineLICM: that could be done on per-pass basis
when (if ever) they get adopted for GlobalISel.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D45640
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334520
91177308-0d34-0410-b5e6-
96231b3b80d8
Konstantin Zhuravlyov [Tue, 12 Jun 2018 18:02:46 +0000 (18:02 +0000)]
AMDHSA: Code object v3 updates
- Do not emit following assembler directives:
- .hsa_code_object_version
- .hsa_code_object_isa
- .amd_amdgpu_isa
- .amd_amdgpu_hsa_metadata
- .amd_amdgpu_pal_metadata
- Do not emit .note entries
- Cleanup and bring in sync kernel descriptor header file
- Emit kernel descriptor into .rodata with appropriate relocations and
alignments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334519
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Tue, 12 Jun 2018 17:43:52 +0000 (17:43 +0000)]
Refactor ExecuteAndWait to take StringRefs.
This simplifies some code which had StringRefs to begin with, and
makes other code more complicated which had const char* to begin
with.
In the end, I think this makes for a more idiomatic and platform
agnostic API. Not all platforms launch process with null terminated
c-string arrays for the environment pointer and argv, but the api
was designed that way because it allowed easy pass-through for
posix-based platforms. There's a little additional overhead now
since on posix based platforms we'll be takign StringRefs which
were constructed from null terminated strings and then copying
them to null terminate them again, but from a readability and
usability standpoint of the API user, I think this API signature
is strictly better.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334518
91177308-0d34-0410-b5e6-
96231b3b80d8
Fangrui Song [Tue, 12 Jun 2018 16:20:44 +0000 (16:20 +0000)]
[MC] [X86] Teach leaq _GLOBAL_OFFSET_TABLE(%rip), %r15 to use R_X86_64_GOTPC32 instead of R_X86_64_PC32
Summary:
This is similar to D46319 (ARM). x86-64 psABI p40 gives an example:
leaq _GLOBAL_OFFSET_TABLE(%rip), %r15 # GOTPC32 reloc
GNU as creates R_X86_64_GOTPC32. However, MC currently emits R_X86_64_PC32.
Reviewers: javed.absar, echristo
Subscribers: kristof.beyls, llvm-commits, peter.smith, grimar
Differential Revision: https://reviews.llvm.org/D47507
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334515
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Berg [Tue, 12 Jun 2018 16:13:11 +0000 (16:13 +0000)]
Utilize new SDNode flag functionality to expand current support for fmul
Summary: This patch originated from D46562 and is a proper subset, with some issues addressed for fmul.
Reviewers: spatel, hfinkel, wristow, arsenm
Reviewed By: spatel
Subscribers: nhaehnle, wdng
Differential Revision: https://reviews.llvm.org/D47911
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334514
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Tue, 12 Jun 2018 16:12:29 +0000 (16:12 +0000)]
[CostModel] Replace ShuffleKind::SK_Alternate with ShuffleKind::SK_Select (PR33744)
As discussed on PR33744, this patch relaxes ShuffleKind::SK_Alternate which requires shuffle masks to only match an alternating pattern from its 2 sources:
e.g. v4f32: <0,5,2,7> or <4,1,6,3>
This seems far too restrictive as most SIMD hardware which will implement it using a general blend/bit-select instruction, so replaces it with SK_Select, permitting elements from either source as long as they are inline:
e.g. v4f32: <0,5,2,7>, <4,1,6,3>, <0,1,6,7>, <4,1,2,3> etc.
This initial patch just updates the name and cost model shuffle mask analysis, later patch reviews will update SLP to better utilise this - it still limits itself to SK_Alternate style patterns.
Differential Revision: https://reviews.llvm.org/D47985
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334513
91177308-0d34-0410-b5e6-
96231b3b80d8
Paul Robinson [Tue, 12 Jun 2018 16:09:03 +0000 (16:09 +0000)]
[DWARFv5] llvm-mc -dwarf-version does not imply -g.
Don't provide the assembler source as the "root file" unless the user
asked to have debug info for the assembler source (with -g).
If the source doesn't provide an explicit ".file 0" then (a) use the
compilation directory as directory #0, and (b) use the file #1 info
for file #0 also.
Differential Revision: https://reviews.llvm.org/D48055
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334512
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 12 Jun 2018 15:48:03 +0000 (15:48 +0000)]
[X86] Remove TB_ALIGN_16 from VEXTRACTF128/VEXTRACTI128 in the memory folding table.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334511
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Tue, 12 Jun 2018 15:14:34 +0000 (15:14 +0000)]
Fix signed/unsigned warning. NFCI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334509
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Tue, 12 Jun 2018 15:12:50 +0000 (15:12 +0000)]
[Hexagon] Make floating point operations expensive for vectorization
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334508
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Tue, 12 Jun 2018 14:47:13 +0000 (14:47 +0000)]
[CostModel] Treat Identity shuffle masks as zero cost
As discussed on D47985, identity shuffle masks should probably be free.
I've limited this to the case where the input and output types all match - but we could probably accept all cases.
Differential Revision: https://reviews.llvm.org/D47986
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334506
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 12 Jun 2018 14:21:51 +0000 (14:21 +0000)]
[x86] move shrunkblend transform to helper function; NFCI
We should be able to obsolete D48043 by easing the constraints
on this existing code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334504
91177308-0d34-0410-b5e6-
96231b3b80d8
Clement Courbet [Tue, 12 Jun 2018 13:28:37 +0000 (13:28 +0000)]
[llvm-exegesis] Sum counter values when several counters are specified for a ProcRes.
Summary: This allows handling memory ports on SNB.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48076
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334502
91177308-0d34-0410-b5e6-
96231b3b80d8
Guillaume Chatelet [Tue, 12 Jun 2018 13:07:16 +0000 (13:07 +0000)]
[llvm-exegesis] Move libpfm linking to LLVMExegesis.
Summary: This patch moves linking of libpfm from different places to a single one.
Reviewers: courbet
Subscribers: mgorny, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D48075
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334499
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Tue, 12 Jun 2018 12:49:36 +0000 (12:49 +0000)]
[SelectionDAG] Provide default expansion for rotates
Implement default legalization of rotates: either in terms of the rotation
in the opposite direction (if legal), or in terms of shifts and ors.
Implement generating of rotate instructions for Hexagon. Hexagon only
supports rotates by an immediate value, so implement custom lowering of
ROTL/ROTR on Hexagon. If a rotate is not legal, use the default expansion.
Differential Revision: https://reviews.llvm.org/D47725
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334497
91177308-0d34-0410-b5e6-
96231b3b80d8
Florian Hahn [Tue, 12 Jun 2018 11:16:56 +0000 (11:16 +0000)]
Use SmallPtrSet explicitly for SmallSets with pointer types (NFC).
Currently SmallSet<PointerTy> inherits from SmallPtrSet<PointerTy>. This
patch replaces such types with SmallPtrSet, because IMO it is slightly
clearer and allows us to get rid of unnecessarily including SmallSet.h
Reviewers: dblaikie, craig.topper
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D47836
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334492
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Dardis [Tue, 12 Jun 2018 10:28:06 +0000 (10:28 +0000)]
[mips] Guard some floating point instructions correctly
Reviewers: smaksimovic, atanasyan, abeserminji
Differential Revision: https://reviews.llvm.org/D47636
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334491
91177308-0d34-0410-b5e6-
96231b3b80d8
Aleksandar Beserminji [Tue, 12 Jun 2018 10:23:49 +0000 (10:23 +0000)]
[mips] Extend LONG_BRANCH_LUi/ADDiu with extra parameter
Extend LONG_BRANCH_LUi and LONG_BRANCH_ADDiu pseudo instructions with
additional flag, so instead of always lowering to lui %hi(...),
addiu %lo(...) or addiu %hi(...), now they can lower to either %lo, %hi,
%higher or %highest depending on the added flag.
Differential Revision: https://reviews.llvm.org/D47941
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334490
91177308-0d34-0410-b5e6-
96231b3b80d8
Luke Geeson [Tue, 12 Jun 2018 09:35:20 +0000 (09:35 +0000)]
[AArch64] Audit on rL333879 to fix FP16 64bit bitpatterns
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334488
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Tue, 12 Jun 2018 09:18:13 +0000 (09:18 +0000)]
[CostModel][X86] Add extra Identity shuffle mask cost tests (D47986)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334486
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 12 Jun 2018 07:32:19 +0000 (07:32 +0000)]
[X86] Add NotMemoryFoldable to the VPCOMPRESS instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@334481
91177308-0d34-0410-b5e6-
96231b3b80d8