Duncan P. N. Exon Smith [Wed, 3 Dec 2014 05:41:20 +0000 (05:41 +0000)]
LoopVectorize: Remove unnecessary RAUW
Remove an unnecessary `MDNode::replaceAllUsesWith()`. In the preceding
line, `TheLoop->setLoopID()` visits all backedges and sets the new loop
ID. This sufficiently updates the loop metadata.
Metadata RAUW is going away as part of PR21532.
llvm-svn: 223210
Kaelyn Takata [Wed, 3 Dec 2014 05:30:54 +0000 (05:30 +0000)]
Handle delayed corrections in a couple more error paths in ParsePostfixExpressionSuffix.
llvm-svn: 223209
Matt Arsenault [Wed, 3 Dec 2014 05:22:39 +0000 (05:22 +0000)]
R600/SI: Fix SIFixSGPRCopies for copies to physical registers
This shows up when operands required to be passed in VCC are copied
to.
llvm-svn: 223208
Matt Arsenault [Wed, 3 Dec 2014 05:22:38 +0000 (05:22 +0000)]
R600/SI: Remove incorrect assertion
This can be a COPY to a physical register, such as VCC
llvm-svn: 223207
Matt Arsenault [Wed, 3 Dec 2014 05:22:35 +0000 (05:22 +0000)]
R600/SI: Remove i1 pseudo VALU ops
Select i1 logical ops directly to 64-bit SALU instructions.
Vector i1 values are always really in SGPRs, with each
bit for each item in the wave. This saves about 4 instructions
when and/or/xoring any condition, and also helps write conditions
that need to be passed in vcc.
This should work correctly now that the SGPR live range
fixing pass works. More work is needed to eliminate the VReg_1
pseudo regclass and possibly the entire SILowerI1Copies pass.
llvm-svn: 223206
Matt Arsenault [Wed, 3 Dec 2014 05:22:32 +0000 (05:22 +0000)]
R600/SI: Fix suspicious indexing
The loop is over the operands of an instruction, and checks the
register with the sub reg index of the dest register. This probably
meant to be checking the sub reg index of the same operand.
llvm-svn: 223205
Matt Arsenault [Wed, 3 Dec 2014 05:22:30 +0000 (05:22 +0000)]
R600/SI: Fix running SILowerI1Copies a second time
llvm-svn: 223204
Matt Arsenault [Wed, 3 Dec 2014 05:22:29 +0000 (05:22 +0000)]
R600/SI: Fix live range error hidden by SIFoldOperands
m0 is treated as a virtual register class with a single register
rather than the physical register it really is. This was updating
the live range of the used virtual copy of m0 from the first ds_read
instruction, and leaving the unused copy unchanged. This resulted in a
"Live segment doesn't end at a valid instruction" verifier error because
the erased instructions. Update the live range of the second copy (which
should be dead).
No test since I'm not sure how to trigger this with SIFoldOperands
enabled.
llvm-svn: 223203
Duncan P. N. Exon Smith [Wed, 3 Dec 2014 04:49:16 +0000 (04:49 +0000)]
ADT: Add SmallVector<>::emplace_back(): fixup
Add missing `void` return type from `!LLVM_HAS_VARIADIC_TEMPLATES` case
in r223201.
llvm-svn: 223202
Duncan P. N. Exon Smith [Wed, 3 Dec 2014 04:45:09 +0000 (04:45 +0000)]
ADT: Add SmallVector<>::emplace_back()
llvm-svn: 223201
Rui Ueyama [Wed, 3 Dec 2014 04:34:20 +0000 (04:34 +0000)]
[PECOFF] Fix a bug in /export option handler.
/export option can be given multiple times to specify multiple
symbols to be exported. /export accepts both decorated and
undecorated name.
If you give both undecorated and decorated name of the same symbol
to /export, they are resolved to the same symbol. In this case,
we need to de-duplicate the exported names, so that we don't have
duplicated items in the export symbol table in a DLL.
We remove duplicate items from a vector. The bug was there.
Because we had pointers pointing to elements of the vector,
after an item is removed, they would point wrong elements.
This patch is to remove these pointers. Added a test for that case.
llvm-svn: 223200
Tom Stellard [Wed, 3 Dec 2014 04:28:32 +0000 (04:28 +0000)]
StructurizeCFG: Use LoopInfo analysis for better loop detection
We were assuming that each back-edge in a region represented a unique
loop, which is not always the case. We need to use LoopInfo to
correctly determine which back-edges are loops.
llvm-svn: 223199
Duncan P. N. Exon Smith [Wed, 3 Dec 2014 04:13:23 +0000 (04:13 +0000)]
NVPTX: Delete dead code
`MDNode` does not inherit from `User`, and it never has a name.
llvm-svn: 223198
Tom Stellard [Wed, 3 Dec 2014 04:08:00 +0000 (04:08 +0000)]
R600/SI: Enable inline assembly
We just needed to remove the assertion in
AMDGPURegisterInfo::getFrameRegister(), which is called when
initializing the parser for inline assembly.
llvm-svn: 223197
Jason Molenda [Wed, 3 Dec 2014 04:02:03 +0000 (04:02 +0000)]
Update setMCJITMemoryManager call to keep in line with llvm r223183.
Patch from Ryan Goodfellow.
llvm-svn: 223196
Peter Zotov [Wed, 3 Dec 2014 03:39:01 +0000 (03:39 +0000)]
[OCaml] [cmake] Disable OCaml bindings if ctypes >=0.3 is not found.
llvm-svn: 223195
Matt Arsenault [Wed, 3 Dec 2014 03:12:13 +0000 (03:12 +0000)]
R600/SI: Change mubuf offsets to print as decimal
This matches SC's behavior.
llvm-svn: 223194
Nick Lewycky [Wed, 3 Dec 2014 02:45:01 +0000 (02:45 +0000)]
Emit the entry block first and the exit block second, then all the blocks in between afterwards. This is what gcc always does, and some out of tree tools depend on that.
llvm-svn: 223193
NAKAMURA Takumi [Wed, 3 Dec 2014 02:40:24 +0000 (02:40 +0000)]
GCRelocateOperands: Try to appease msc17.
llvm-svn: 223192
Peter Collingbourne [Wed, 3 Dec 2014 02:37:10 +0000 (02:37 +0000)]
Update test to check for prologue instead of prefix
llvm-svn: 223191
Peter Collingbourne [Wed, 3 Dec 2014 02:08:51 +0000 (02:08 +0000)]
UBSan now uses prologue data instead of prefix data
As the semantics of prefix data has changed. See D6454.
Patch by Ben Gamari!
Test Plan: Testsuite
Differential Revision: http://reviews.llvm.org/D6489
llvm-svn: 223190
Peter Collingbourne [Wed, 3 Dec 2014 02:08:38 +0000 (02:08 +0000)]
Prologue support
Patch by Ben Gamari!
This redefines the `prefix` attribute introduced previously and
introduces a `prologue` attribute. There are a two primary usecases
that these attributes aim to serve,
1. Function prologue sigils
2. Function hot-patching: Enable the user to insert `nop` operations
at the beginning of the function which can later be safely replaced
with a call to some instrumentation facility
3. Runtime metadata: Allow a compiler to insert data for use by the
runtime during execution. GHC is one example of a compiler that
needs this functionality for its tables-next-to-code functionality.
Previously `prefix` served cases (1) and (2) quite well by allowing the user
to introduce arbitrary data at the entrypoint but before the function
body. Case (3), however, was poorly handled by this approach as it
required that prefix data was valid executable code.
Here we redefine the notion of prefix data to instead be data which
occurs immediately before the function entrypoint (i.e. the symbol
address). Since prefix data now occurs before the function entrypoint,
there is no need for the data to be valid code.
The previous notion of prefix data now goes under the name "prologue
data" to emphasize its duality with the function epilogue.
The intention here is to handle cases (1) and (2) with prologue data and
case (3) with prefix data.
References
----------
This idea arose out of discussions[1] with Reid Kleckner in response to a
proposal to introduce the notion of symbol offsets to enable handling of
case (3).
[1] http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-May/073235.html
Test Plan: testsuite
Differential Revision: http://reviews.llvm.org/D6454
llvm-svn: 223189
NAKAMURA Takumi [Wed, 3 Dec 2014 02:05:51 +0000 (02:05 +0000)]
ExceptionDemo: Let setMCJITMemoryManager() take unique_ptr, since r223183.
llvm-svn: 223188
Ahmed Bougacha [Wed, 3 Dec 2014 02:03:26 +0000 (02:03 +0000)]
[X86][MC] Intel syntax: accept implicit memory operand sizes larger than 80.
The X86AsmParser intel handling was refactored in r216481, making it
try each different memory operand size to see which one matches.
Operand sizes larger than 80 ("[xyz]mmword ptr") were forgotten, which
led to an "invalid operand" error for code such as:
movdqa [rax], xmm0
llvm-svn: 223187
Nico Weber [Wed, 3 Dec 2014 01:25:49 +0000 (01:25 +0000)]
Add support for has_feature(cxx_alignof) and has_feature(c_alignof).
r142020 added support for has_feature(cxx_alignas). This does the same for
alignof.
llvm-svn: 223186
Nico Weber [Wed, 3 Dec 2014 01:21:41 +0000 (01:21 +0000)]
Fix incorrect codegen for devirtualized calls to virtual overloaded operators.
Consider this program:
struct A {
virtual void operator-() { printf("base\n"); }
};
struct B final : public A {
virtual void operator-() override { printf("derived\n"); }
};
int main() {
B* b = new B;
-static_cast<A&>(*b);
}
Before this patch, clang saw the virtual call to A::operator-(), figured out
that it can be devirtualized, and then just called A::operator-() directly,
without going through the vtable. Instead, it should've looked up which
operator-() the call devirtualizes to and should've called that.
For regular virtual member calls, clang gets all this right already. So
instead of giving EmitCXXOperatorMemberCallee() all the logic that
EmitCXXMemberCallExpr() already has, cut the latter function into two pieces,
call the second piece EmitCXXMemberOrOperatorMemberCallExpr(), and use it also
to generate code for calls to virtual member operators.
This way, virtual overloaded operators automatically don't get devirtualized
if they have covariant returns (like it was done for regular calls in r218602),
etc.
This also happens to fix (or at least improve) codegen for explicit constructor
calls (`A a; a.A::A()`) in MS mode with -fsanitize-address-field-padding=1.
(This adjustment for virtual operator calls seems still wrong with the MS ABI.)
llvm-svn: 223185
Richard Smith [Wed, 3 Dec 2014 01:05:50 +0000 (01:05 +0000)]
PR21706: -Wunsequenced was missing warnings when leaving a sequenced region that contained side effects.
llvm-svn: 223184
Lang Hames [Wed, 3 Dec 2014 00:51:19 +0000 (00:51 +0000)]
[MCJIT] Unique-ptrify the RTDyldMemoryManager member of MCJIT. NFC.
llvm-svn: 223183
Hal Finkel [Wed, 3 Dec 2014 00:19:17 +0000 (00:19 +0000)]
[PowerPC] Fix readcyclecounter to be custom expanded for all 32-bit targets
We need to use the custom expansion of readcyclecounter on all 32-bit targets
(even those with 64-bit registers). This should fix the ppc64 buildbot.
llvm-svn: 223182
Kostya Serebryany [Wed, 3 Dec 2014 00:08:41 +0000 (00:08 +0000)]
[asan] fix four asan tests to run in use-after-return mode
llvm-svn: 223181
Tim Northover [Tue, 2 Dec 2014 23:53:43 +0000 (23:53 +0000)]
AArch64: strengthen Darwin ABI alignment assumptions
A global variable without an explicit alignment specified should be assumed to
be ABI-aligned according to its type, like on other platforms. This allows us
to use better memory operations when accessing it.
rdar://
18533701
llvm-svn: 223180
David Majnemer [Tue, 2 Dec 2014 23:44:40 +0000 (23:44 +0000)]
FullProduct should be _FullProduct
llvm-svn: 223179
Pete Cooper [Tue, 2 Dec 2014 23:34:23 +0000 (23:34 +0000)]
Use a typed enum instead of 'unsigned char' for packed field. NFC.
This makes it easier to debug Twine as the 'Kind' fields now show their enum values in lldb and not escaped characters.
llvm-svn: 223178
Kaelyn Takata [Tue, 2 Dec 2014 23:32:20 +0000 (23:32 +0000)]
Ensure typos in the default values of template parameters get diagnosed.
llvm-svn: 223177
David Majnemer [Tue, 2 Dec 2014 23:30:26 +0000 (23:30 +0000)]
Intrin: shrx_u64 should be _shrx_u64
llvm-svn: 223176
David Majnemer [Tue, 2 Dec 2014 23:30:24 +0000 (23:30 +0000)]
Intrin: Add _umul128
Implement _umul128; it provides the high and low halves of a 128-bit
multiply. We can simply use our __int128 arithmetic to implement this,
we generate great code for it:
movq %rdx, %rax
mulq %rcx
movq %rdx, (%r8)
retq
Differential Revision: http://reviews.llvm.org/D6486
llvm-svn: 223175
Jason Molenda [Tue, 2 Dec 2014 23:21:05 +0000 (23:21 +0000)]
Mark the armv7 q0-q3 and q8-q15 registers as volatile (not callee
preserved) in the ABI.
Realistically lldb isn't able to track register saves of any of
the neon regs right now so we should probably mark all of the
regs as unavailable when you're not on stack frame 0...
<rdar://problem/
19115127>
llvm-svn: 223174
Justin Bogner [Tue, 2 Dec 2014 23:15:30 +0000 (23:15 +0000)]
InstrProf: Use the same names for variables as we use in the profile
There's no need to use different names for the local variables than we
use in the profile itself, and it's a bit simpler and easier to debug
if we're consistent.
llvm-svn: 223173
Tim Northover [Tue, 2 Dec 2014 23:13:39 +0000 (23:13 +0000)]
AArch64: don't be too greedy when folding :lo12: accesses into mem ops.
This frequently leads to cases like:
ldr xD, [xN, :lo12:var]
add xA, xN, :lo12:var
ldr xD, [xA, #8]
where the ADD would have been needed anyway, and the two distinct addressing
modes can prevent the formation of an ldp. Because of how we handle ADRP
(aggressively forming an ADRP/ADD pseudo-inst at ISel time), this pattern also
results in duplicated ADRP instructions (one on its own to cover the ldr, and
one combined with the add).
llvm-svn: 223172
Michael Zolotukhin [Tue, 2 Dec 2014 22:59:06 +0000 (22:59 +0000)]
PR21302. Vectorize only bottom-tested loops.
rdar://problem/
18886083
llvm-svn: 223171
Michael Zolotukhin [Tue, 2 Dec 2014 22:59:02 +0000 (22:59 +0000)]
Apply loop-rotate to several vectorizer tests.
Such loops shouldn't be vectorized due to the loops form.
After applying loop-rotate (+simplifycfg) the tests again start to check
what they are intended to check.
llvm-svn: 223170
Fariborz Jahanian [Tue, 2 Dec 2014 22:42:52 +0000 (22:42 +0000)]
Another warning with no group name bites the dust.
rdar://
19116886
llvm-svn: 223168
Justin Bogner [Tue, 2 Dec 2014 22:38:52 +0000 (22:38 +0000)]
InstrProf: Remove some pointless indirection (NFC)
It doesn't make much sense to have std::unique_ptrs of std::string and
std::vector. Avoid some useless indirection by using these types
directly.
llvm-svn: 223166
Simon Pilgrim [Tue, 2 Dec 2014 22:31:23 +0000 (22:31 +0000)]
[X86][SSE] Keep 4i32 vector insertions in integer domain on SSE4.1 targets
4i32 shuffles for single insertions into zero vectors lowers to X86vzmovl which was using (v)blendps - causing domain switch stalls. This patch fixes this by using (v)pblendw instead.
The updated tests on test/CodeGen/X86/sse41.ll still contain a domain stall due to the use of insertps - I'm looking at fixing this in a future patch.
Differential Revision: http://reviews.llvm.org/D6458
llvm-svn: 223165
Alexey Samsonov [Tue, 2 Dec 2014 22:20:11 +0000 (22:20 +0000)]
Replace InternalScopedBuffer<char> with InternalScopedString where applicable.
Summary: No functionality change.
Test Plan: make check-all
Reviewers: kcc
Reviewed By: kcc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D6472
llvm-svn: 223164
Chris Matthews [Tue, 2 Dec 2014 22:19:21 +0000 (22:19 +0000)]
Give lit a --xunit-xml-output option for saving results in xunit format
--xunit-xml-output saves test results to disk in JUnit's xml format. This will allow Jenkins to report the details of a lit run.
Based on a patch by David Chisnall.
llvm-svn: 223163
Kaelyn Takata [Tue, 2 Dec 2014 22:05:35 +0000 (22:05 +0000)]
Diagnose TypoExprs in a couple of error cases in ParsePostfixExpressionSuffix.
Also have CorrectDelayedTyposInExpr check that the Expr* isn't null
before trying to access its members. Fixes PR21679.
llvm-svn: 223162
Hal Finkel [Tue, 2 Dec 2014 22:01:00 +0000 (22:01 +0000)]
[PowerPC] Implement readcyclecounter for PPC32
We've long supported readcyclecounter on PPC64, but it is easier there (the
read of the 64-bit time-base register can be accomplished via a single
instruction). This now provides an implementation for PPC32 as well. On PPC32,
the time-base register is still 64 bits, but can only be read 32 bits at a time
via two separate SPRs. The ISA manual explains how to do this properly (it
involves re-reading the upper bits and looping if the counter has wrapped while
being read).
This requires PPC to implement a custom integer splitting legalization for the
READCYCLECOUNTER node, turning it into a target-specific SDAG node, which then
gets turned into a pseudo-instruction, which is then expanded to the necessary
sequence (which has three SPR reads, the comparison and the branch).
Thanks to Paul Hargrove for pointing out to me that this was still unimplemented.
llvm-svn: 223161
Tom Stellard [Tue, 2 Dec 2014 22:00:07 +0000 (22:00 +0000)]
R600/SI: Emit amd_kernel_code_t header for AMDGPU environment
llvm-svn: 223160
Eric Christopher [Tue, 2 Dec 2014 21:57:15 +0000 (21:57 +0000)]
Make sure that the TargetOptions operator== is checking the
full contents of the class.
llvm-svn: 223159
Alexey Samsonov [Tue, 2 Dec 2014 21:40:41 +0000 (21:40 +0000)]
Add missing #include to fix Android build.
llvm-svn: 223157
Lang Hames [Tue, 2 Dec 2014 21:36:24 +0000 (21:36 +0000)]
[AArch64][Stackmaps] Optimize stackmap shadows on AArch64.
Reduce the number of nops emitted for stackmap shadows on AArch64 by counting
non-stackmap instructions up to the next branch target towards the requested
shadow.
<rdar://problem/
14959522>
llvm-svn: 223156
Zachary Turner [Tue, 2 Dec 2014 21:32:44 +0000 (21:32 +0000)]
Fix broken test suite on Windows after r223091.
Differential Revision: http://reviews.llvm.org/D6484
Reviewed by: Oleksiy Vyalov
llvm-svn: 223155
Tom Stellard [Tue, 2 Dec 2014 21:28:53 +0000 (21:28 +0000)]
R600/SI: Move more information into SIProgramInfo struct
llvm-svn: 223154
Eric Christopher [Tue, 2 Dec 2014 21:09:01 +0000 (21:09 +0000)]
Add bindings for the rest of the MCJIT options that we previously
had support for. We're still missing a binding for an MCJIT
memory manager.
llvm-svn: 223153
Tobias Grosser [Tue, 2 Dec 2014 21:04:20 +0000 (21:04 +0000)]
checkout_isl: Do not fail in presence of an old CLooG checkout
This should help our buildbots and may also simplify life for other people.
llvm-svn: 223152
Matt Arsenault [Tue, 2 Dec 2014 21:02:20 +0000 (21:02 +0000)]
R600: Cleanup some tests and add missing testcases
llvm-svn: 223151
Philip Reames [Tue, 2 Dec 2014 21:01:48 +0000 (21:01 +0000)]
Restructure some assertion checking based on post commit feedback by Aaron and Tom.
llvm-svn: 223150
Nico Weber [Tue, 2 Dec 2014 20:41:18 +0000 (20:41 +0000)]
Wrap to 80 columns. No behavior change.
llvm-svn: 223149
Daniel Sanders [Tue, 2 Dec 2014 20:40:27 +0000 (20:40 +0000)]
[mips] Fix passing of small structures for big-endian O32.
Summary:
Like N32/N64, they must be passed in the upper bits of the register.
The new code could be merged with the existing if-statements but I've
refrained from doing this since it will make porting the O32 implementation
to tablegen harder later.
Reviewers: vmedic
Reviewed By: vmedic
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D6463
llvm-svn: 223148
Roman Divacky [Tue, 2 Dec 2014 20:03:22 +0000 (20:03 +0000)]
Introduce CPUStringIsValid() into MCSubtargetInfo and use it for ARM .cpu parsing.
Previously .cpu directive in ARM assembler didnt switch to the new CPU and
therefore acted as a nop. This implemented real action for .cpu and eg.
allows to assembler FreeBSD kernel with -integrated-as.
llvm-svn: 223147
Tobias Grosser [Tue, 2 Dec 2014 19:49:40 +0000 (19:49 +0000)]
Make checkout isl script executable
llvm-svn: 223146
Alexey Samsonov [Tue, 2 Dec 2014 19:48:40 +0000 (19:48 +0000)]
Simplify Symbolizer::SymbolizePC() interface.
Return a linked list of AddressInfo objects, instead of using an array of
these objects as an output parameter. This simplifies the code in callers
of this function (especially TSan).
Fix a few memory leaks from internal allocator, when the returned
AddressInfo objects were not properly cleared.
llvm-svn: 223145
Tom Stellard [Tue, 2 Dec 2014 19:45:05 +0000 (19:45 +0000)]
R600/SI: Refactor AMDGPUAsmPrinter::EmitProgramInfoSI()
llvm-svn: 223144
Philip Reames [Tue, 2 Dec 2014 19:37:00 +0000 (19:37 +0000)]
[Statepoints 4/4] Statepoint infrastructure for garbage collection: Documentation
This is the fourth and final patch in the statepoint series. It contains the documentation for the statepoint intrinsics and their usage.
There's definitely still room to improve the documentation here, but I wanted to get this landed so it was available for others. There will likely be a series of small cleanup changes over the next few weeks as we work to clarify and revise the documentation. If you have comments or questions, please feel free to discuss them either in this commit thread, the original review thread, or on llvmdev. Comments are more than welcome.
Reviewed by: atrick, ributzka
Differential Revision: http://reviews.llvm.org/D5683
llvm-svn: 223143
Philip Reames [Tue, 2 Dec 2014 19:28:57 +0000 (19:28 +0000)]
Appease a build bot complaining about an unused variable that's used in an assertion.
llvm-svn: 223142
Johannes Doerfert [Tue, 2 Dec 2014 19:26:58 +0000 (19:26 +0000)]
Drop Cloog support
This commit drops the Cloog support for Polly. The scripts and
documentation are changed to only use isl as prerequisity. In the code
all Cloog specific parts have been removed and all relevant tests have
been ported to the isl backend when it was created.
llvm-svn: 223141
JF Bastien [Tue, 2 Dec 2014 19:19:59 +0000 (19:19 +0000)]
Make le64 DescriptionString consistent with other targets.
Summary:
In particular, remove the defaults and reorder fields so it matches the result of DataLayout::getStringDescription().
Change by David Neto.
Reviewers: dschuff, sdt
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D6482
llvm-svn: 223140
Reid Kleckner [Tue, 2 Dec 2014 18:59:08 +0000 (18:59 +0000)]
cmake: Remove MAXPATHLEN define as autoconf does not provide it
Presumably it was added to the CMake system when MAXPATHLEN was still
used by code built for Windows. Currently only lib/Support/Path.inc uses
MAXPATHLEN, and it should be available on all Unices.
llvm-svn: 223139
Reid Kleckner [Tue, 2 Dec 2014 18:58:38 +0000 (18:58 +0000)]
Remove '#undef const' from config.h.cmake to sync with autoconf
This define was removed from config.h.in when Rafael removed our use of
libtool.
llvm-svn: 223138
Philip Reames [Tue, 2 Dec 2014 18:50:36 +0000 (18:50 +0000)]
[Statepoints 3/4] Statepoint infrastructure for garbage collection: SelectionDAGBuilder
This is the third patch in a small series. It contains the CodeGen support for lowering the gc.statepoint intrinsic sequences (223078) to the STATEPOINT pseudo machine instruction (223085). The change also includes the set of helper routines and classes for working with gc.statepoints, gc.relocates, and gc.results since the lowering code uses them.
With this change, gc.statepoints should be functionally complete. The documentation will follow in the fourth change, and there will likely be some cleanup changes, but interested parties can start experimenting now.
I'm not particularly happy with the amount of code or complexity involved with the lowering step, but at least it's fairly well isolated. The statepoint lowering code is split into it's own files and anyone not working on the statepoint support itself should be able to ignore it.
During the lowering process, we currently spill aggressively to stack. This is not entirely ideal (and we have plans to do better), but it's functional, relatively straight forward, and matches closely the implementations of the patchpoint intrinsics. Most of the complexity comes from trying to keep relocated copies of values in the same stack slots across statepoints. Doing so avoids the insertion of pointless load and store instructions to reshuffle the stack. The current implementation isn't as effective as I'd like, but it is functional and 'good enough' for many common use cases.
In the long term, I'd like to figure out how to integrate the statepoint lowering with the register allocator. In principal, we shouldn't need to eagerly spill at all. The register allocator should do any spilling required and the statepoint should simply record that fact. Depending on how challenging that turns out to be, we may invest in a smarter global stack slot assignment mechanism as a stop gap measure.
Reviewed by: atrick, ributzka
llvm-svn: 223137
Fariborz Jahanian [Tue, 2 Dec 2014 18:42:51 +0000 (18:42 +0000)]
This patch fixes a crash involving use of predefined
expressions. It fixes crash when mangling name for block's helper
function used inside a constructor/destructor.
rdar://
19065361.
llvm-svn: 223136
Bruno Cardoso Lopes [Tue, 2 Dec 2014 18:31:53 +0000 (18:31 +0000)]
[SwitchLowering] Handle destinations on multiple phi instructions
Follow up from r222926. Also handle multiple destinations from merged
cases on multiple and subsequent phi instructions.
rdar://problem/
19106978
llvm-svn: 223135
Samuel Benzaquen [Tue, 2 Dec 2014 18:29:00 +0000 (18:29 +0000)]
Skip some unnecessary type checks.
Summary:
Skip some unnecessary type checks wrt DynTypedNodes.
Add DynTypedNode::getUnchecked() to skip the runtime check when the type
is known.
Speed up DynTypedNode::operator== by using isSame() instead of
isBaseOf().
Skip the type check in MatcherInterface<T>::matches(). All calls come
from DynTypedMatcher::matches(), which already did the type check.
This change speeds up our clang-tidy benchmark by ~4%.
Reviewers: klimek
Subscribers: klimek, cfe-commits
Differential Revision: http://reviews.llvm.org/D6468
llvm-svn: 223134
Ahmed Bougacha [Tue, 2 Dec 2014 18:09:51 +0000 (18:09 +0000)]
[MachineCSE] Clear kill-flag on registers imp-def'd by the CSE'd instruction.
Go through implicit defs of CSMI and MI, and clear the kill flags on
their uses in all the instructions between CSMI and MI.
We might have made some of the kill flags redundant, consider:
subs ... %NZCV<imp-def> <- CSMI
csinc ... %NZCV<imp-use,kill> <- this kill flag isn't valid anymore
subs ... %NZCV<imp-def> <- MI, to be eliminated
csinc ... %NZCV<imp-use,kill>
Since we eliminated MI, and reused a register imp-def'd by CSMI
(here %NZCV), that register, if it was killed before MI, should have
that kill flag removed, because it's lifetime was extended.
Also, add an exhaustive testcase for the motivating example.
Reviewed by: Juergen Ributzka <juergen@apple.com>
llvm-svn: 223133
Philip Reames [Tue, 2 Dec 2014 18:06:10 +0000 (18:06 +0000)]
Remove unneccessary code introduced with 223101.
llvm-svn: 223132
Zachary Turner [Tue, 2 Dec 2014 17:57:54 +0000 (17:57 +0000)]
Disable warning 4530 for MSVC builds.
We compile with exceptions off for LLVM and all other LLVM
subprojects, so this brings parity to LLD and disables this
warning.
Reviewed by: Rui Ueyama
llvm-svn: 223131
Tom Stellard [Tue, 2 Dec 2014 17:41:43 +0000 (17:41 +0000)]
R600/SI: Set correct number of user sgprs for HSA runtime
We don't support scratch buffers yet with HSA.
llvm-svn: 223130
Peter Zotov [Tue, 2 Dec 2014 17:35:26 +0000 (17:35 +0000)]
[OCaml] Add Llvm.mdnull.
Patch by Gideon Smeding <gideon.smeding@3ds.com>.
llvm-svn: 223129
JF Bastien [Tue, 2 Dec 2014 17:30:19 +0000 (17:30 +0000)]
libc++: support NaCl when building thread.cpp
Summary: NaCl shouldn't include sysctl.h when trying to determine std::thread::hardware_concurrency, it should instead use sysconf(_SC_NPROCESSORS_ONLN) through unistd.h. No test needs to be changed, since hardware_concurrency.pass.cpp already tests that std::thread::hardware_concurrency > 0.
Test Plan: make check-libcxx
Reviewers: dschuff, danalbert
Subscribers: jfb, cfe-commits
Differential Revision: http://reviews.llvm.org/D6470
llvm-svn: 223128
Sanjay Patel [Tue, 2 Dec 2014 17:25:27 +0000 (17:25 +0000)]
fix typo in comment
llvm-svn: 223127
Tim Northover [Tue, 2 Dec 2014 17:15:22 +0000 (17:15 +0000)]
AArch64: make register block rules apply to vector types too.
The blocking code originated in ARM, which is more aggressive about casting
types to a canonical representative before doing anything else, so I missed out
most vector HFAs and broke the ABI. This should fix it.
llvm-svn: 223126
Tom Stellard [Tue, 2 Dec 2014 17:05:41 +0000 (17:05 +0000)]
R600/SI: Set the ATC bit on all resource descriptors for the HSA runtime
llvm-svn: 223125
Tom Stellard [Tue, 2 Dec 2014 16:45:47 +0000 (16:45 +0000)]
Triple: Add AMDHSA operating system type
This operating system type represents the AMD HSA runtime,
and will be required by the R600 backend in order to generate
correct code for this runtime.
llvm-svn: 223124
Anton Korobeynikov [Tue, 2 Dec 2014 16:04:58 +0000 (16:04 +0000)]
Fix invalid calling convention used for libcalls on ARM.
ARM ABI specifies that all the libcalls use soft FP ABI
(even hard FP binaries). These days clang emits _mulsc3 / _muldc3
calls with default (C) calling convention which would be translated
into AAPCS_VFP LLVM calling and thus the result of complex
multiplication will be bogus.
Introduce a way for a target to specify explicitly calling
convention for libcalls. Right now this is temporary correctness
fix. Ultimately, we'll end with intrinsic for complex
multiplication and all calling convention decisions for libcalls
will be put into backend.
llvm-svn: 223123
Viktor Kutuzov [Tue, 2 Dec 2014 15:04:39 +0000 (15:04 +0000)]
[Tsan] Fix the atomic_race.cc test to pass on systems with high loads
Differential Revision: http://reviews.llvm.org/D6478
llvm-svn: 223122
Viktor Kutuzov [Tue, 2 Dec 2014 14:59:51 +0000 (14:59 +0000)]
[Tsan] Do not flush all streams on exit
Differential Revision: http://reviews.llvm.org/D6462
llvm-svn: 223121
Serge Pavlov [Tue, 2 Dec 2014 14:52:20 +0000 (14:52 +0000)]
Reverted r223114, it caused failure on on clang-native-arm-cortex-a9.
llvm-svn: 223120
Bruno Cardoso Lopes [Tue, 2 Dec 2014 14:22:34 +0000 (14:22 +0000)]
[LICM] Avoind store sinking if no preheader is available
Load instructions are inserted into loop preheaders when sinking stores
and later removed if not used by the SSA updater. Avoid sinking if the
loop has no preheader and avoid crashes. This fixes one more side effect
of not handling indirectbr instructions properly on LoopSimplify.
llvm-svn: 223119
Daniel Jasper [Tue, 2 Dec 2014 14:21:16 +0000 (14:21 +0000)]
clang-format: Escape '*' in generated flag documentation.
llvm-svn: 223118
Daniel Jasper [Tue, 2 Dec 2014 13:24:51 +0000 (13:24 +0000)]
clang-format: Add option to suppress operator alignment.
With alignment:
int aaaaaa = aa
+
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
*
cccccccccccccccccccccccccccccccc;
Without alignment:
int aaaaaa = aa
+
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
*
cccccccccccccccccccccccccccccccc;
This fixes llvm.org/PR21666.
llvm-svn: 223117
Justin Holewinski [Tue, 2 Dec 2014 12:58:24 +0000 (12:58 +0000)]
[NVPTX] Fix type error for some builtins in BuiltinsNVPTX.def
llvm-svn: 223116
Asiri Rathnayake [Tue, 2 Dec 2014 12:09:55 +0000 (12:09 +0000)]
Remove unused function.
Removing an unused function which is causing one of the build bots to fail.
This was introduced in the commit r223113. A proper cleanup of the so_imm
tblgen defintion (made redundant by the mod_imm definition) needs to happen
soon.
llvm-svn: 223115
Serge Pavlov [Tue, 2 Dec 2014 11:06:09 +0000 (11:06 +0000)]
Emit warning if define or undef reserved identifier or keyword.
Summary:
This change implements warnings if macro name is identical to a keyword or
reserved identifier. The warnings are different depending on the "danger"
of the operation. Defining macro that replaces a keyword is on by default.
Other cases produce warning that is off by default but can be turned on
using option -Wreserved-id-macro.
This change fixes PR11488.
Reviewers: rnk
Reviewed By: rnk
Subscribers: rnk, cfe-commits
Differential Revision: http://reviews.llvm.org/D6194
llvm-svn: 223114
Asiri Rathnayake [Tue, 2 Dec 2014 10:53:20 +0000 (10:53 +0000)]
Add support for ARM modified-immediate assembly syntax.
Certain ARM instructions accept 32-bit immediate operands encoded as a 8-bit
integer value (0-255) and a 4-bit rotation (0-30, even). Current ARM assembly
syntax support in LLVM allows the decoded (32-bit) immediate to be specified
as a single immediate operand for such instructions:
mov r0, #
4278190080
The ARMARM defines an extended assembly syntax allowing the encoding to be made
more explicit, as in:
mov r0, #255, #8 ; (same 32-bit value as above)
The behaviour of the two instructions can be different w.r.t flags, which is
documented under "Modified immediate constants" in ARMARM. This patch enables
support for this extended syntax at the MC layer.
llvm-svn: 223113
Will Newton [Tue, 2 Dec 2014 09:49:09 +0000 (09:49 +0000)]
Add ARM relocations to ELFYAML
Tested with check-all with no regressions.
llvm-svn: 223112
Daniel Jasper [Tue, 2 Dec 2014 09:46:56 +0000 (09:46 +0000)]
clang-format: precedence-based indentation when breaking before operators.
Before:
bool value = aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
+
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
+
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
== aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
*
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
+
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
&& aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
*
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
> ccccccccccccccccccccccccccccccccccccccccc;
After:
bool value = aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
+
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
+
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
== aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
*
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
+
bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
&& aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
*
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
> ccccccccccccccccccccccccccccccccccccccccc;
Not particularly pretty, but can probably help to uncover bugs. And if this
bugs somebody, parentheses can help.
llvm-svn: 223111
Charlie Turner [Tue, 2 Dec 2014 08:22:29 +0000 (08:22 +0000)]
Emit Tag_ABI_FP_denormal correctly in fast-math mode.
The default ARM floating-point mode does not support IEEE 754 mode exactly. Of
relevance to this patch is that input denormals are flushed to zero. The way in
which they're flushed to zero depends on the architecture,
* For VFPv2, it is implementation defined as to whether the sign of zero is
preserved.
* For VFPv3 and above, the sign of zero is always preserved when a denormal
is flushed to zero.
When FP support has been disabled, the strategy taken by this patch is to
assume the software support will mirror the behaviour of the hardware support
for the target *if it existed*. That is, for architectures which can only have
VFPv2, it is assumed the software will flush to positive zero. For later
architectures it is assumed the software will flush to zero preserving sign.
Change-Id: Icc5928633ba222a4ba3ca8c0df44a440445865fd
llvm-svn: 223110
Sonam Kumari [Tue, 2 Dec 2014 05:29:47 +0000 (05:29 +0000)]
[signext.ll] Removal Of Duplicate Test Cases
Removed the duplicate test case existing in signext.ll file.
llvm-svn: 223109
Bob Wilson [Tue, 2 Dec 2014 05:17:21 +0000 (05:17 +0000)]
Re-apply "Revert r166370 and r166540 now that Xcode 4.6 has been available for a while."
This reverts commit r176892.
I had reverted this a while back to give Chromium more time to update, and
Nico says it should be OK now.
llvm-svn: 223108