Ahmed Bougacha [Wed, 21 Aug 2013 07:28:24 +0000 (07:28 +0000)]
MC CFG: Split MCBasicBlocks to mirror atom splitting.
When an MCTextAtom is split, all MCBasicBlocks backed by it are
automatically split, with a fallthrough between both blocks, and
the successors moved to the second block.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188881
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:28:17 +0000 (07:28 +0000)]
MC CFG: Add a few needed methods, mainly MCModule::findFirstAtomAfter.
While there, do some minor cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188880
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:28:13 +0000 (07:28 +0000)]
MC: ObjectSymbolizer can now recognize external function stubs.
Only implemented in the Mach-O ObjectSymbolizer.
The testcase sadly introduces a new binary.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188879
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:28:07 +0000 (07:28 +0000)]
MC: Refactor ObjectSymbolizer to make relocation/section info generation lazy.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188878
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:28:02 +0000 (07:28 +0000)]
MC CFG: Add entrypoint address to MCModule.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188877
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:27:59 +0000 (07:27 +0000)]
MC CFG: Add more MCFunction container methods (find, empty).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188876
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:27:55 +0000 (07:27 +0000)]
MC CFG: Keep pointer to parent MCModule in created MCFunctions.
Also, drive-by cleaning around createFunction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188875
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:27:50 +0000 (07:27 +0000)]
MC CFG: Don't insert preds/succs again.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188874
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:27:47 +0000 (07:27 +0000)]
MC CFG: Remap enough for the inserted instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188873
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:27:44 +0000 (07:27 +0000)]
MC CFG: uint64_t -> size_t for vector size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188872
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 21 Aug 2013 07:27:40 +0000 (07:27 +0000)]
MC CFG: Add a getter for MCDataAtom's data array.
While there, switch to new-style documentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188871
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Wed, 21 Aug 2013 06:13:34 +0000 (06:13 +0000)]
DebugInfo: Do not use the DWARF Version for the .debug_pubnames or .debug_pubtypes version field
Summary:
LLVM would generate DWARF with version 3 in the .debug_pubname and
.debug_pubtypes version fields. This would lead SGI dwarfdump to fail
parsing the DWARF with (in the instance of .debug_pubnames) would exit
with:
dwarfdump ERROR: dwarf_get_globals: DW_DLE_PUBNAMES_VERSION_ERROR (123)
This fixes PR16950.
Reviewers: echristo, dblaikie
Reviewed By: echristo
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D1454
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188869
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 21 Aug 2013 05:57:45 +0000 (05:57 +0000)]
Synchronize VEX JIT encoding code with the MCJIT version. Fix a bug in the MCJIT code where CurOp was being incremented even if the operand it was pointing at wasn't used. Maybe only matters if there are any EVEX_K instructions that aren't VEX_4V.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188868
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Wed, 21 Aug 2013 05:03:10 +0000 (05:03 +0000)]
In LLVM FMA3 operands are dst, src1, src2, src3, however dst is not encoded as it is always src1. This was causing the encoding of the operands to be off by one.
Patch by Chris Bieneman.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188866
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Wed, 21 Aug 2013 05:02:12 +0000 (05:02 +0000)]
Add the FMA3 feature in order to test FMA encoding using the old jit.
Patch by Chris Bieneman!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188865
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 21 Aug 2013 03:57:57 +0000 (03:57 +0000)]
Rename mattr names for AVX-512 to from avx-512 -> avx512f, avx-512-pfi -> av512pf, avx-512-cdi -> avx512cd, avx-512-eri->avx512er. This matches better with official docs and what gcc patches appearto be using. I didn't touch the has* functions or the feature flag names to avoid change the td and lowering file while commits are still happening.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188859
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 21 Aug 2013 02:37:25 +0000 (02:37 +0000)]
X86TargetMachine.cpp: Clarify to emit GOT in i686-{cygming|win32}-elf for mcjit.
I suppose all "lli -use-mcjit i686-*" should require GOT, (and to fail.)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188856
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 21 Aug 2013 02:37:14 +0000 (02:37 +0000)]
lli/RecordingMemoryManager.cpp: Make it complain if _GLOBAL_OFFSET_TABLE_ were not provided.
FIXME: Would it be responsible to provide GOT?
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188855
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Wed, 21 Aug 2013 01:20:11 +0000 (01:20 +0000)]
Move #includes from .h to .cpp file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188852
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Wed, 21 Aug 2013 01:18:46 +0000 (01:18 +0000)]
[micromips] Print instruction alias "not" if the last operand of a nor is zero.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188851
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 20 Aug 2013 23:52:00 +0000 (23:52 +0000)]
Move registering the execution of a basic block to the beginning rather than the end.
There are situations which can affect the correctness (or at least expectation)
of the gcov output. For instance, if a call to __gcov_flush() occurs within a
block before the execution count is registered and then the program aborts in
some way, then that block will not be marked as executed. This is not normally
what the user expects.
If we move the code that's registering when a block is executed to the
beginning, we can catch these types of situations.
PR16893
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188849
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 20 Aug 2013 23:47:25 +0000 (23:47 +0000)]
[mips] Add support for mfhc1 and mthc1.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188848
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 20 Aug 2013 23:38:40 +0000 (23:38 +0000)]
[mips] Add support for calling convention CC_MipsO32_FP64, which is used when the
size of floating point registers is 64-bit.
Test case will be added when support for mfhc1 and mthc1 is added.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188847
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 20 Aug 2013 23:21:55 +0000 (23:21 +0000)]
[mips] Remove predicates that were incorrectly or unnecessarily added.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188845
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Tue, 20 Aug 2013 23:04:15 +0000 (23:04 +0000)]
Add some constantness.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188844
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Tue, 20 Aug 2013 23:00:25 +0000 (23:00 +0000)]
Use -disable-output and to suppress output and don't use a temporary file unless we need one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188843
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 20 Aug 2013 22:58:56 +0000 (22:58 +0000)]
[mips] Define register class FGRH32 for the high half of the 64-bit floating
point registers. We will need this register class later when we add
definitions for instructions mfhc1 and mthc1. Also, remove sub-register indices
sub_fpeven and sub_fpodd and use sub_lo and sub_hi instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188842
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Tue, 20 Aug 2013 22:52:02 +0000 (22:52 +0000)]
Fix include guards.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188841
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Tue, 20 Aug 2013 21:21:45 +0000 (21:21 +0000)]
SLPVectorizer: Fix invalid iterator errors
Update iterator when the SLP vectorizer changes the instructions in the basic
block by restarting the traversal of the basic block.
Patch by Yi Jiang!
Fixes PR 16899.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188832
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Tue, 20 Aug 2013 21:20:04 +0000 (21:20 +0000)]
Teach ConstantFolding about pointer address spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188831
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 20 Aug 2013 21:08:22 +0000 (21:08 +0000)]
[mips] Resolve register classes dynamically using ptr_rc to reduce the number of
load/store instructions defined. Previously, we were defining load/store
instructions for each pointer size (32 and 64-bit), but now we need just one
definition.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188830
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Tue, 20 Aug 2013 20:53:09 +0000 (20:53 +0000)]
Add an option which permits the user to specify using a bitmask, that various
functions be compiled as mips32, without having to add attributes. This
is useful in certain situations where you don't want to have to edit the
function attributes in the source. For now it's only an option used for
the compiler developers when debugging the mips16 port.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188826
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Tue, 20 Aug 2013 20:46:51 +0000 (20:46 +0000)]
[mips] Guard micromips instructions with predicate InMicroMips. Also, fix
assembler predicate HasStdEnd so that it is false when the target is micromips.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188824
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Tue, 20 Aug 2013 19:12:42 +0000 (19:12 +0000)]
ARM: Fix fast-isel copy/paste-o.
Update testcase to be more careful about checking register
values. While regexes are general goodness for these sorts of
testcases, in this example, the registers are constrained by
the calling convention, so we can and should check their
explicit values.
rdar://
14779513
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188819
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 18:13:48 +0000 (18:13 +0000)]
Still more MCJIT PIC test XFAILs
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188815
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 17:01:35 +0000 (17:01 +0000)]
Clarifying two MCJIT PIC tests as XFAIL on i686-pc-linux
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188814
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 16:42:22 +0000 (16:42 +0000)]
Removing duplicate XFAIL markers
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188812
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 15:47:04 +0000 (15:47 +0000)]
Marking two more MCJIT PIC tests as XFAIL on i686
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188808
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 15:36:04 +0000 (15:36 +0000)]
Marking MCJIT PIC tests as XFAIL on arm
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188807
91177308-0d34-0410-b5e6-
96231b3b80d8
Vladimir Medic [Tue, 20 Aug 2013 13:33:18 +0000 (13:33 +0000)]
Fix style issues in AsmParser.cpp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188798
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Tue, 20 Aug 2013 11:00:29 +0000 (11:00 +0000)]
AVX-512: Added more patterns for VMOVSS, VMOVSD, VMOVD, VMOVQ
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188786
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Tue, 20 Aug 2013 09:41:47 +0000 (09:41 +0000)]
[mips][msa] Removed fcge, fcgt, fsge, fsgt
These instructions were present in a draft spec but were removed before
publication.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188782
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Tue, 20 Aug 2013 09:40:35 +0000 (09:40 +0000)]
[SystemZ] Update README
We now use MVST, CLST and SRST for the obvious cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188781
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Tue, 20 Aug 2013 09:38:48 +0000 (09:38 +0000)]
[SystemZ] Use SRST to optimize memchr
SystemZTargetLowering::emitStringWrapper() previously loaded the character
into R0 before the loop and made R0 live on entry. I'd forgotten that
allocatable registers weren't allowed to be live across blocks at this stage,
and it confused LiveVariables enough to cause a miscompilation of f3 in
memchr-02.ll.
This patch instead loads R0 in the loop and leaves LICM to hoist it
after RA. This is actually what I'd tried originally, but I went for
the manual optimisation after noticing that R0 often wasn't being hoisted.
This bug forced me to go back and look at why, now fixed as r188774.
We should also try to optimize null checks so that they test the CC result
of the SRST directly. The select between null and the SRST GPR result could
then usually be deleted as dead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188779
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Tue, 20 Aug 2013 09:27:31 +0000 (09:27 +0000)]
memcmp is not a valid way to compare structs with padding in them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188778
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Tue, 20 Aug 2013 09:22:54 +0000 (09:22 +0000)]
[mips][msa] Added insve
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188777
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Tue, 20 Aug 2013 09:14:46 +0000 (09:14 +0000)]
Fix test typo and add usual "br %r14" test
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188775
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Tue, 20 Aug 2013 09:11:13 +0000 (09:11 +0000)]
Fix overly pessimistic shortcut in post-RA MachineLICM
Post-RA LICM keeps three sets of registers: PhysRegDefs, PhysRegClobbers
and TermRegs. When it sees a definition of R it adds all aliases of R
to the corresponding set, so that when it needs to test for membership
it only needs to test a single register, rather than worrying about
aliases there too. E.g. the final candidate loop just has:
unsigned Def = Candidates[i].Def;
if (!PhysRegClobbers.test(Def) && ...) {
to test whether register Def is multiply defined.
However, there was also a shortcut in ProcessMI to make sure we didn't
add candidates if we already knew that they would fail the final test.
This shortcut was more pessimistic than the final one because it
checked whether _any alias_ of the defined register was multiply defined.
This is too conservative for targets that define register pairs.
E.g. on z, R0 and R1 are sometimes used as a pair, so there is a
128-bit register that aliases both R0 and R1. If a loop used
R0 and R1 independently, and the definition of R0 came first,
we would be able to hoist the R0 assignment (because that used
the final test quoted above) but not the R1 assignment (because
that meant we had two definitions of the paired R0/R1 register
and would fail the shortcut in ProcessMI).
This patch just uses the same check for the ProcessMI shortcut as
we use in the final candidate loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188774
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Tue, 20 Aug 2013 08:57:11 +0000 (08:57 +0000)]
ARM: implement some simple f64 materializations.
Previously we used a const-pool load for virtually all 64-bit floating values.
Actually, we can get quite a few common values (including 0.0, 1.0) via "vmov"
instructions of one stripe or another.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188773
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 08:56:28 +0000 (08:56 +0000)]
[stackprotector] Small cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188772
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 08:56:26 +0000 (08:56 +0000)]
[stackprotector] Small Bit of computation hoisting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188771
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 08:56:23 +0000 (08:56 +0000)]
[stackprotector] Added significantly longer comment to FindPotentialTailCall to make clear its relationship to llvm::isInTailCallPosition.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188770
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 08:46:16 +0000 (08:46 +0000)]
Removed trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188769
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 08:46:13 +0000 (08:46 +0000)]
[stackprotector] Removed stale TODO.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188768
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Tue, 20 Aug 2013 08:38:21 +0000 (08:38 +0000)]
[mips][msa] Added and.v, bmnz.v, bmz.v, bsel.v, nor.v, or.v, xor.v
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188767
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 08:36:53 +0000 (08:36 +0000)]
[stackprotector] Added support for emitting the llvm intrinsic stack protector check.
rdar://
13935163
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188766
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 08:36:50 +0000 (08:36 +0000)]
[stackprotector] Refactor out the end of isInTailCallPosition into the function returnTypeIsEligibleForTailCall.
This allows me to use returnTypeIsEligibleForTailCall in the stack protector pass.
rdar://
13935163
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188765
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 07:17:27 +0000 (07:17 +0000)]
Remove unused variables that crept in.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188761
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Tue, 20 Aug 2013 07:00:16 +0000 (07:00 +0000)]
Teach selectiondag how to handle the stackprotectorcheck intrinsic.
Previously, generation of stack protectors was done exclusively in the
pre-SelectionDAG Codegen LLVM IR Pass "Stack Protector". This necessitated
splitting basic blocks at the IR level to create the success/failure basic
blocks in the tail of the basic block in question. As a result of this,
calls that would have qualified for the sibling call optimization were no
longer eligible for optimization since said calls were no longer right in
the "tail position" (i.e. the immediate predecessor of a ReturnInst
instruction).
Then it was noticed that since the sibling call optimization causes the
callee to reuse the caller's stack, if we could delay the generation of
the stack protector check until later in CodeGen after the sibling call
decision was made, we get both the tail call optimization and the stack
protector check!
A few goals in solving this problem were:
1. Preserve the architecture independence of stack protector generation.
2. Preserve the normal IR level stack protector check for platforms like
OpenBSD for which we support platform specific stack protector
generation.
The main problem that guided the present solution is that one can not
solve this problem in an architecture independent manner at the IR level
only. This is because:
1. The decision on whether or not to perform a sibling call on certain
platforms (for instance i386) requires lower level information
related to available registers that can not be known at the IR level.
2. Even if the previous point were not true, the decision on whether to
perform a tail call is done in LowerCallTo in SelectionDAG which
occurs after the Stack Protector Pass. As a result, one would need to
put the relevant callinst into the stack protector check success
basic block (where the return inst is placed) and then move it back
later at SelectionDAG/MI time before the stack protector check if the
tail call optimization failed. The MI level option was nixed
immediately since it would require platform specific pattern
matching. The SelectionDAG level option was nixed because
SelectionDAG only processes one IR level basic block at a time
implying one could not create a DAG Combine to move the callinst.
To get around this problem a few things were realized:
1. While one can not handle multiple IR level basic blocks at the
SelectionDAG Level, one can generate multiple machine basic blocks
for one IR level basic block. This is how we handle bit tests and
switches.
2. At the MI level, tail calls are represented via a special return
MIInst called "tcreturn". Thus if we know the basic block in which we
wish to insert the stack protector check, we get the correct behavior
by always inserting the stack protector check right before the return
statement. This is a "magical transformation" since no matter where
the stack protector check intrinsic is, we always insert the stack
protector check code at the end of the BB.
Given the aforementioned constraints, the following solution was devised:
1. On platforms that do not support SelectionDAG stack protector check
generation, allow for the normal IR level stack protector check
generation to continue.
2. On platforms that do support SelectionDAG stack protector check
generation:
a. Use the IR level stack protector pass to decide if a stack
protector is required/which BB we insert the stack protector check
in by reusing the logic already therein. If we wish to generate a
stack protector check in a basic block, we place a special IR
intrinsic called llvm.stackprotectorcheck right before the BB's
returninst or if there is a callinst that could potentially be
sibling call optimized, before the call inst.
b. Then when a BB with said intrinsic is processed, we codegen the BB
normally via SelectBasicBlock. In said process, when we visit the
stack protector check, we do not actually emit anything into the
BB. Instead, we just initialize the stack protector descriptor
class (which involves stashing information/creating the success
mbbb and the failure mbb if we have not created one for this
function yet) and export the guard variable that we are going to
compare.
c. After we finish selecting the basic block, in FinishBasicBlock if
the StackProtectorDescriptor attached to the SelectionDAGBuilder is
initialized, we first find a splice point in the parent basic block
before the terminator and then splice the terminator of said basic
block into the success basic block. Then we code-gen a new tail for
the parent basic block consisting of the two loads, the comparison,
and finally two branches to the success/failure basic blocks. We
conclude by code-gening the failure basic block if we have not
code-gened it already (all stack protector checks we generate in
the same function, use the same failure basic block).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188755
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 20 Aug 2013 05:23:59 +0000 (05:23 +0000)]
Fix formatting. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188746
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 20 Aug 2013 05:22:42 +0000 (05:22 +0000)]
Add AVX-512 and related features to the CPUID detection code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188745
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 20 Aug 2013 04:24:14 +0000 (04:24 +0000)]
Move AVX and non-AVX replication inside a couple multiclasses to avoid repeating each instruction for both individually.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188743
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 20 Aug 2013 04:22:09 +0000 (04:22 +0000)]
Add an error check for a typo I accidentally made in a td file that caused an assert to fire.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188742
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Tue, 20 Aug 2013 03:12:23 +0000 (03:12 +0000)]
[PowerPC] More refactoring prior to real PPC emitPrologue/Epilogue changes.
(Patch committed on behalf of Mark Minich, whose log entry follows.)
This is a continuation of the refactorings performed in svn rev 188573
(see that rev's comments for more detail).
This is my stage 2 refactoring: I combined the emitPrologue() &
emitEpilogue() PPC32 & PPC64 code into a single flow, simplifying a
lot of the code since in essence the PPC32 & PPC64 code generation
logic is the same, only the instruction forms are different (in most
cases). This simplification is necessary because my functional changes
(yet to come) add significant complexity, and without the
simplification of my stage 2 refactoring, the overall complexity of
both emitPrologue() & emitEpilogue() would have become almost
intractable for most mortal programmers (like me).
This submission was intended to be a pure refactoring (no functional
changes whatsoever). However, in the process of combining the PPC32 &
PPC64 flows, I spotted a difference that I believe is a bug (see svn
rev 186478 line 863, or svn rev 188573 line 888): This line appears to
be restoring the BP with the original FP content, not the original BP
content. When I merged the 32-bit and 64-bit code, I used the
corresponding code from the 64-bit flow, which I believe uses the
correct offset (BPOffset) for this operation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188741
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 01:50:50 +0000 (01:50 +0000)]
Marking MCJIT PIC tests as XFAIL on AArch64
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188740
91177308-0d34-0410-b5e6-
96231b3b80d8
Venkatraman Govindaraju [Tue, 20 Aug 2013 01:26:14 +0000 (01:26 +0000)]
[Sparc] Use HWEncoding instead of unused Num field in Sparc register definitions. Also, correct the definitions of RETL and RET instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188738
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 00:37:33 +0000 (00:37 +0000)]
Fixing XPASSes among MCJIT PIC test on i686
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188736
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 00:22:03 +0000 (00:22 +0000)]
Second attempt to mark Large/PIC MCJIT test as XFAIL for PowerPC64
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188735
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Tue, 20 Aug 2013 00:14:50 +0000 (00:14 +0000)]
Marking two MCJIT PIC tests as XFAIL on Darwin
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188734
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Mon, 19 Aug 2013 23:52:53 +0000 (23:52 +0000)]
Trying again with PIC tests for MCJIT
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188730
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 19 Aug 2013 23:35:46 +0000 (23:35 +0000)]
Add a llvm.copysign intrinsic
This adds a llvm.copysign intrinsic; We already have Libfunc recognition for
copysign (which is turned into the FCOPYSIGN SDAG node). In order to
autovectorize calls to copysign in the loop vectorizer, we need a corresponding
intrinsic as well.
In addition to the expected changes to the language reference, the loop
vectorizer, BasicTTI, and the SDAG builder (the intrinsic is transformed into
an FCOPYSIGN node, just like the function call), this also adds FCOPYSIGN to a
few lists in LegalizeVector{Ops,Types} so that vector copysigns can be
expanded.
In TargetLoweringBase::initActions, I've made the default action for FCOPYSIGN
be Expand for vector types. This seems correct for all in-tree targets, and I
think is the right thing to do because, previously, there was no way to generate
vector-values FCOPYSIGN nodes (and most targets don't specify an action for
vector-typed FCOPYSIGN).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188728
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 19 Aug 2013 23:35:24 +0000 (23:35 +0000)]
Don't form PPC CTR-based loops around a copysignl call
copysign/copysignf never become function calls (because the SDAG expansion code
does not lower to the corresponding function call, but rather directly
implements the associated logic), but copysignl almost always is lowered into a
call to the requested libm functon (and, thus, might clobber CTR).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188727
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Mon, 19 Aug 2013 23:27:43 +0000 (23:27 +0000)]
Adding PIC support for ELF on x86_64 platforms
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188726
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Mon, 19 Aug 2013 23:13:33 +0000 (23:13 +0000)]
Introduce non-const overloads for GlobalAlias::{get,resolve}AliasedGlobal.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188725
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Mon, 19 Aug 2013 22:47:55 +0000 (22:47 +0000)]
Use pop_back_val() instead of both back() and pop_back().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188723
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Mon, 19 Aug 2013 22:17:40 +0000 (22:17 +0000)]
Teach InstCombine visitGetElementPtr about address spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188721
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Mon, 19 Aug 2013 22:17:34 +0000 (22:17 +0000)]
Cleanup visitGetElementPtr to make address space change easier
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188720
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Mon, 19 Aug 2013 22:17:18 +0000 (22:17 +0000)]
commonPointerCast cleanups to make address space change easier
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188719
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Mon, 19 Aug 2013 22:12:00 +0000 (22:12 +0000)]
Make sure that pop_back_val() result is used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188717
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Mon, 19 Aug 2013 22:05:07 +0000 (22:05 +0000)]
Reverting r188709 until I can figure out the proper way to XFAIL it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188715
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Mon, 19 Aug 2013 21:43:16 +0000 (21:43 +0000)]
Fix assert with GEP ptr vector indexing structs
Also fix it calculating the wrong value. The struct index
is not a ConstantInt, so it was being interpreted as an array
index.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188713
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 19 Aug 2013 21:41:38 +0000 (21:41 +0000)]
Use less verbose code and update comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188711
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Mon, 19 Aug 2013 21:40:31 +0000 (21:40 +0000)]
Revert non-test parts of r188507
Re-add the inboundsless tests I didn't add originally
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188710
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Mon, 19 Aug 2013 21:08:35 +0000 (21:08 +0000)]
Adding tests for PIC with MCJIT
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188709
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Mon, 19 Aug 2013 21:07:38 +0000 (21:07 +0000)]
Turn on pubnames by default on linux.
Until gdb supports the new accelerator tables we should add the
pubnames section so that gdb_index can be generated from gold
at link time. On darwin we already emit the accelerator tables
and so don't need to worry about pubnames.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188708
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Mon, 19 Aug 2013 20:25:26 +0000 (20:25 +0000)]
Suppress an annoying CMake warning in ChooseMSVCCRT.cmake
Warning was:
Argument not separated from preceding token by whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188701
91177308-0d34-0410-b5e6-
96231b3b80d8
Paul Redmond [Mon, 19 Aug 2013 20:01:35 +0000 (20:01 +0000)]
Improve the widening of integral binary vector operations
- split WidenVecRes_Binary into WidenVecRes_Binary and WidenVecRes_BinaryCanTrap
- WidenVecRes_BinaryCanTrap preserves the original behaviour for operations
that can trap
- WidenVecRes_Binary simply widens the operation and improves codegen for
3-element vectors by allowing widening and promotion on x86 (matches the
behaviour of unary and ternary operation widening)
- use WidenVecRes_Binary for operations on integers.
Reviewed by: nrotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188699
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Kaylor [Mon, 19 Aug 2013 19:38:06 +0000 (19:38 +0000)]
Adding comments to document RuntimeDyld relocation handling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188697
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Mon, 19 Aug 2013 19:08:03 +0000 (19:08 +0000)]
[mips] Fix instruction definitions that were incorrectly marked as code-gen-only.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188690
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Mon, 19 Aug 2013 19:02:33 +0000 (19:02 +0000)]
Add definition of __warn_unused_result__ attribute. It will be used in the
futher commits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188689
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Mon, 19 Aug 2013 19:00:35 +0000 (19:00 +0000)]
Introduce SpecialCaseList::isIn overload for GlobalAliases.
Differential Revision: http://llvm-reviews.chandlerc.com/D1437
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188688
91177308-0d34-0410-b5e6-
96231b3b80d8
Mihai Popa [Mon, 19 Aug 2013 15:02:25 +0000 (15:02 +0000)]
Thumb2 add immediate alias for SP
The Thumb2 add immediate is in fact defined for SP. The manual is misleading as it points to a different section for add immediate with SP, however the encoding is the same as for add immediate with register only with the SP operand hard coded. As such add immediate with SP and add immediate with register can safely be treated as the same instruction.
All the patch does is adjust a register constraint on an instruction alias.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188676
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Mon, 19 Aug 2013 13:26:14 +0000 (13:26 +0000)]
AVX-512: added arithmetic and logical operations.
ADD, SUB, MUL integer and FP types. OR, AND, XOR.
Added embeded broadcast form for these instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188673
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Mon, 19 Aug 2013 12:56:58 +0000 (12:56 +0000)]
[SystemZ] Add negative integer absolute (load negative)
For now this matches the equivalent of (neg (abs ...)), which did hit a few
times in projects/test-suite. We should probably also match cases where
absolute-like selects are used with reversed arguments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188671
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Mon, 19 Aug 2013 12:48:54 +0000 (12:48 +0000)]
[SystemZ] Add integer absolute (load positive)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188670
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Mon, 19 Aug 2013 12:42:31 +0000 (12:42 +0000)]
[SystemZ] Add support for sibling calls
This first cut is pretty conservative. The final argument register (R6)
is call-saved, so we would need to make sure that the R6 argument to a
sibling call is the same as the R6 argument to the calling function,
which seems worth keeping as a separate patch.
Saying that integer truncations are free means that we no longer
use the extending instructions LGF and LLGF for spills in int-conv-09.ll
and int-conv-10.ll. Instead we treat the registers as 64 bits wide and
truncate them to 32-bits where necessary. I think it's unlikely we'd
use LGF and LLGF for spills in other situations for the same reason,
so I'm removing the tests rather than replacing them. The associated
code is generic and applies to many more instructions than just
LGF and LLGF, so there is no corresponding code removal.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188669
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Kuperstein [Mon, 19 Aug 2013 06:55:47 +0000 (06:55 +0000)]
Adds missing TLI check for library simplification of
* pow(x, 0.5) -> fabs(sqrt(x))
* pow(2.0, x) -> exp2(x)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188656
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 19 Aug 2013 06:55:37 +0000 (06:55 +0000)]
Add ExpandFloatOp_FCOPYSIGN to handle ppcf128-related expansions
We had previously been asserting when faced with a FCOPYSIGN f64, ppcf128 node
because there was no way to expand the FCOPYSIGN node. Because ppcf128 is the
sum of two doubles, and the first double must have the larger magnitude, we
can take the sign from the first double. As a result, in addition to fixing the
crash, this is also an optimization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188655
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Mon, 19 Aug 2013 06:55:01 +0000 (06:55 +0000)]
AVX-512: compiler intrinsics
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188654
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 19 Aug 2013 05:01:02 +0000 (05:01 +0000)]
Add the PPC fcpsgn instruction
Modern PPC cores support a floating-point copysign instruction, and we can use
this to lower the FCOPYSIGN node (which is created from calls to the libm
copysign function). A couple of extra patterns are necessary because the
operand types of FCOPYSIGN need not agree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188653
91177308-0d34-0410-b5e6-
96231b3b80d8