Chandler Carruth [Mon, 28 Apr 2014 08:52:44 +0000 (08:52 +0000)]
[inliner] Significantly improve the compile time in cases like PR19499
by avoiding inlining massive switches merely because they have no
instructions in them. These switches still show up where we fail to form
lookup tables, and in those cases they are actually going to cause
a very significant code size hit anyways, so inlining them is not the
right call. The right way to fix any performance regressions stemming
from this is to enhance the switch-to-lookup-table logic to fire in more
places.
This makes PR19499 about 5x less bad. It uncovers a second compile time
problem in that test case that is unrelated (surprisingly!).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207403
91177308-0d34-0410-b5e6-
96231b3b80d8
Hao Liu [Mon, 28 Apr 2014 07:34:27 +0000 (07:34 +0000)]
[ARM64]Fix a bug cannot select UQSHL/SQSHL with constant i64 shift amount.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207399
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 28 Apr 2014 05:57:50 +0000 (05:57 +0000)]
Convert more SelectionDAG functions to use ArrayRef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207397
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Mon, 28 Apr 2014 04:05:08 +0000 (04:05 +0000)]
[C++] Use 'nullptr'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207394
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Mon, 28 Apr 2014 03:34:48 +0000 (03:34 +0000)]
MC: range-loopify
Use C++11 range-based loops rather than explicit constructors. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207393
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Mon, 28 Apr 2014 01:57:46 +0000 (01:57 +0000)]
Use raw_ostream and Format.h on Windows so that we don't have to roll
our own portability system to cope without snprintf.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207389
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Mon, 28 Apr 2014 01:24:35 +0000 (01:24 +0000)]
Update the Windows TimeValue formatting to match the new formatting on
Unix-like OSes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207388
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Mon, 28 Apr 2014 01:24:32 +0000 (01:24 +0000)]
Update tests to use the new format of printing a TimeValue. It's a bit
odd to have the output of 'llvm-ar tv' depend on the format of
TimeValue::str(), but that's what we have today. If anyone needs the
output to remain compatible with GNU ar or old versions of llvm-ar, just
shout and I'll switch the code to manually format its times.
Note that there isn't a portable format -- Mac and GNU have different
formats at least (thanks Rafael!) so...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207387
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 27 Apr 2014 23:59:25 +0000 (23:59 +0000)]
Teach the pass manager's execution dump to print the current time before
each line. This is particularly nice for tracking which run of
a particular pass over a particular function was slow.
This also required making the TimeValue string much more useful. First,
there is a standard format for writing out a date and time. Let's use
that rather than strings that would have to be parsed. Second, actually
output the nanosecond resolution that timevalue claims to have.
This is proving useful working on PR19499, so I figured it would be
generally useful to commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207385
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 27 Apr 2014 23:57:57 +0000 (23:57 +0000)]
CodeGen/AsmPrinter.h: Fix \param in r207369. [-Wdocumentation]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207384
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 23:22:43 +0000 (23:22 +0000)]
Convert AddNodeIDNode and SelectionDAG::getNodeIfExiists to use ArrayRef<SDValue>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207383
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 27 Apr 2014 20:23:58 +0000 (20:23 +0000)]
Add emitThumbSet to the arm target streamer.
This fixes the asm printer implementation and lets the parser be unaware of
what .thumb_set is.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207381
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 19:40:43 +0000 (19:40 +0000)]
Fix an assert I accidentally broke to hopefully fix the build bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207380
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 19:21:20 +0000 (19:21 +0000)]
Convert SelectionDAGISel::MorphNode to use ArrayRef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207379
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 19:21:16 +0000 (19:21 +0000)]
Convert SelectionDAG::MorphNodeTo to use ArrayRef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207378
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 19:21:11 +0000 (19:21 +0000)]
Convert SelectionDAG::SelectNodeTo to use ArrayRef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207377
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 19:21:06 +0000 (19:21 +0000)]
Convert one last signature of getNode to take an ArrayRef of SDUse.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207376
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 19:21:02 +0000 (19:21 +0000)]
Convert SDNode constructor to use ArrayRef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207375
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 19:20:57 +0000 (19:20 +0000)]
Convert SelectionDAG::getMergeValues to use ArrayRef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207374
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 27 Apr 2014 19:20:47 +0000 (19:20 +0000)]
Const-correct SelectionDAG::getAtomic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207373
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Sun, 27 Apr 2014 18:50:45 +0000 (18:50 +0000)]
Clarify the doxygen comment for AsmPrinter::EmitDwarfRegOpPiece and add
default arguments to the function.
No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207372
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 27 Apr 2014 18:47:54 +0000 (18:47 +0000)]
X86TTI: Adjust sdiv cost now that we can lower it on plain SSE2.
Includes a fix for a horrible typo that caused all SDIV costs to be
slightly off :)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207371
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 27 Apr 2014 18:47:41 +0000 (18:47 +0000)]
X86: If SSE4.1 is missing lower SMUL_LOHI of v4i32 to pmuludq and fix up the high parts.
This is more expensive than pmuldq but still cheaper than scalarizing the whole thing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207370
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Sun, 27 Apr 2014 18:25:45 +0000 (18:25 +0000)]
Debug info: Refactor EmitDwarfRegOpPiece to be a member function of
AsmPrinter.
No functional change.
http://reviews.llvm.org/D3373
rdar://problem/
15928306
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207369
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Sun, 27 Apr 2014 18:25:40 +0000 (18:25 +0000)]
Debug Info: Prepare DebugLocEntry to handle more than a single value per
entry. This is in preparation for generic DW_OP_piece support.
No functional change so far.
http://reviews.llvm.org/D3373
rdar://problem/
15928306
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207368
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 27 Apr 2014 17:23:37 +0000 (17:23 +0000)]
Make getOrCreateSymbolData non virtual.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207367
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sun, 27 Apr 2014 17:10:46 +0000 (17:10 +0000)]
Avoid using MCSymbolData on the asm streamer.
Only the object streamers need to track if a symbol should be marked thumb or
not. This ports the ELF case. The COFF case is not ported since it is currently
not working for some other reason (I will report a bug).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207366
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 16:10:57 +0000 (16:10 +0000)]
MC: duplicate .file test for WoA (SVN r207341)
Since the COFF tests are dependent on X86, duplicate the test for ARM. Use the
default check prefix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207365
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 27 Apr 2014 14:54:59 +0000 (14:54 +0000)]
Make helper functions static.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207359
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sun, 27 Apr 2014 14:47:23 +0000 (14:47 +0000)]
Remove redundant explicit default initialization of non-trivially constructed member.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207357
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 27 Apr 2014 11:59:44 +0000 (11:59 +0000)]
Add the default constructor DwarfAccelTable::DataArray() to initialize (MCSymbol*)StrSym explicitly.
It will fix crash in codegen on msvc x64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207356
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sun, 27 Apr 2014 11:59:33 +0000 (11:59 +0000)]
Revert r206989, "Mark llvm/test/BugPoint/compile-custom.ll as XFAIL:vg_leak." It has been fixed since r207265.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207355
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 27 Apr 2014 11:54:45 +0000 (11:54 +0000)]
Update test not to check for a shuffle of an all-zero vector.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207354
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 27 Apr 2014 11:41:06 +0000 (11:41 +0000)]
SelectionDAG: Aggressively fold shuffles of constant splats.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207352
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 05:28:10 +0000 (05:28 +0000)]
ARM: MSVC does not support = default
Explicitly "implement" the destructor as MSVC does not support defaulted methods
yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207350
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 04:54:16 +0000 (04:54 +0000)]
MC: restore behaviour of defaulting to ELF
This restores the previous behaviour of just assuming that if you dont specify a
valid triple that you really meant the default triple with an ELF object file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207349
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 04:29:36 +0000 (04:29 +0000)]
tests: Windows ARM now supports object emission
Update lit.cfg with the fact that LLVM can now generate WoA PE/COFF objects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207347
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 04:29:32 +0000 (04:29 +0000)]
COFF: move ARM COFF test to ARM directory
The COFF tests all assume X86. Just move the new COFF tests under ARM to
appease the build bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207346
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 03:48:22 +0000 (03:48 +0000)]
Add WoA object file emission support
Introduce support for WoA PE/COFF object file emission from LLVM. Add the new
target specific PE/COFF Streamer (ARMWinCOFFStreamer) that handles the ARM
specific behaviour of PE/COFF object emission. ARM exception information is not
yet emitted and is a TODO item.
The ARM specific object writer (ARMWinCOFFObjectWriter) handles the ARM specific
relocation handling in conjunction with the WinCOFFObjectWriter in the MC layer.
The MC layer needs to be updated to deal with the relocation adjustments.
Branch relocations are adjusted by 4 bytes (unlikely their ELF counterparts).
Minor tweaks to switch multiple conditional checks into equivalent switch
statements. The ObjectFileInfo is updated to relax the object file setup for
Windows COFF. Move the architecture checks into an assertion. Windows COFF is
currently only supported on x86, x86_64, and ARM (thumb). Rather than
defaulting to ELF, we will refuse to generate an object file. This is better
though as you do not get an (arbitrary) object file which is different from the
request.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207345
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 03:48:12 +0000 (03:48 +0000)]
MC: create X86WinCOFFStreamer for target specific behaviour
This introduces a target specific streamer, X86WinCOFFStreamer, which handles
the target specific behaviour (e.g. WinEH). This is mostly to ensure that
differences between ARM and X86 remain disjoint and do not accidentally cross
boundaries. This is the final staging change for enabling object emission for
Windows on ARM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207344
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 03:48:05 +0000 (03:48 +0000)]
MC: rename WinCOFFStreamer and move declaration out-of-line
This is in preparation for promoting WinCOFFStreamer to a base class which will
be shared by the X86 and ARM specific target COFF streamers. Also add a new
getOrCreateSymbolData interface (like MCELFStreamer) for the ARM COFF Streamer.
This makes the COFFStreamer more similar to the ELFStreamer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207343
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 03:48:01 +0000 (03:48 +0000)]
MC: style tweaks to WinCOFFStreamer
Stylistic changes to prepare for splitting up the COFFStreamer into target
specific streamers. Tweak some assertion messages. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207342
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Sun, 27 Apr 2014 03:47:57 +0000 (03:47 +0000)]
ARM: Support SingleParameterDotFile on WoA
Currently, the integrated assembler is the only choice for assembling Windows on
ARM binaries. IAS supports the .file <filename> directive which emits the file
symbol into the resulting object binary. Mark the GNU COFF information to
indicate support for this feature.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207341
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sun, 27 Apr 2014 01:59:50 +0000 (01:59 +0000)]
[LCG] Re-organize the methods for mutating a call graph to make their
API requirements much more obvious.
The key here is that there are two totally different use cases for
mutating the graph. Prior to doing any SCC formation, it is very easy to
mutate the graph. There may be users that want to do small tweaks here,
and then use the already-built graph for their SCC-based operations.
This method remains on the graph itself and is documented carefully as
being cheap but unavailable once SCCs are formed.
Once SCCs are formed, and there is some in-flight DFS building them, we
have to be much more careful in how we mutate the graph. These mutation
operations are sunk onto the SCCs themselves, which both simplifies
things (the code was already there!) and helps make it obvious that
these interfaces are only applicable within that context. The other
primary constraint is that the edge being mutated is actually related to
the SCC on which we call the method. This helps make it obvious that you
cannot arbitrarily mutate some other SCC.
I've tried to write much more complete documentation for the interesting
mutation API -- intra-SCC edge removal. Currently one aspect of this
documentation is a lie (the result list of SCCs) but we also don't even
have tests for that API. =[ I'm going to add tests and fix it to match
the documentation next.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207339
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 23:09:49 +0000 (23:09 +0000)]
DAGCombiner: Simplify code a bit, make more transforms work with vectors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207338
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 22:59:28 +0000 (22:59 +0000)]
[LCG] Add some pedantry to the use of ptrdiff_t to appease build bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207337
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 22:51:31 +0000 (22:51 +0000)]
[LCG] Eliminate more boiler plate by using the iterator facade base
class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207336
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 22:43:56 +0000 (22:43 +0000)]
[LCG] Switch the node iterator to use the new fancy adaptor base. This
is *much* cleaner, makes the iterator a full random access iterator,
etc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207335
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 26 Apr 2014 22:37:45 +0000 (22:37 +0000)]
DwarfDebug: Roll argument into call.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207334
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 26 Apr 2014 22:12:18 +0000 (22:12 +0000)]
DebugInfo: Fix and test a regression caused by r207263 causing the DW_AT_object_pointer to go missing on blocks
Noticed by inspection. Test coverage added.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207333
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 20:10:49 +0000 (20:10 +0000)]
Mark the growing path in SmallVector::push_back as cold.
It's vital for performance that the cold path of push_back isn't inlined.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207331
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Apr 2014 19:29:47 +0000 (19:29 +0000)]
Replace std::vector with SmallVector for some small, known size vectors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207330
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Apr 2014 19:29:41 +0000 (19:29 +0000)]
Convert getMemIntrinsicNode to take ArrayRef of SDValue instead of pointer and size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207329
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Liew [Sat, 26 Apr 2014 19:05:45 +0000 (19:05 +0000)]
Trivial test commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207328
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Apr 2014 18:35:24 +0000 (18:35 +0000)]
Convert SelectionDAG::getNode methods to use ArrayRef<SDValue>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207327
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 26 Apr 2014 18:35:13 +0000 (18:35 +0000)]
Remove an unused version of getMemIntrinsicNode and getNode. Additionally, these were calling makeVTList with the pointers passed in which would were unlikely to belong to SelectionDAG and likely would have just been stack pointers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207326
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 26 Apr 2014 18:25:07 +0000 (18:25 +0000)]
Include C++ source for debug info test case committed in r207323
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207324
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 26 Apr 2014 17:27:38 +0000 (17:27 +0000)]
DWARF Type Units: Avoid emitting type units under fission if the type requires an address.
Since there's no way to ensure the type unit in the .dwo and the type
unit skeleton in the .o are correlated, this cannot work.
This implementation is a bit inefficient for a few reasons, called out
in comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207323
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 16:26:41 +0000 (16:26 +0000)]
Print X86ISD::PMULDQ nodes properly in debug output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207322
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 26 Apr 2014 16:26:41 +0000 (16:26 +0000)]
DwarfDebug: Minor refactoring around type unit construction
Sinking addition of the declaration attribute down to where the
signature is added. So that if the signature is not added neither is the
declaration attribute (this will come in handy when aborting type unit
construction to instead emit the type into the CU directly in some
cases)
Pull out type unit identifier hashing just to simplify the function a
little, it'll be getting longer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207321
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 14:53:05 +0000 (14:53 +0000)]
X86TTI: i16/i32 vector div with a constant (splat) divisor are reasonably cheap now.
Turn vectorization back on.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207320
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 14:12:19 +0000 (14:12 +0000)]
X86: Lower SMUL_LOHI of v4i32 to pmuldq when SSE4.1 is available.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207318
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 13:01:03 +0000 (13:01 +0000)]
X86: Add patterns for MULHU/MULHS of v8i16 and v16i16.
This gets us pretty code for divs of i16 vectors. Turn the existing
intrinsics into the corresponding nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207317
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 13:00:53 +0000 (13:00 +0000)]
Rip out X86-specific vector SDIV lowering, make the corresponding DAGCombiner transform work on vectors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207316
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 12:06:28 +0000 (12:06 +0000)]
DAGCombiner: Turn divs of vector splats into vectorized multiplications.
Otherwise the legalizer would just scalarize everything. Support for
mulhi in the targets isn't that great yet so on most targets we get
exactly the same scalarized output. Add a test for x86 vector udiv.
I had to disable the mulhi nodes on ARM because there aren't any patterns
for it. As far as I know ARM has instructions for getting the high part of
a multiply so this should be fixed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207315
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 26 Apr 2014 12:06:11 +0000 (12:06 +0000)]
X86: Custom lower v4i32 UMUL_LOHI into 2 pmuludqs.
Test will follow soon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207314
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Zolotukhin [Sat, 26 Apr 2014 09:56:41 +0000 (09:56 +0000)]
Revert r206749 till a final decision about the intrinsics is made.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207313
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 09:45:55 +0000 (09:45 +0000)]
[LCG] Rather than removing nodes from the SCC entry set when we process
them, just skip over any DFS-numbered nodes when finding the next root
of a DFS. This allows the entry set to just be a vector as we populate
it from a uniqued source. It also removes the possibility for a linear
scan of the entry set to actually do the removal which can make things
go quadratic if we get unlucky.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207312
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 09:28:00 +0000 (09:28 +0000)]
[LCG] Rotate the full SCC finding algorithm to avoid round-trips through
the DFS stack for leaves in the call graph. As mentioned in my previous
commit, this is particularly interesting for graphs which have high fan
out but low connectivity resulting in many leaves. For such graphs, this
can remove a large % of the DFS stack traffic even though it doesn't
make the stack much smaller.
It's a bit easier to formulate this for the full algorithm because that
one stops completely for each SCC. For example, I was able to directly
eliminate the "Recurse" boolean used to continue an outer loop from the
inner loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207311
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 09:06:53 +0000 (09:06 +0000)]
[LCG] Hoist the main DFS loop out of the edge removal function. This
makes working through the worklist much cleaner, and makes it possible
to avoid the 'bool-to-continue-the-outer-loop' hack. Not a huge
difference, but I think this is approaching as polished as I can make
it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207310
91177308-0d34-0410-b5e6-
96231b3b80d8
Gerolf Hoflehner [Sat, 26 Apr 2014 05:58:11 +0000 (05:58 +0000)]
RecursivelyDeleteTriviallyDeadInstructions() could remove
more than 1 instruction. The caller need to be aware of this
and adjust instruction iterators accordingly.
rdar://
16679376
Repaired r207302.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207309
91177308-0d34-0410-b5e6-
96231b3b80d8
Gerolf Hoflehner [Sat, 26 Apr 2014 05:43:41 +0000 (05:43 +0000)]
Restore CloneFunction.cpp which got accidently
overwritten by previous backout of r207303
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207308
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 03:36:42 +0000 (03:36 +0000)]
[LCG] In the incremental SCC re-formation, lift the node currently being
processed in the DFS out of the stack completely. Keep it exclusively in
a variable. Re-shuffle some code structure to make this easier. This can
have a very dramatic effect in some cases because call graphs tend to
look like a high fan-out spanning tree. As a consequence, there are
a large number of leaf nodes in the graph, and this technique causes
leaf nodes to never even go into the stack. While this only reduces the
max depth by 1, it may cause the total number of round trips through the
stack to drop by a lot.
Now, most of this isn't really relevant for the incremental version. =]
But I wanted to prototype it first here as this variant is in ways more
complex. As long as I can get the code factored well here, I'll next
make the primary walk look the same. There are several refactorings this
exposes I think.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207306
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 03:36:37 +0000 (03:36 +0000)]
[LCG] Special case the removal of self edges. These don't impact the SCC
graph in any way because we don't track edges in the SCC graph, just
nodes. This also lets us add a nice assert about the invariant that
we're working on at least a certain number of nodes within the SCC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207305
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Sat, 26 Apr 2014 02:58:04 +0000 (02:58 +0000)]
[DAG] During DAG legalization keep opaque constants even after expanding.
The included test case would return the incorrect results, because the expansion
of an shift with a constant shift amount of 0 would generate undefined behavior.
This is because ExpandShiftByConstant assumes that all shifts by constants with
a value of 0 have already been optimized away. This doesn't happen for opaque
constants and usually this isn't a problem, because opaque constants won't take
this code path - they are not supposed to. In the case that the opaque constant
has to be expanded by the legalizer, the legalizer would drop the opaque flag.
In this case we hit the limitations of ExpandShiftByConstant and create incorrect
code.
This commit fixes the legalizer by not dropping the opaque flag when expanding
opaque constants and adding an assertion to ExpandShiftByConstant to catch this
not supported case in the future.
This fixes <rdar://problem/
16718472>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207304
91177308-0d34-0410-b5e6-
96231b3b80d8
Gerolf Hoflehner [Sat, 26 Apr 2014 02:03:17 +0000 (02:03 +0000)]
Revert commit r207302 since build failures
have been reported.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207303
91177308-0d34-0410-b5e6-
96231b3b80d8
Gerolf Hoflehner [Sat, 26 Apr 2014 01:19:16 +0000 (01:19 +0000)]
RecursivelyDeleteTriviallyDeadInstructions() could remove
more than 1 instruction. The caller need to be aware of this
and adjust instruction iterators accordingly.
rdar://
16679376
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207302
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Sat, 26 Apr 2014 01:11:26 +0000 (01:11 +0000)]
[X86] Implement TargetLowering::getScalingFactorCost hook.
Scaling factors are not free on X86 because every "complex" addressing mode
breaks the related instruction into 2 allocations instead of 1.
<rdar://problem/
16730541>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207301
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 26 Apr 2014 01:03:46 +0000 (01:03 +0000)]
[LCG] Refactor the duplicated code I added in my last commit here into
a helper function. Also factor the other two places where we did the
same thing into the helper function. =] Much cleaner this way. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207300
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrea Di Biagio [Sat, 26 Apr 2014 01:03:22 +0000 (01:03 +0000)]
[InstCombine][X86] Teach how to fold calls to SSE2/AVX2 packed logical shift
right intrinsics.
A packed logical shift right with a shift count bigger than or equal to the
element size always produces a zero vector. In all other cases, it can be
safely replaced by a 'lshr' instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207299
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Sat, 26 Apr 2014 00:53:26 +0000 (00:53 +0000)]
Add missing include guards and missing #include, found by modules build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207298
91177308-0d34-0410-b5e6-
96231b3b80d8
Filipe Cabecinhas [Sat, 26 Apr 2014 00:02:37 +0000 (00:02 +0000)]
Appease the almighty buildbots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207295
91177308-0d34-0410-b5e6-
96231b3b80d8
Filipe Cabecinhas [Fri, 25 Apr 2014 23:51:17 +0000 (23:51 +0000)]
Optimization for certain shufflevector by using insertps.
Summary:
If we're doing a v4f32/v4i32 shuffle on x86 with SSE4.1, we can lower
certain shufflevectors to an insertps instruction:
When most of the shufflevector result's elements come from one vector (and
keep their index), and one element comes from another vector or a memory
operand.
Added tests for insertps optimizations on shufflevector.
Added support and tests for v4i32 vector optimization.
Reviewers: nadav
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D3475
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207291
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 23:16:58 +0000 (23:16 +0000)]
Revert "blockfreq: Approximate irreducible control flow"
This reverts commit r207286. It causes an ICE on the
cmake-llvm-x86_64-linux buildbot [1]:
llvm/lib/Analysis/BlockFrequencyInfo.cpp: In lambda function:
llvm/lib/Analysis/BlockFrequencyInfo.cpp:182:1: internal compiler error: in get_expr_operands, at tree-ssa-operands.c:1035
[1]: http://bb.pgr.jp/builders/cmake-llvm-x86_64-linux/builds/12093/steps/build_llvm/logs/stdio
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207287
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 23:08:57 +0000 (23:08 +0000)]
blockfreq: Approximate irreducible control flow
Previously, irreducible backedges were ignored. With this commit,
irreducible SCCs are discovered on the fly, and modelled as loops with
multiple headers.
This approximation specifies the headers of irreducible sub-SCCs as its
entry blocks and all nodes that are targets of a backedge within it
(excluding backedges within true sub-loops). Block frequency
calculations act as if we insert a new block that intercepts all the
edges to the headers. All backedges and entries to the irreducible SCC
point to this imaginary block. This imaginary block has an edge (with
even probability) to each header block.
The result is now reasonable enough that I've added a number of
testcases for irreducible control flow. I've outlined in
`BlockFrequencyInfoImpl.h` ways to improve the approximation.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207286
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Fri, 25 Apr 2014 23:00:25 +0000 (23:00 +0000)]
Unbreak the gdb buildbot by not lowering dbg.declare intrinsics for arrays.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207284
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 25 Apr 2014 22:23:54 +0000 (22:23 +0000)]
Make sure that rangelists are also relative to the compile unit
low_pc similar to location lists.
Fixes PR19563
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207283
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 25 Apr 2014 22:22:01 +0000 (22:22 +0000)]
R600: Fix function name printing in LowerCall
v2: Check both ExternalSymbol and GlobalAddress
Patch by: Jan Vesely <jan.vesely@rutgers.edu>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207282
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 22:21:35 +0000 (22:21 +0000)]
DwarfAccelTable: Store the string symbol in the accelerator table to avoid duplicate lookup.
This also avoids the need for subtly side-effecting calls to manifest
strings in the string table at the point where items are added to the
accelerator tables.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207281
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Roeder [Fri, 25 Apr 2014 21:46:51 +0000 (21:46 +0000)]
Add an -mattr option to the gold plugin to support subtarget features in LTO
This adds support for an -mattr option to the gold plugin and to llvm-lto. This
allows the caller to specify details of the subtarget architecture, like +aes,
or +ssse3 on x86. Note that this requires a change to the include/llvm-c/lto.h
interface: it adds a function lto_codegen_set_attr and it increments the
version of the interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207279
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Fri, 25 Apr 2014 21:42:35 +0000 (21:42 +0000)]
Fix missing include
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207278
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 21:34:35 +0000 (21:34 +0000)]
Encapsulate the DWARF string pool in a separate type.
Pulls out some more code from some of the rather monolithic DWARF
classes. Unlike the address table, the string table won't move up into
DwarfDebug - each DWARF file has its own string table (but there can be
only one address table).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207277
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Fri, 25 Apr 2014 21:30:03 +0000 (21:30 +0000)]
[DWARF parser] Cleanup code in DWARFDebugAranges.
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207276
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Fri, 25 Apr 2014 21:10:56 +0000 (21:10 +0000)]
[DWARF parser] Cleanup code in DWARFDebugAbbrev.
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207274
91177308-0d34-0410-b5e6-
96231b3b80d8
Adam Nemet [Fri, 25 Apr 2014 21:02:21 +0000 (21:02 +0000)]
[LoopStrengthReduce] Don't trim formula that uses a subset of required registers
Consider this use from the new testcase:
LSR Use: Kind=ICmpZero, Offsets={0}, widest fixup type: i32
reg({1000,+,-1}<nw><%for.body>)
-3003 + reg({3,+,3}<nw><%for.body>)
-1001 + reg({1,+,1}<nuw><nsw><%for.body>)
-1000 + reg({0,+,1}<nw><%for.body>)
-3000 + reg({0,+,3}<nuw><%for.body>)
reg({-1000,+,1}<nw><%for.body>)
reg({-3000,+,3}<nsw><%for.body>)
This is the last use we consider for a solution in SolveRecurse, so CurRegs is
a large set. (CurRegs is the set of registers that are needed by the
previously visited uses in the in-progress solution.)
ReqRegs is {
{3,+,3}<nw><%for.body>,
{1,+,1}<nuw><nsw><%for.body>
}
This is the intersection of the regs used by any of the formulas for the
current use and CurRegs.
Now, the code requires a formula to contain *all* these regs (the comment is
simply wrong), otherwise the formula is immediately disqualified. Obviously,
no formula for this use contains two regs so they will all get disqualified.
The fix modifies the check to allow the formula in this case. The idea is
that neither of these formulae is introducing any new registers which is the
point of this early pruning as far as I understand.
In terms of set arithmetic, we now allow formulas whose used regs are a subset
of the required regs not just the other way around.
There are few more loops in the test-suite that are now successfully LSRed. I
have benchmarked those and found very minimal change.
Fixes <rdar://problem/
13965777>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207271
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 20:52:08 +0000 (20:52 +0000)]
SCC: Use the reference typedef
Actually use the `reference` typedef, and remove the private
redefinition of `pointer` since it has no users.
Using `reference` exposes a problem with r207257, which specified the
wrong `value_type` to `iterator_facade_base` (fixed that too).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207270
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Fri, 25 Apr 2014 20:49:25 +0000 (20:49 +0000)]
This reapplies r207235 with an additional bugfixes caught by the msan
buildbot - do not insert debug intrinsics before phi nodes.
Debug info for optimized code: Support variables that are on the stack and
described by DBG_VALUEs during their lifetime.
Previously, when a variable was at a FrameIndex for any part of its
lifetime, this would shadow all other DBG_VALUEs and only a single
fbreg location would be emitted, which in fact is only valid for a small
range and not the entire lexical scope of the variable. The included
dbg-value-const-byref testcase demonstrates this.
This patch fixes this by
Local
- emitting dbg.value intrinsics for allocas that are passed by reference
- dropping all dbg.declares (they are now fully lowered to dbg.values)
SelectionDAG
- renamed constructors for SDDbgValue for better readability.
- fix UserValue::match() to handle indirect values correctly
- not inserting an MMI table entries for dbg.values that describe allocas.
- lowering dbg.values that describe allocas into *indirect* DBG_VALUEs.
CodeGenPrepare
- leaving dbg.values for an alloca were they are (see comment)
Other
- regenerated/updated instcombine.ll testcase and included source
rdar://problem/
16679879
http://reviews.llvm.org/D3374
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207269
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 20:19:11 +0000 (20:19 +0000)]
MCAssembler: Simplify implementation of const variants of getSymbolData by calling one implementation from the other.
Code review feedback by Rafael Espindola on r207124.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207266
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 20:15:16 +0000 (20:15 +0000)]
BugPoint: Fix some memory leaks.
Patch by Kostya Serebryany.
unique_ptr would be nice, but it's a bit too much work for an area I'm
not familiar with, nor invested in, unfortunately.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207265
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 20:02:24 +0000 (20:02 +0000)]
DwarfUnit: Remove unused function
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207264
91177308-0d34-0410-b5e6-
96231b3b80d8