Craig Topper [Sun, 27 Dec 2015 06:55:08 +0000 (06:55 +0000)]
[AVX-512] Remove alernate integer forms for VPERMILPS and VPERMILPD. There no tests for them and I don't see any way to select them anyway. If they are really needed they should be implemented as patterns and not full fledged instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256462
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Sun, 27 Dec 2015 06:26:41 +0000 (06:26 +0000)]
Make the test properly constrained
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256460
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 27 Dec 2015 06:18:57 +0000 (06:18 +0000)]
InstrProfTest.cpp: Don't assume string literals are always merged.
MSC18 Debug didn't merge them.
FIXME: I tweaked just to appease a builder. Almost string literals should be addressed identically there.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256459
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Sun, 27 Dec 2015 06:18:48 +0000 (06:18 +0000)]
Try to passify buildbot
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256458
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 27 Dec 2015 06:14:33 +0000 (06:14 +0000)]
Prune the feature "tls". No one is using it since TLS is enabled for Cygwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256457
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Sun, 27 Dec 2015 06:07:26 +0000 (06:07 +0000)]
[X86, Win64] Use a frame pointer if pushf is emitted
A frame pointer must be used if stack pointer is modified after the
prologue. LLVM will emit pushf/popf if we need to save/restore the
FLAGS register, requiring us to have a frame pointer for the function.
There is a small twist: this sequence might exist in user code via
inline-assembly. For now, conservatively assume that such functions
require a frame pointer. For real world justification, please see
clang's implementation of __readeflags.
This fixes PR25945.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256456
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Sun, 27 Dec 2015 06:07:12 +0000 (06:07 +0000)]
[WinEH] Add comments explaining the EH tables
This is aids in debugging WinEH, similar functionality is present for
DWARF EH.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256455
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Sat, 26 Dec 2015 21:44:55 +0000 (21:44 +0000)]
[x86] lower calls to llvm.maxnum.v4f32 using maxps
This is a follow-on to:
http://reviews.llvm.org/rL255700
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256454
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Dec 2015 20:13:33 +0000 (20:13 +0000)]
[X86] Fix an unused variable warning in released builds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256453
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Dec 2015 19:48:43 +0000 (19:48 +0000)]
[X86] Add support for printing shuffle comments for AVX512 PSHUFB instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256452
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Dec 2015 19:48:37 +0000 (19:48 +0000)]
[X86] Fold some variable declarations and initializations into if statements. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256451
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Dec 2015 11:40:48 +0000 (11:40 +0000)]
Fix safepoint intrinsic signatures in test.
Should bring back the bots after r256443.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256450
91177308-0d34-0410-b5e6-
96231b3b80d8
Chen Li [Sat, 26 Dec 2015 07:54:32 +0000 (07:54 +0000)]
[gc.statepoint] Change gc.statepoint intrinsic's return type to token type instead of i32 type
Summary: This patch changes gc.statepoint intrinsic's return type to token type instead of i32 type. Using token types could prevent LLVM to merge different gc.statepoint nodes into PHI nodes and cause further problems with gc relocations. The patch also changes the way on how gc.relocate and gc.result look for their corresponding gc.statepoint on unwind path. The current implementation uses the selector value extracted from a { i8*, i32 } landingpad as a hook to find the gc.statepoint, while the patch directly uses a token type landingpad (http://reviews.llvm.org/D15405) to find the gc.statepoint.
Reviewers: sanjoy, JosephTremoulet, pgavlin, igor-laevsky, mjacob
Subscribers: reames, mjacob, sanjoy, llvm-commits
Differential Revision: http://reviews.llvm.org/D15662
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256443
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Dec 2015 04:58:05 +0000 (04:58 +0000)]
Add test case for r256433. "[X86] Fix shuffle decoding for variable VPERMIL to be tolerant of the Constant type not matching due to folding in the constant pool and to get VPERMILPD correct."
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256435
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Dec 2015 04:56:51 +0000 (04:56 +0000)]
Revert r256432 "Test"
This is the test case for r256433, but it got committed incorrectly in my local repo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256434
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Dec 2015 04:50:07 +0000 (04:50 +0000)]
[X86] Fix shuffle decoding for variable VPERMIL to be tolerant of the Constant type not matching due to folding in the constant pool and to get VPERMILPD correct.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256433
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Dec 2015 04:50:01 +0000 (04:50 +0000)]
Test
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256432
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 23:27:57 +0000 (23:27 +0000)]
[X86] Fix copy and paste typo from pasting from another Makefile to restore code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256431
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 22:22:16 +0000 (22:22 +0000)]
[X86] Put back the include path to the main X86 sources in the AsmParser library to fix the bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256430
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 22:10:11 +0000 (22:10 +0000)]
[X86] Remove X86CodeGen dependency from the AsmParser library.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256429
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 22:10:08 +0000 (22:10 +0000)]
[X86] Move getX86SubSuperRegisterOrZero to X86MCTargetDesc.cpp so it can be used by AsmParser library without depending on X86CodeGen library.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256428
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 22:10:01 +0000 (22:10 +0000)]
Remove extra forward declarations and scrub includes for all in tree InstPrinters. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256427
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 22:09:49 +0000 (22:09 +0000)]
[X86] Move AVX512 STATIC_ROUNDING enum to X86BaseInfo.h to fix a layering violation in AsmParser.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256426
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 22:09:45 +0000 (22:09 +0000)]
[X86] Replace MVT::SimpleValueType in the AsmParser library and getX86SubSuperRegister with just an unsigned representing size.
This a is step towards fixing a layering violation so the X86 AsmParser won't depending on CodeGen types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256425
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 19:44:16 +0000 (19:44 +0000)]
[X86] Don't pass the default value to the High argument of getX86SubSuperRegister. Most place don't care about this argument. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256424
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Fri, 25 Dec 2015 18:16:45 +0000 (18:16 +0000)]
[llvm-objdump] Use stderr and not stdout for fatal errors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256423
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 17:07:32 +0000 (17:07 +0000)]
[X86] getX86SubSuperRegisterOrZero shouldn't call getX86SubSuperRegister recursively. It should call itself instead. Otherwise it might fire an assertion when it was designed not too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256422
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 17:07:30 +0000 (17:07 +0000)]
[X86] Add missing X86II::MRM_C4, MRM_C5, etc. encodings to getMemoryOperandNo. These aren't used by any instructions, but could be someday. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256421
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 17:07:27 +0000 (17:07 +0000)]
[X86] Use assert instead of if and llvm_unreachable. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256420
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 17:07:24 +0000 (17:07 +0000)]
[X86] Minor identation fixes. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256419
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Fri, 25 Dec 2015 09:37:26 +0000 (09:37 +0000)]
[CodeGen] Use generic printAsOperand machinery instead of hand rolling it
We already know how to properly print out basic blocks in
printAsOperand, we should not roll it ourselves in
AsmPrinter::EmitBasicBlockStart. No functionality change is intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256413
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 04:06:20 +0000 (04:06 +0000)]
[IR] Mark the Type subclass helper methods 'inline' and move their definitions to DerivedTypes.h so they can be inlined by the compiler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256406
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Dec 2015 02:04:17 +0000 (02:04 +0000)]
[Transforms] Use asserts instead of ifs around llvm_unreachable. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256405
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 25 Dec 2015 00:31:02 +0000 (00:31 +0000)]
[WebAssembly] Fix handling of COPY instructions in WebAssemblyRegStackify.
Move RegStackify after coalescing and teach it to use LiveIntervals instead
of depending on SSA form. This avoids a problem where a register in a COPY
instruction is stackified and then subsequently coalesced with a register
that is not stackified.
This also puts it after the scheduler, which allows us to simplify the
EXPR_STACK constraint, as we no longer have instructions being reordered
after stackification and before coloring.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256402
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 24 Dec 2015 21:17:56 +0000 (21:17 +0000)]
[InstCombine] transform more extract/insert pairs into shuffles (PR2109)
This is an extension of the shuffle combining from r203229:
http://reviews.llvm.org/rL203229
The idea is to widen a short input vector with undef elements so the
existing shuffle transform for extract/insert can kick in.
The motivation is to finally solve PR2109:
https://llvm.org/bugs/show_bug.cgi?id=2109
For that example, the IR becomes:
%1 = bitcast <2 x i32>* %P to <2 x float>*
%ld1 = load <2 x float>, <2 x float>* %1, align 8
%2 = shufflevector <2 x float> %ld1, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef>
%i2 = shufflevector <4 x float> %A, <4 x float> %2, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
ret <4 x float> %i2
And x86 SSE output improves from:
movq (%rdi), %xmm1 ## xmm1 = mem[0],zero
movdqa %xmm1, %xmm2
shufps $229, %xmm2, %xmm2 ## xmm2 = xmm2[1,1,2,3]
shufps $48, %xmm0, %xmm1 ## xmm1 = xmm1[0,0],xmm0[3,0]
shufps $132, %xmm1, %xmm0 ## xmm0 = xmm0[0,1],xmm1[0,2]
shufps $32, %xmm0, %xmm2 ## xmm2 = xmm2[0,0],xmm0[2,0]
shufps $36, %xmm2, %xmm0 ## xmm0 = xmm0[0,1],xmm2[2,0]
retq
To the almost optimal:
movhpd (%rdi), %xmm0
Note: There's a tension in the existing transform related to generating
arbitrary shufflevector masks. We avoid that in other places in InstCombine
because we're scared that codegen can't handle strange masks, but it looks
like we're ok with producing those here. I purposely chose weird insert/extract
indexes for the regression tests to see the effect in these cases.
For PowerPC+Altivec, AArch64, and X86+SSE/AVX, I think the codegen is equal or
better for these examples.
Differential Revision: http://reviews.llvm.org/D15096
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256394
91177308-0d34-0410-b5e6-
96231b3b80d8
Dave Bartolomeo [Thu, 24 Dec 2015 19:17:54 +0000 (19:17 +0000)]
Fix signed/unsigned warning in Line.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256390
91177308-0d34-0410-b5e6-
96231b3b80d8
Dave Bartolomeo [Thu, 24 Dec 2015 19:15:56 +0000 (19:15 +0000)]
Remove unused constants from TypeTableBuilder.cpp.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256389
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Seurer [Thu, 24 Dec 2015 18:54:35 +0000 (18:54 +0000)]
Fix case of path name
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256388
91177308-0d34-0410-b5e6-
96231b3b80d8
Dave Bartolomeo [Thu, 24 Dec 2015 18:51:35 +0000 (18:51 +0000)]
Fix CodeView library name and non-CMake builds
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256387
91177308-0d34-0410-b5e6-
96231b3b80d8
Dave Bartolomeo [Thu, 24 Dec 2015 18:25:54 +0000 (18:25 +0000)]
Fix initializer order warning in TypeRecord.h
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256386
91177308-0d34-0410-b5e6-
96231b3b80d8
Dave Bartolomeo [Thu, 24 Dec 2015 18:12:38 +0000 (18:12 +0000)]
LLVM CodeView library
Summary: This diff is the initial implementation of the LLVM CodeView library. There is much more work to be done, namely a CodeView dumper and tests. This patch should help others make progress on the LLVM->CodeView debug info emission while I continue with the implementation of the dumper and tests.
This library implements support for emitting debug info in the CodeView format. This phase of the implementation only includes support for CodeView type records. Clients that need to emit type records will use a class derived from TypeTableBuilder. TypeTableBuilder provides member functions for writing each kind of type record; each of these functions eventually calls the writeRecord virtual function to emit the actual bits of the record. Derived classes override writeRecord to implement the folding of duplicate records and the actual emission to the appropriate destination. LLVMCodeView provides MemoryTypeTableBuilder, which creates the table in memory. In the future, other classes derived from TypeTableBuilder will write to other destinations, such as the type stream in a PDB.
The rest of the types in LLVMCodeView define the actual CodeView type records and all of the supporting enums and other types used in the type records. The TypeIndex class is of particular interest, because it is used by clients as a handle to a type in the type table.
The library provides a relatively low-level interface based on the actual on-disk format of CodeView. For example, type records refer to other type records by TypeIndex, rather than by an actual pointer to the referent record. This allows clients to emit type records one at a time, rather than having to keep the entire transitive closure of type records in memory until everything has been emitted. At some point, having a higher-level interface layered on top of this one may be useful for debuggers and other tools that want a more holistic view of the debug info. The lower-level interface should be sufficient for compilers and linkers to do the debug info manipulation that they need to do efficiently.
Reviewers: rnk, majnemer
Subscribers: silvas, rnk, jevinskie, llvm-commits
Differential Revision: http://reviews.llvm.org/D14961
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256385
91177308-0d34-0410-b5e6-
96231b3b80d8
Marina Yatsina [Thu, 24 Dec 2015 12:09:51 +0000 (12:09 +0000)]
[X86][ms-inline asm] Add support for memory operands that include structs
Add ability to reference struct symbols in memory operands.
Test case will be added on the clang side (review http://reviews.llvm.org/D15749)
Differential Revision: http://reviews.llvm.org/D15748
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256381
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 24 Dec 2015 10:03:37 +0000 (10:03 +0000)]
[ProfileData] Make helper function static.
No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256375
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 24 Dec 2015 10:03:35 +0000 (10:03 +0000)]
[FunctionImport] Move pass into anonymous namespace.
No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256374
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Dec 2015 09:08:08 +0000 (09:08 +0000)]
Add a missing const qualifier on the context instruction. This somehow
has always been missing. =/
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256371
91177308-0d34-0410-b5e6-
96231b3b80d8
Asaf Badouh [Thu, 24 Dec 2015 08:25:00 +0000 (08:25 +0000)]
[X86][PKU] Add {RD,WR}PKRU encoding
Differential Revision: http://reviews.llvm.org/D15711
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256366
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Thu, 24 Dec 2015 08:12:22 +0000 (08:12 +0000)]
AVX-512: Kreg set 0/1 optimization
The patterns that set a mask register to 0/1
KXOR %kn, %kn, %kn / KXNOR %kn, %kn, %kn
are replaced with
KXOR %k0, %k0, %kn / KXNOR %k0, %k0, %kn - AVX-512 targets optimization.
KNL does not recognize dependency-breaking idioms for mask registers,
so kxnor %k1, %k1, %k2 has a RAW dependence on %k1.
Using %k0 as the undef input register is a performance heuristic based
on the assumption that %k0 is used less frequently than the other mask
registers, since it is not usable as a write mask.
Differential Revision: http://reviews.llvm.org/D15739
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256365
91177308-0d34-0410-b5e6-
96231b3b80d8
Igor Breger [Thu, 24 Dec 2015 07:11:53 +0000 (07:11 +0000)]
AVX512: VPMOVM2B/W/D/Q intrinsic implementation.
Differential Revision: http://reviews.llvm.org//D15747
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256364
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 24 Dec 2015 05:20:40 +0000 (05:20 +0000)]
Use range-based for loops. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256363
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 24 Dec 2015 05:14:55 +0000 (05:14 +0000)]
AMDGPU: Fix getRegisterBitWidth for vectors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256362
91177308-0d34-0410-b5e6-
96231b3b80d8
Nico Weber [Thu, 24 Dec 2015 04:01:06 +0000 (04:01 +0000)]
Revert r256336, it caused PR25939
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256361
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Thu, 24 Dec 2015 03:18:18 +0000 (03:18 +0000)]
AMDGPU/SI: Fix encoding of flat instructions on VI
Reviewers: arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D15735
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256360
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Thu, 24 Dec 2015 02:41:55 +0000 (02:41 +0000)]
AMDGPU/SI: Remove non-existent flat instructions
Reviewers: arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D15734
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256357
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 24 Dec 2015 02:31:20 +0000 (02:31 +0000)]
Address Sanjoy's review comments to r256326
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256356
91177308-0d34-0410-b5e6-
96231b3b80d8
JF Bastien [Wed, 23 Dec 2015 23:56:13 +0000 (23:56 +0000)]
WebAssembly: remove 'external' from test
Summary: Linker testing was sad at seeing an unresolved external symbol. For now don't do that: it's valid but we're not playing with multi-file linking yet, and the LLVM tests are used as hacky sanity tests for single-file linking (the GCC torture tests are much better for this purpose). Another solution would be to use '.extern' to make the intent explicit (don't simple-file link this, there's an unresolved symbol), some assemblers use '.extern' while others ignore it, so we wouldn't really be inventing anything new.
Reviewers: sunfish, kripken
Subscribers: jfb, llvm-commits, dschuff
Differential Revision: http://reviews.llvm.org/D15753
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256353
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 23 Dec 2015 23:44:28 +0000 (23:44 +0000)]
[Statepoints] Use Indirect operands for spill slots
Teach the statepoint lowering code to emit Indirect stackmap entries for spill inserted by StatepointLowering (i.e. SelectionDAG), but Direct stackmap entries for in-IR allocas which represent manual stack slots. This is what the docs call for (http://llvm.org/docs/StackMaps.html#stack-map-format), but we've been emitting both as Direct. This was pointed out recently on the mailing list as a bug. It also blocks http://reviews.llvm.org/D15632 which extends the lowering to handle vector-of-pointers since only Indirect references can encode a variable sized slot.
To implement this, I introduced a new flag on the StackObject class used to maintian information about stack slots. I original considered (and prototyped in http://reviews.llvm.org/D15632), the idea of using the existing isSpillSlot flag, but end up deciding that was a bit too risky and that the cost of adding a new flag was low. Having the new flag will also allow us - in the future - to emit better comments in verbose assembly which indicate where a particular stack spill around a call comes from. (deopt, gc, regalloc).
Differential Revision: http://reviews.llvm.org/D15759
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256352
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Wed, 23 Dec 2015 21:51:13 +0000 (21:51 +0000)]
llvm-dwarfdump: Add support for dumping .dSYM bundles.
This replicates the logic of Darwin dwarfdump for manually opening up
.dSYM bundles without introducing any new dependencies.
<rdar://problem/
20491670>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256350
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Wed, 23 Dec 2015 19:29:34 +0000 (19:29 +0000)]
[llvm-readobj] Use stderr and not stdout for error messages.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256347
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 23 Dec 2015 19:16:04 +0000 (19:16 +0000)]
[MemOperands] Clarify code around dropping memory operands [NFC]
Clarify a comment about what it means to drop memory operands from an instruction. While I'm adding change the name of the method slightly to make it a bit more clear what's going on when reading calling code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256346
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Wed, 23 Dec 2015 18:27:23 +0000 (18:27 +0000)]
[Function] Properly remove use when clearing personality
Summary:
We need to actually remove the use of the personality function,
otherwise we can run into trouble if we want to e.g. delete
the personality function because ther's no way to get rid of
its uses. Do this by resetting to ConstantPointerNull value
that the operands are set to when first allocated.
Reviewers: vsk, dexonsmith
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15752
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256345
91177308-0d34-0410-b5e6-
96231b3b80d8
JF Bastien [Wed, 23 Dec 2015 18:18:53 +0000 (18:18 +0000)]
Fix SCEV r256338.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256344
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Wed, 23 Dec 2015 17:48:14 +0000 (17:48 +0000)]
[SCEV] Fix getLoopBackedgeTakenCounts
The way `getLoopBackedgeTakenCounts` is written right now isn't
correct. It will try to compute and store the BE counts of a Loop
#{child loop} number of times (which may be zero).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256338
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 23 Dec 2015 17:29:33 +0000 (17:29 +0000)]
[LIR] General refactoring to simplify code and the ease future code review.
Move several checks into isLegalStores. Also, delineate between those stores
that are memset-able and those that are memcpy-able.
http://reviews.llvm.org/D15683
Patch by Haicheng Wu <haicheng@codeaurora.org>!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256336
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 23 Dec 2015 17:05:57 +0000 (17:05 +0000)]
[MachineLICM] Fix handling of memoperands
As far as I can tell, the correct interpretation of an empty memoperands list is that we didn't have sufficient room to store information about the MachineInstr, NOT that the MachineInstr doesn't access any particular bit of memory. This appears to be fairly consistent in a number of places, but I'm not 100% sure of this interpretation. I'd really appreciate someone more knowledgeable confirming my reading of the code.
This patch fixes two latent bugs in MachineLICM - given the above assumption - and adds comments to document the meaning and required handling. I don't have test cases; these were noticed by inspection.
Differential Revision: http://reviews.llvm.org/D15730
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256335
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 23 Dec 2015 13:10:07 +0000 (13:10 +0000)]
[X86][AVX] Only shuffle the lower half of vectors if the upper half is undefined
First step towards making better use of AVX's implicit zeroing of the upper half of a 256-bit vector by instructions that only act on the lower 128-bit vector - discussed on D14151.
As well as the fact that 128-bit shuffle instructions are generally more capable, this can be performant for older CPUs with 128-bit ALUs (e.g. Jaguar, Sandy Bridge) that must treat 256-bit vectors as multiple micro-ops.
Moved the similar subvector extraction shuffle combines from PerformShuffleCombine256 to lowerVectorShuffle as well.
Note: I've avoided combining shuffles that reference elements from the upper halves of the input vectors - this may be reviewed in future work as well (AVX1 would probably always gain, but AVX2 does have some cross-lane shuffle instructions).
Differential Revision: http://reviews.llvm.org/D15477
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256332
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Wed, 23 Dec 2015 09:58:46 +0000 (09:58 +0000)]
[OperandBundles] Have GlobalsModRef play nice with operand bundles
A call site's use of a Value might not correspond to an argument
operand but to a bundle operand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256329
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Wed, 23 Dec 2015 09:58:43 +0000 (09:58 +0000)]
[OperandBundles] Have TailCallElim play nice with operand bundles
A call site's use of a Value might not correspond to an argument
operand but to a bundle operand.
This fixes PR25928.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256328
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Wed, 23 Dec 2015 09:58:41 +0000 (09:58 +0000)]
[OperandBundles] Have InstCombine play nice with operand bundles
Don't assume a call's use corresponds to an argument operand, it might
correspond to a bundle operand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256327
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Wed, 23 Dec 2015 09:58:36 +0000 (09:58 +0000)]
[OperandBundles] Have DeadArgElim play nice with operand bundles
A call site's use of a Value might not correspond to an argument
operand but to a bundle operand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256326
91177308-0d34-0410-b5e6-
96231b3b80d8
Igor Breger [Wed, 23 Dec 2015 08:06:50 +0000 (08:06 +0000)]
AVX512BW: Enable packed word shift for 512bit vector. Enable lowering scalar immidiate shift v64i8 .Fix predicate for AVX1/2 shifts.
Differential Revision: http://reviews.llvm.org/D15713
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256324
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Wed, 23 Dec 2015 03:59:04 +0000 (03:59 +0000)]
[WinEH] Don't visit the same catchswitch twice
We visited the same catchswitch twice because it was both the child of
another funclet and the predecessor of a cleanuppad.
Instead, change the numbering algorithm to only recurse if the unwind
destination of the inner funclet agrees with the unwind destination of
the catchswitch.
This fixes PR25926.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256317
91177308-0d34-0410-b5e6-
96231b3b80d8
Nico Weber [Wed, 23 Dec 2015 02:38:31 +0000 (02:38 +0000)]
win: Pass /W4 in front of all the -wd flags.
This should fix many many -Wunused-parameter warnings in self-host builds on
Windows after r255382. cl.exe doesn't care about the order of /W4 and
/wd flags, but clang-cl currently does (just like -Wno-foo -Wall order
matters for clang). We might want to change how clang-cl behaves in
the future, but until then this change makes self-host builds much more
silent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256315
91177308-0d34-0410-b5e6-
96231b3b80d8
Paul Robinson [Wed, 23 Dec 2015 01:57:31 +0000 (01:57 +0000)]
Form reform for MCDwarf.
MCDwarf emits a canned abbreviation table, but was not emitting proper
forms for DWARF version 4, which is the default after r249655.
Differential Revision: http://reviews.llvm.org/D15732
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256313
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 23 Dec 2015 01:42:15 +0000 (01:42 +0000)]
[GC] Make GCStrategy::isGCManagedPointer a type predicate not a value predicate [NFC]
Reasons:
1) The existing form was a form of false generality. None of the implemented GCStrategies use anything other than a type. Its becoming more and more clear we're going to need some type of strong GC pointer in the type system and we shouldn't pretend otherwise at this point.
2) The API was awkward when applied to vectors-of-pointers. The old one could have been made to work, but calling isGCManagedPointer(Ty->getScalarType()) is much cleaner than the Value alternatives.
3) The rewriting implementation effectively assumes the type based predicate as well. We should be consistent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256312
91177308-0d34-0410-b5e6-
96231b3b80d8
Nico Weber [Wed, 23 Dec 2015 01:04:53 +0000 (01:04 +0000)]
Unbreak LLVM_ENABLE_THREADS=OFF builds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256308
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Wed, 23 Dec 2015 00:22:04 +0000 (00:22 +0000)]
[WebAssembly] Add a TODO comment for a possible future optimization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256306
91177308-0d34-0410-b5e6-
96231b3b80d8
Manuel Jacob [Wed, 23 Dec 2015 00:19:45 +0000 (00:19 +0000)]
[RS4GC] Fix base pair printing for constants.
Previously, "%" + name of the value was printed for each derived and base
pointer. This is correct for instructions, but wrong for e.g. globals.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256305
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 22 Dec 2015 23:57:37 +0000 (23:57 +0000)]
Provide a way to specify inliner's attribute compatibility and merging.
This reapplies r256277 with two changes:
- In emitFnAttrCompatCheck, change FuncName's type to std::string to fix
a use-after-free bug.
- Remove an unnecessary install-local target in lib/IR/Makefile.
Original commit message for r252949:
Provide a way to specify inliner's attribute compatibility and merging
rules using table-gen. NFC.
This commit adds new classes CompatRule and MergeRule to Attributes.td,
which are used to generate code to check attribute compatibility and
merge attributes of the caller and callee.
rdar://problem/
19836465
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256304
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Tue, 22 Dec 2015 23:45:55 +0000 (23:45 +0000)]
[BPI] Fix two potential divide-by-zero operations that are introduced in r256263.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256303
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 22 Dec 2015 23:45:49 +0000 (23:45 +0000)]
Disable use list order on the gold plugin.
It turns out that his is *really* slow. With this change the link of
clang with plugin-opt=emit-llvm goes from 41 to 26 seconds.
We can add an option to enable it again if needed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256302
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 22 Dec 2015 23:45:21 +0000 (23:45 +0000)]
[WebAssembly] Trim unneeded #includes. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256301
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 22 Dec 2015 23:39:16 +0000 (23:39 +0000)]
[WebAssembly] Minor code simplification. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256300
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 22 Dec 2015 23:37:37 +0000 (23:37 +0000)]
Add an OperandNamespace field to Target.td's Operand.
For targets to add their own operand types as needed, as advertised in
Operand's comment, they need to be able to specify an alternate namespace
for OperandType names too. This matches the RegisterOperand class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256299
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Tue, 22 Dec 2015 23:12:41 +0000 (23:12 +0000)]
[unittest] Use Support/thread.h instead of <thread> (second try)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256292
91177308-0d34-0410-b5e6-
96231b3b80d8
Nico Weber [Tue, 22 Dec 2015 23:12:00 +0000 (23:12 +0000)]
lit: Limit number of processes on Windows to 32.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256291
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Tue, 22 Dec 2015 23:09:08 +0000 (23:09 +0000)]
[unittest] Use Support/Thread.h instead of <thread> to fix the Windows build
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256290
91177308-0d34-0410-b5e6-
96231b3b80d8
Changpeng Fang [Tue, 22 Dec 2015 20:55:23 +0000 (20:55 +0000)]
AMDGPU/SI: Use flat for global load/store when targeting HSA
Summary:
For some reason doing executing an MUBUF instruction with the addr64
bit set and a zero base pointer in the resource descriptor causes
the memory operation to be dropped when the shader is executed using
the HSA runtime.
This kind of MUBUF instruction is commonly used when the pointer is
stored in VGPRs. The base pointer field in the resource descriptor
is set to zero and and the pointer is stored in the vaddr field.
This patch resolves the issue by only using flat instructions for
global memory operations when targeting HSA. This is an overly
conservative fix as all other configurations of MUBUF instructions
appear to work.
NOTE: re-commit by fixing a failure in Codegen/AMDGPU/llvm.dbg.value.ll
Reviewers: tstellarAMD
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D15543
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256282
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 22 Dec 2015 20:43:30 +0000 (20:43 +0000)]
Also add unnamed_addr to functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256281
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 22 Dec 2015 20:29:09 +0000 (20:29 +0000)]
Revert r256277 and r256279.
Some of the bots failed again.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256280
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 22 Dec 2015 20:06:50 +0000 (20:06 +0000)]
Add a .td file I forgot to add in r256277.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256279
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 22 Dec 2015 20:00:05 +0000 (20:00 +0000)]
Provide a way to specify inliner's attribute compatibility and merging.
This reapplies r252990 and r252949. I've added member function getKind
to the Attr classes which returns the enum or string of the attribute.
Original commit message for r252949:
Provide a way to specify inliner's attribute compatibility and merging
rules using table-gen. NFC.
This commit adds new classes CompatRule and MergeRule to Attributes.td,
which are used to generate code to check attribute compatibility and
merge attributes of the caller and callee.
rdar://problem/
19836465
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256277
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 22 Dec 2015 19:50:22 +0000 (19:50 +0000)]
Delete dead GlobalAliases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256276
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 22 Dec 2015 19:46:44 +0000 (19:46 +0000)]
Revert "AMDGPU/SI: Use flat for global load/store when targeting HSA"
This reverts commit r256273.
It broke CodeGen/AMDGPU/llvm.dbg.value.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256275
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 22 Dec 2015 19:38:07 +0000 (19:38 +0000)]
Merge duplicated code.
The code for deleting dead global variables and functions was
duplicated.
This is in preparation for also deleting dead global aliases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256274
91177308-0d34-0410-b5e6-
96231b3b80d8
Changpeng Fang [Tue, 22 Dec 2015 19:32:28 +0000 (19:32 +0000)]
AMDGPU/SI: Use flat for global load/store when targeting HSA
Summary:
For some reason doing executing an MUBUF instruction with the addr64
bit set and a zero base pointer in the resource descriptor causes
the memory operation to be dropped when the shader is executed using
the HSA runtime.
This kind of MUBUF instruction is commonly used when the pointer is
stored in VGPRs. The base pointer field in the resource descriptor
is set to zero and and the pointer is stored in the vaddr field.
This patch resolves the issue by only using flat instructions for
global memory operations when targeting HSA. This is an overly
conservative fix as all other configurations of MUBUF instructions
appear to work.
Reviewers: tstellarAMD
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D15543
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256273
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 22 Dec 2015 19:26:18 +0000 (19:26 +0000)]
Use early continue to reduce indentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256272
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Tue, 22 Dec 2015 19:16:50 +0000 (19:16 +0000)]
Simplify iterator management. NFC.
Not passing an iterator to processGlobal will allow it to work with
other GlobalValues.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256271
91177308-0d34-0410-b5e6-
96231b3b80d8
Paul Robinson [Tue, 22 Dec 2015 18:59:02 +0000 (18:59 +0000)]
Add advice on choosing reviewers
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256265
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Tue, 22 Dec 2015 18:56:14 +0000 (18:56 +0000)]
[BPI] Replace weights by probabilities in BPI.
This patch removes all weight-related interfaces from BPI and replace
them by probability versions. With this patch, we won't use edge weight
anymore in either IR or MC passes. Edge probabilitiy is a better
representation in terms of CFG update and validation.
Differential revision: http://reviews.llvm.org/D15519
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256263
91177308-0d34-0410-b5e6-
96231b3b80d8
Manuel Jacob [Tue, 22 Dec 2015 18:44:45 +0000 (18:44 +0000)]
Remove deprecated llvm.experimental.gc.result.{int,float,ptr} intrinsics.
Summary:
These were deprecated 11 months ago when a generic
llvm.experimental.gc.result intrinsic, which works for all types, was added.
Reviewers: sanjoy, reames
Subscribers: sanjoy, chenli, llvm-commits
Differential Revision: http://reviews.llvm.org/D15719
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256262
91177308-0d34-0410-b5e6-
96231b3b80d8