OSDN Git Service
David Blaikie [Wed, 2 Dec 2015 06:21:34 +0000 (06:21 +0000)]
[llvm-dwp] Emit a rather fictional debug_cu_index
This is very rudimentary support for debug_cu_index, but it is enough to
allow llvm-dwarfdump to find the offsets for contributions and
correctly dump debug_info.
It will need to actually find the real signature of the unit and build
the real hash table with the right number of buckets, as per the DWP
specification.
It will also need to be expanded to cover the tu_index as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254489
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Wed, 2 Dec 2015 06:21:28 +0000 (06:21 +0000)]
DebugInfo\DWARF: Privatize some accidentally public members
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254488
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 2 Dec 2015 05:24:38 +0000 (05:24 +0000)]
[X86] Fix weird identation. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254487
91177308-0d34-0410-b5e6-
96231b3b80d8
Mehdi Amini [Wed, 2 Dec 2015 04:34:28 +0000 (04:34 +0000)]
Change ModuleLinker to take a set of GlobalValues to import instead of a single one
For efficiency reason, when importing multiple functions for the same Module,
we can avoid reparsing it every time.
Differential Revision: http://reviews.llvm.org/D15102
From: Mehdi Amini <mehdi.amini@apple.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254486
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Wed, 2 Dec 2015 02:49:37 +0000 (02:49 +0000)]
[libFuzzer] add a test that is built with -fsanitize-coverage=trace-bb
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254484
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Wed, 2 Dec 2015 02:37:13 +0000 (02:37 +0000)]
[sanitizer coverage] when adding a bb trace instrumentation, do it instead, not in addition to, regular coverage. Do the regular coverage in the run-time instead
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254482
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Wed, 2 Dec 2015 02:07:00 +0000 (02:07 +0000)]
[X86] Fix a think-o when checking if the eflags needs to be preserved.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254480
91177308-0d34-0410-b5e6-
96231b3b80d8
Mehdi Amini [Wed, 2 Dec 2015 02:00:29 +0000 (02:00 +0000)]
Modify FunctionImport to take a callback to load modules
When linking static archive, there is no individual module files to
load. Instead they can be mmap'ed and could be initialized from a
buffer directly. The callback provide flexibility to override the
scheme for loading module from the summary.
Differential Revision: http://reviews.llvm.org/D15101
From: Mehdi Amini <mehdi.amini@apple.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254479
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Wed, 2 Dec 2015 01:22:54 +0000 (01:22 +0000)]
[X86] Make sure the prologue does not clobber EFLAGS when it lives accross it.
This is a superset of the fix done in r254448.
This fixes PR25607.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254478
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Wed, 2 Dec 2015 00:33:54 +0000 (00:33 +0000)]
AArch64: fix 128-bit shifts
We mustn't introduce a shift of exactly 64-bits for any inputs, since that's an
UNDEF value (and worse, it's not what you want with the natural Arch64
implementation).
The generated code is pretty horrific, but I couldn't come up with an obviously
better alternative (if the amount is constant EXTR could help). Turns out
128-bit shifts are just nasty.
rdar://
22491037
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254475
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 23:06:26 +0000 (23:06 +0000)]
Use default member initializers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254473
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Tue, 1 Dec 2015 23:05:27 +0000 (23:05 +0000)]
Define member operator delete
For the struct with trailing objects, define
a member operator delete. Without this, the program
will fail when -fsized-deallocation option is used
where the wrong size will be passed to the global
delete operator.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254471
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 1 Dec 2015 23:04:05 +0000 (23:04 +0000)]
AMDGPU: Error on addrspacecasts that aren't actually implemented
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254469
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 1 Dec 2015 23:04:00 +0000 (23:04 +0000)]
AMDGPU: Implement isNoopAddrSpaceCast
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254468
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 23:01:51 +0000 (23:01 +0000)]
Remove unnecessary getter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254466
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 22:40:40 +0000 (22:40 +0000)]
Pass down the dst GV to linkGlobalValueBody. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254465
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Tue, 1 Dec 2015 21:50:20 +0000 (21:50 +0000)]
Fix a bug in IfConversion.cpp.
The bug is introduced in r254377 which failed some tests on ARM, where a new
probability is assigned to a successor but the provided BB may not be a
successor.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254463
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Tue, 1 Dec 2015 21:48:52 +0000 (21:48 +0000)]
ARM: Change ArchCheck field to uint64_t
The values in this field are compared against getAvailableFeatures()
which returns an uint64_t. This was causing problems in an internal
branch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254462
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 1 Dec 2015 20:31:08 +0000 (20:31 +0000)]
AMDGPU: Disallow flat_scr in SI assembler
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254459
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Tue, 1 Dec 2015 20:26:26 +0000 (20:26 +0000)]
[PGO] Add support for reading multiple versions of indexed profile format profile data
Profile readers using incompatible on-disk hash table format can now share the same
implementation and interfaces.
Differential Revision: http://reviews.llvm.org/D15100
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254458
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 20:23:19 +0000 (20:23 +0000)]
Delete unused includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254457
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Tue, 1 Dec 2015 20:20:49 +0000 (20:20 +0000)]
IR: Clean up some duplicated code in ConstantDataSequential creation. NFC
ConstantDataArray::getImpl and ConstantDataVector::getImpl had a lot
of copy pasta in how they handled sequences of constants. Break that
out into a couple of simple functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254456
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 20:11:43 +0000 (20:11 +0000)]
clang-format LinkModules.cpp.
Most of the file has been changed recently and was already clang-format
clean.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254454
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 1 Dec 2015 19:57:43 +0000 (19:57 +0000)]
use range-based for loops; NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254453
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 1 Dec 2015 19:57:17 +0000 (19:57 +0000)]
AMDGPU: Optimize VOP2 operand legalization
Don't use commuteInstruction, and don't commute if
doing so will not improve legality. Skip the more
complex checks for literal operands and constant bus restrictions,
which are not a concern for VOP2 instructions because src1
does not accept SGPRs or constants and few implicitly
read vcc.
This gets called quite a few times and the
attempts at commuting are a significant fraction
of the time spent in SIFixSGPRCopies, so it's
somewhat worthwhile to optimize. With this patch and others
leading up to it, this reduces the compile time of SIFixSGPRCopies
on some of the LuxMark 2 kernels from ~8ms to ~5ms on my system.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254452
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 19:50:54 +0000 (19:50 +0000)]
Use references now that it is natural to do so.
The linker never takes ownership of a module or changes which module it
is refering to, making it natural to use references.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254449
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Tue, 1 Dec 2015 19:49:31 +0000 (19:49 +0000)]
[X86] Make sure the prologue does not clobber EFLAGS when it lives accross it.
This fixes PR25629.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254448
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Tue, 1 Dec 2015 19:47:32 +0000 (19:47 +0000)]
Use nullptr (NFC)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254447
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 1 Dec 2015 19:32:35 +0000 (19:32 +0000)]
don't repeat function/variable names in comments; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254445
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Tue, 1 Dec 2015 19:25:11 +0000 (19:25 +0000)]
Fix Thumb1 epilogue generation
Summary:
This had been broken for a very long time, but nobody noticed until
D14357 enabled shrink-wrapping by default.
Reviewers: jroelofs, qcolombet
Subscribers: tyomitch, llvm-commits, rengolin
Differential Revision: http://reviews.llvm.org/D14986
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254444
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 1 Dec 2015 19:19:18 +0000 (19:19 +0000)]
fix typo; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254442
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 1 Dec 2015 19:17:58 +0000 (19:17 +0000)]
[llvm-dwp] Deduplicate strings in the debug_str.dwo section
Also, ensure that references to those strings in debug_str_offsets.dwo
correctly refer to the deduplicated strings.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254441
91177308-0d34-0410-b5e6-
96231b3b80d8
Weiming Zhao [Tue, 1 Dec 2015 19:17:49 +0000 (19:17 +0000)]
[AArch64] Fix a corner case in BitFeild select
Summary:
When not useful bits, BitWidth becomes 0 and APInt will not be happy.
See https://llvm.org/bugs/show_bug.cgi?id=25571
We can just mark the operand as IMPLICIT_DEF is none bits of it is used.
Reviewers: t.p.northover, jmolloy
Subscribers: gberry, jmolloy, mgrang, aemerson, llvm-commits, rengolin
Differential Revision: http://reviews.llvm.org/D14803
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254440
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 1 Dec 2015 19:08:39 +0000 (19:08 +0000)]
AMDGPU: Report extractelement as free in cost model
The cost for scalarized operations is computed as N * (scalar operation
cost + 1 extractelement + 1 insertelement). This partially fixes
inflating the cost of scalarized operations since every operation is
scalarized and free. I don't think we want any cost asociated with
scalarization, but for now insertelement is still counted. I'm not sure
if we should pretend that insertelement is also free, or add a way
to compute a custom scalarization cost.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254438
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Tue, 1 Dec 2015 19:06:36 +0000 (19:06 +0000)]
[Verifier] Improve error for cross-module refs
By including the module name in the error message.
This makes the error message much more useful and
saves a trip to the debugger.
Reviewers: dexonsmith
Subscribers: dexonsmith, llvm-commits
Differential Revision: http://reviews.llvm.org/D14473
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254437
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 18:50:35 +0000 (18:50 +0000)]
Delete dead code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254436
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 18:46:19 +0000 (18:46 +0000)]
Use a forwarding constructor instead of an init method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254435
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 18:41:30 +0000 (18:41 +0000)]
Delete the setModule method from the Linker.
It was only used from LTO for a debug feature, and LTO can just create
another linker.
It is pretty odd to have a method to reset the module in the middle of a
link. It would make IdentifiedStructTypes inconsistent with the Module
for example.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254434
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 1 Dec 2015 18:07:07 +0000 (18:07 +0000)]
[llvm-dwp] Correctly update debug_str_offsets.dwo when linking dwo files
This doesn't deduplicate strings in the debug_str section, nor does it
properly wire up the index so that debug_info can /find/ these strings,
but it does correct the str_offsets specifically.
Follow up patches to address those related/next issues.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254431
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Tue, 1 Dec 2015 17:45:22 +0000 (17:45 +0000)]
AMDGPU/SI: Remove REGISTER_STORE/REGISTER_LOAD code which is now dead
Reviewers: arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D15050
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254427
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Tue, 1 Dec 2015 17:45:17 +0000 (17:45 +0000)]
AMDGPU: Use the default strings for data emission directives
Summary:
This makes the assembly output look nicer and there is no reason to
have custom strings for these.
Reviewers: arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D14671
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254426
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 1 Dec 2015 17:27:55 +0000 (17:27 +0000)]
[x86] add a convenience method to check for FMA capability; NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254425
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 17:17:04 +0000 (17:17 +0000)]
Make appending var linking less of a special case.
It has to be a bit special because:
* materializeInitFor is not really supposed to call replaceAllUsesWith.
The caller has a plain variable with Dst and expects just the
initializer to be set, not for it to be removed.
* Calling mutateType as we used to do before gets some type
inconsistency which breaks the bitcode writer.
* If linkAppendingVarProto create a dest decl with the correct type to
avoid the above problems, it needs to put the original dst init in
some side table for materializeInitFor to use.
In the end the simplest solution seems to be to just have
linkAppendingVarProto do all the work and set ValueMap[SrcGV to avoid
recursion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254424
91177308-0d34-0410-b5e6-
96231b3b80d8
Teresa Johnson [Tue, 1 Dec 2015 17:12:10 +0000 (17:12 +0000)]
[ThinLTO] Wrap dbgs() output in DEBUG macro
Missed in a couple places.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254422
91177308-0d34-0410-b5e6-
96231b3b80d8
Teresa Johnson [Tue, 1 Dec 2015 16:45:23 +0000 (16:45 +0000)]
[ThinLTO] Remove stale comment (NFC)
Stale as of r254036 which added basic profitability check.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254421
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 15:46:46 +0000 (15:46 +0000)]
Simplify test. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254419
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 1 Dec 2015 15:19:48 +0000 (15:19 +0000)]
Bring r254336 back:
The difference is that now we don't error on out-of-comdat access to
internal global values. We copy them instead. This seems to match the
expectation of COFF linkers (see pr25686).
Original message:
Start deciding earlier what to link.
A traditional linker is roughly split in symbol resolution and
"copying
stuff".
The two tasks are badly mixed in lib/Linker.
This starts splitting them apart.
With this patch there are no direct call to linkGlobalValueBody or
linkGlobalValueProto. Everything is linked via WapValue.
This also includes a few fixes:
* A GV goes undefined if the comdat is dropped (comdat11.ll).
* We error if an internal GV goes undefined (comdat13.ll).
* We don't link an unused comdat.
The first two match the behavior of an ELF linker. The second one is
equivalent to running globaldce on the input.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254418
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 1 Dec 2015 14:26:35 +0000 (14:26 +0000)]
[LIR] Push check into helper function. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254416
91177308-0d34-0410-b5e6-
96231b3b80d8
Yury Gribov [Tue, 1 Dec 2015 13:24:48 +0000 (13:24 +0000)]
Fix "WARNING: Title underline too short." introduced by r254404.
Patch by Max Ostapenko.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254413
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Tue, 1 Dec 2015 12:43:46 +0000 (12:43 +0000)]
AVX-512: fixed asm string of vsqrtss
(vvsqrtss was generated before)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254411
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Tue, 1 Dec 2015 12:35:03 +0000 (12:35 +0000)]
AVX-512: regenerated test for avx512 arithmetics, NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254410
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Tue, 1 Dec 2015 12:30:40 +0000 (12:30 +0000)]
Fixed a failure in getSpaltValue()
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254409
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Tue, 1 Dec 2015 12:08:36 +0000 (12:08 +0000)]
Fixed a failure in cost calculation for vector GEP
Cost calculation for vector GEP failed with due to invalid cast to GEP index operand.
The bug is fixed, added a test.
http://reviews.llvm.org/D14976
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254408
91177308-0d34-0410-b5e6-
96231b3b80d8
Hrvoje Varga [Tue, 1 Dec 2015 11:59:21 +0000 (11:59 +0000)]
[mips][microMIPS] Implement RECIP.fmt, RINT.fmt, ROUND.L.fmt, ROUND.W.fmt, SEL.fmt, SELEQZ.fmt, SELNEQZ.fmt and CLASS.fmt
Differential Revision: http://reviews.llvm.org/D13885
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254405
91177308-0d34-0410-b5e6-
96231b3b80d8
Yury Gribov [Tue, 1 Dec 2015 11:40:55 +0000 (11:40 +0000)]
Introduce new @llvm.get.dynamic.area.offset.i{32, 64} intrinsics.
The @llvm.get.dynamic.area.offset.* intrinsic family is used to get the offset
from native stack pointer to the address of the most recent dynamic alloca on
the caller's stack. These intrinsics are intendend for use in combination with
@llvm.stacksave and @llvm.restore to get a pointer to the most recent dynamic
alloca. This is useful, for example, for AddressSanitizer's stack unpoisoning
routines.
Patch by Max Ostapenko.
Differential Revision: http://reviews.llvm.org/D14983
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254404
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Tue, 1 Dec 2015 11:05:39 +0000 (11:05 +0000)]
Allow known and unknown probabilities coexist in MBB's successor list.
Previously it is not allowed for each MBB to have successors with both known and
unknown probabilities. However, this may be too strict as at this stage we could
not always guarantee that. It is better to remove this restriction now, and I
will work on validating MBB's successors' probabilities first (for example,
check if the sum is approximate one).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254402
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Tue, 1 Dec 2015 10:48:51 +0000 (10:48 +0000)]
[AArch64] Add ARMv8.2-A Statistical Profiling Extension
The Statistical Profiling Extension is an optional extension to
ARMv8.2-A. Since it is an optional extension, I have added the
FeatureSPE subtarget feature to control it. The assembler-visible parts
of this extension are the new "psb csync" instruction, which is
equivalent to "hint #17", and a number of system registers.
Differential Revision: http://reviews.llvm.org/D15021
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254401
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Tue, 1 Dec 2015 10:33:56 +0000 (10:33 +0000)]
[ARM] Add ARMv8.2-A to TargetParser
Add ARMv8.2-A to TargetParser, so that it can be used by the clang
command-line options and the .arch directive.
Most testing of this will be done in clang, checking that the
command-line options that this enables work.
Differential Revision: http://reviews.llvm.org/D15037
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254400
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Tue, 1 Dec 2015 10:23:06 +0000 (10:23 +0000)]
[ARM] Add subtarget features for ARMv8.2-A
This adds subtarget features for ARMv8.2-A, which builds on (and
requires the features from) ARMv8.1-A. Most assembler-visible features
of ARMv8.2-A are system instructions, and are all required parts of the
architecture, so just depend on the HasV8_2aOps subtarget feature.
There is also one large, optional feature, which adds 16-bit floating
point versions of all existing floating-point instructions (VFP and
SIMD), this is represented by the FeatureFullFP16 subtarget feature.
Differential Revision: http://reviews.llvm.org/D15036
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254399
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 1 Dec 2015 10:07:41 +0000 (10:07 +0000)]
llvm/test/DebugInfo/X86/safestack-byval.ll: Give an explicit triple for now. It crashes for targeting *-win32.
Also revert r254375 and r254361.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254397
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 1 Dec 2015 10:07:37 +0000 (10:07 +0000)]
Move llvm/test/DebugInfo/Generic/safestack-byval.ll to X86. It depends on x86-64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254396
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 1 Dec 2015 07:49:27 +0000 (07:49 +0000)]
Introduce a range version of std::find, and use in SCEV
Reviewers: dblaikie, pcc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15064
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254391
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 1 Dec 2015 07:49:23 +0000 (07:49 +0000)]
Introduce a range version of std::any_of, and use it in SCEV
Reviewers: dblaikie, pcc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15063
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254390
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:13:16 +0000 (06:13 +0000)]
[X86] Fix patterns for memory forms of FP FSUBR and FDIVR. They need to have memory on the left hand side of the fsub/fdiv operations in their patterns.
Not sure how to test this. I noticed by inspection in the isel tables where the same pattern tried to produce DIV and DIVR or SUB and SUBR.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254388
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:13:15 +0000 (06:13 +0000)]
[X86] Use range-based for loops. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254387
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:13:13 +0000 (06:13 +0000)]
[X86] Use array_lengthof instead of calculating manually. Also change index types to size_t to match.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254386
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:13:10 +0000 (06:13 +0000)]
[Hexagon] Use std::begin() and std::end() instead of doing the same manually. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254385
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:13:08 +0000 (06:13 +0000)]
[Hexagon] Use array_lengthof and const correct and type correct the array and array size. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254384
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:13:06 +0000 (06:13 +0000)]
Use array_lengthof instead of manually calculating it. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254383
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:13:04 +0000 (06:13 +0000)]
[Hexagon] Use ArrayRef to avoid needing to calculate an array size. Interestingly the original code may have had a bug because it was passing the byte size of a uint16_t array instead of the number of entries.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254382
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:13:01 +0000 (06:13 +0000)]
[ARM] Use range-based for loops to avoid the need for calculating an array size that I would have otherwise cconverted to array_lengthof. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254381
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 1 Dec 2015 06:12:59 +0000 (06:12 +0000)]
Use array_lengthof instead of manually calculating it. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254380
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Tue, 1 Dec 2015 05:33:24 +0000 (05:33 +0000)]
[Windows] Partially revert r254363 until I can test the right fix.
Reported by: David Blaikie
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254378
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Tue, 1 Dec 2015 05:29:22 +0000 (05:29 +0000)]
Replace all weight-based interfaces in MBB with probability-based interfaces, and update all uses of old interfaces.
(This is the second attempt to submit this patch. The first caused two assertion
failures and was reverted. See https://llvm.org/bugs/show_bug.cgi?id=25687)
The patch in http://reviews.llvm.org/D13745 is broken into four parts:
1. New interfaces without functional changes (http://reviews.llvm.org/D13908).
2. Use new interfaces in SelectionDAG, while in other passes treat probabilities
as weights (http://reviews.llvm.org/D14361).
3. Use new interfaces in all other passes.
4. Remove old interfaces.
This patch is 3+4 above. In this patch, MBB won't provide weight-based
interfaces any more, which are totally replaced by probability-based ones.
The interface addSuccessor() is redesigned so that the default probability is
unknown. We allow unknown probabilities but don't allow using it together
with known probabilities in successor list. That is to say, we either have a
list of successors with all known probabilities, or all unknown
probabilities. In the latter case, we assume each successor has 1/N
probability where N is the number of successors. An assertion checks if the
user is attempting to add a successor with the disallowed mixed use as stated
above. This can help us catch many misuses.
All uses of weight-based interfaces are now updated to use probability-based
ones.
Differential revision: http://reviews.llvm.org/D14973
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254377
91177308-0d34-0410-b5e6-
96231b3b80d8
Colin LeMahieu [Tue, 1 Dec 2015 04:56:25 +0000 (04:56 +0000)]
[Hexagon] Disabling failing safestack test
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254375
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Tue, 1 Dec 2015 04:20:06 +0000 (04:20 +0000)]
RegisterPressure: If we do not collect dead defs the list must be empty
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254372
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Tue, 1 Dec 2015 04:20:04 +0000 (04:20 +0000)]
RegisterPressure: Remove support for recede()/advance() at MBB boundaries
Nobody was checking the returnvalue of recede()/advance() so we can
simply replace this code with asserts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254371
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Tue, 1 Dec 2015 04:20:01 +0000 (04:20 +0000)]
RegisterPressure: There is no need to make getCurSlot() public
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254370
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Tue, 1 Dec 2015 04:19:58 +0000 (04:19 +0000)]
RegisterPressure: There is no need to make discoverLive{In|Out} public
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254369
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Tue, 1 Dec 2015 04:19:56 +0000 (04:19 +0000)]
RegisterPressure: Split RegisterOperands analysis code from result object; NFC
This is in preparation to expose the RegisterOperands class as
RegisterPressure API.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254368
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Tue, 1 Dec 2015 03:49:42 +0000 (03:49 +0000)]
Revert r254348: "Replace all weight-based interfaces in MBB with probability-based interfaces, and update all uses of old interfaces."
and the follow-up r254356: "Fix a bug in MachineBlockPlacement that may cause assertion failure during BranchProbability construction."
Asserts were firing in Chromium builds. See PR25687.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254366
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Tue, 1 Dec 2015 02:38:42 +0000 (02:38 +0000)]
[Windows] Follow-up r254363, remove return.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254364
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Tue, 1 Dec 2015 02:35:04 +0000 (02:35 +0000)]
[Windows] Simplify assertion code. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254363
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 1 Dec 2015 02:14:33 +0000 (02:14 +0000)]
Squelch unused variable warning in SIRegisterInfo.cpp.
Patch by Justin Lebar
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254362
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 1 Dec 2015 01:15:03 +0000 (01:15 +0000)]
llvm/test/DebugInfo/Generic/safestack-byval.ll is using tls.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254361
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 1 Dec 2015 01:14:58 +0000 (01:14 +0000)]
check-llvm: Introduce the new feature "tls".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254360
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 1 Dec 2015 01:07:20 +0000 (01:07 +0000)]
[llvm-dwp] Add missing Makefile for the old configure+make build
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254358
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 1 Dec 2015 00:57:05 +0000 (00:57 +0000)]
[llvm-dwp] Add missing dependency from llvm tests on the llvm-dwp tool
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254357
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Tue, 1 Dec 2015 00:55:42 +0000 (00:55 +0000)]
Fix a bug in MachineBlockPlacement that may cause assertion failure during BranchProbability construction.
The root cause is the rounding behavior in BranchProbability construction. We may consider to use truncation instead in the future.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254356
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 1 Dec 2015 00:48:39 +0000 (00:48 +0000)]
[llvm-dwp] Initial partial prototype
This just concatenates the common DWP sections without doing any of the
fancy DWP things like:
1) update str_offsets
2) deduplicating strings
3) merging/creating cu/tu_index
Patches for these will follow shortly.
(also not sure about target triple/object file type for this tool - do I
really need a whole triple just to write an object file that contains
purely static/hardcoded bytes in each section? & I guess I should just
pick it based on the first input, maybe, rather than hardcoding for now
- but we only produce .dwo on ELF platforms with objcopy for now anyway)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254355
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 1 Dec 2015 00:48:34 +0000 (00:48 +0000)]
llvm-dwp: Initial layout
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254354
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Tue, 1 Dec 2015 00:40:05 +0000 (00:40 +0000)]
[safestack] Protect byval function arguments.
Detect unsafe byval function arguments and move them to the unsafe
stack.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254353
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Tue, 1 Dec 2015 00:34:30 +0000 (00:34 +0000)]
Extend debug info for function parameters in SDAG.
SDAG currently can emit debug location for function parameters when
an llvm.dbg.declare points to either a function argument SSA temp,
or to an AllocaInst. This change extends this logic by adding a
fallback case when neither of the above is true.
This is required for SafeStack, which may copy the contents of a
byval function argument into something that is not an alloca, and
then describe the target as the new location of the said argument.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254352
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Tue, 1 Dec 2015 00:06:13 +0000 (00:06 +0000)]
[safestack] Fix handling of array allocas.
The current code does not take alloca array size into account and,
as a result, considers any access past the first array element to be
unsafe.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254350
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Tue, 1 Dec 2015 00:02:51 +0000 (00:02 +0000)]
Replace all weight-based interfaces in MBB with probability-based interfaces, and update all uses of old interfaces.
The patch in http://reviews.llvm.org/D13745 is broken into four parts:
1. New interfaces without functional changes (http://reviews.llvm.org/D13908).
2. Use new interfaces in SelectionDAG, while in other passes treat probabilities
as weights (http://reviews.llvm.org/D14361).
3. Use new interfaces in all other passes.
4. Remove old interfaces.
This patch is 3+4 above. In this patch, MBB won't provide weight-based
interfaces any more, which are totally replaced by probability-based ones.
The interface addSuccessor() is redesigned so that the default probability is
unknown. We allow unknown probabilities but don't allow using it together
with known probabilities in successor list. That is to say, we either have a
list of successors with all known probabilities, or all unknown
probabilities. In the latter case, we assume each successor has 1/N
probability where N is the number of successors. An assertion checks if the
user is attempting to add a successor with the disallowed mixed use as stated
above. This can help us catch many misuses.
All uses of weight-based interfaces are now updated to use probability-based
ones.
Differential revision: http://reviews.llvm.org/D14973
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254348
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 30 Nov 2015 23:54:19 +0000 (23:54 +0000)]
This reverts commit r254336 and r254344.
They broke a bot and I am debugging why.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254347
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 30 Nov 2015 23:05:25 +0000 (23:05 +0000)]
Disable a consistency check.
Trying to figure out why it fails on a bot but passes locally.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254344
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 30 Nov 2015 22:39:36 +0000 (22:39 +0000)]
[InstCombine] add tests to show potential vector IR shuffle transforms
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254342
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Mon, 30 Nov 2015 22:22:06 +0000 (22:22 +0000)]
[X86][FMA4] Prefer FMA4 to FMA
We currently output FMA instructions on targets which support both FMA4 + FMA (i.e. later Bulldozer CPUS bdver2/bdver3/bdver4).
This patch flips this so FMA4 is preferred; this is for several reasons:
1 - FMA4 is non-destructive reducing the need for mov instructions.
2 - Its more straighforward to commute and fold inputs (although the recent work on FMA has reduced this difference).
3 - All supported targets have FMA4 performance equal or better to FMA - Piledriver (bdver2) in particular has half the throughput when executing FMA instructions.
Its looks like no future AMD processor lines will support FMA4 after the Bulldozer series so we're not causing problems for later CPUs.
Differential Revision: http://reviews.llvm.org/D14997
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254339
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 30 Nov 2015 22:01:43 +0000 (22:01 +0000)]
Start deciding earlier what to link.
A traditional linker is roughly split in symbol resolution and "copying
stuff".
The two tasks are badly mixed in lib/Linker.
This starts splitting them apart.
With this patch there are no direct call to linkGlobalValueBody or
linkGlobalValueProto. Everything is linked via WapValue.
This also includes a few fixes:
* A GV goes undefined if the comdat is dropped (comdat11.ll).
* We error if an internal GV goes undefined (comdat13.ll).
* We don't link an unused comdat.
The first two match the behavior of an ELF linker. The second one is
equivalent to running globaldce on the input.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254336
91177308-0d34-0410-b5e6-
96231b3b80d8