Commit Graph

94653 Commits

Author SHA1 Message Date
Richard Sandiford
0e4044c233 [SystemZ] Add LOAD AND TEST instructions
Just the definitions and MC support.  The next patch uses them for codegen.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187719 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-05 11:00:53 +00:00
Richard Sandiford
66fbb47818 [SystemZ] Split out comparison elimination into a separate pass
Perhaps predictably, doing comparison elimination on the fly during
SystemZLongBranch turned out to be a bad idea.  The next patches make
use of LOAD AND TEST and BRANCH ON COUNT, both of which require
changes to earlier instructions.

No functionality change intended.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187718 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-05 10:58:53 +00:00
Elena Demikhovsky
13e6e9171f AVX-512 set: added mask operations, lowering BUILD_VECTOR for i1 vector types.
Added intrinsics and tests.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187717 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-05 08:52:21 +00:00
Nadav Rotem
f18d7f8cb8 Update the release notes about the status of the vectorizers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187714 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-05 04:31:05 +00:00
Nadav Rotem
055028f6c8 Update the docs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187713 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-05 04:27:34 +00:00
Reed Kotler
25278aa26f Add the saving of S2. This is needed for some of the floating point
helper functions. This can be optimized out later when the remaining
parts of the helper function work is moved into the Mips16HardFloat pass.
For now it forces us to use the 32 bit save/restore instructions instead
of the 16 bit ones.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187712 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 23:56:53 +00:00
Bob Wilson
9c6d857c61 Remove "lto_on_osx" xfails, now that -rdynamic works on Darwin.
Note that this will require a recent version of the linker for Darwin
builds with LTO to pass these tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187711 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 23:55:24 +00:00
Bob Wilson
0cffe07771 Build with the $RDYNAMIC flag on Darwin as well as other platforms.
Part of <rdar://problem/14620988>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187710 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 22:06:11 +00:00
Benjamin Kramer
75311b7b4d X86: Turn fp selects into mask operations.
double test(double a, double b, double c, double d) { return a<b ? c : d; }

before:
_test:
	ucomisd	%xmm0, %xmm1
	ja	LBB0_2
	movaps	%xmm3, %xmm2
LBB0_2:
	movaps	%xmm2, %xmm0

after:
_test:
	cmpltsd	%xmm1, %xmm0
	andpd	%xmm0, %xmm2
	andnpd	%xmm3, %xmm0
	orpd	%xmm2, %xmm0

Small speedup on Benchmarks/SmallPT

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187706 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 12:05:16 +00:00
Elena Demikhovsky
f3e3417e65 AVX-512 set: added VEXTRACTPS instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187705 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 10:46:07 +00:00
Tim Northover
b04578c21b X86: specify CPU on new test to fix atom buildbot
Apparently Atoms use lea for stack adjustment, which we weren't
looking for.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187704 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 10:00:45 +00:00
Tim Northover
a54b66277e X86: correct tail return address calculation
Due to the weird and wondeful usual arithmetic conversions, some
calculations involving negative values were getting performed in
uint32_t and then promoted to int64_t, which is really not a good
idea.

Patch by Katsuhiro Ueno.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187703 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 09:35:57 +00:00
Benjamin Kramer
a2b0c335d8 AsmParser: Store MacroLikeBodies on the side so they don't get leaked.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187702 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 09:06:29 +00:00
Reed Kotler
8a7f9de9d4 Clean up code for Mips16 large frame handling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187701 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-04 01:13:25 +00:00
Benjamin Kramer
af00feb1a6 PPCAsmParser: Stop leaking names.
Store them in a place that gets cleaned up properly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187700 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-03 22:43:29 +00:00
Benjamin Kramer
bd194ced30 Unbreak llvm-rtdyld build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187699 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-03 22:18:45 +00:00
Benjamin Kramer
782fdcefa5 MachObjectFile: Don't leak on error.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187698 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-03 22:16:37 +00:00
Benjamin Kramer
a757e93626 llvm-rtdyld: Don't leak memory managers.
Dyld never outlives MemMgr, just put both on the stack.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187697 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-03 22:16:31 +00:00
Benjamin Kramer
527692a594 ARMAsmParser: Plug a leak.
Using an object to do the cleanup may look like overkill, but it's safer and nicer than putting deletes everywhere.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187696 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-03 22:16:24 +00:00
Benjamin Kramer
bc2160f7c9 Stop leaking register infos in the disassemblers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187695 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-03 22:16:16 +00:00
Hal Finkel
5cad12d12a Fix PPC64 64-bit GPR inline asm constraint matching
Internally, the PowerPC backend names the 32-bit GPRs R[0-9]+, and names the
64-bit parent GPRs X[0-9]+. When matching inline assembly constraints with
explicit register names, on PPC64 when an i64 MVT has been requested, we need
to follow gcc's convention of using r[0-9]+ to refer to the 64-bit (parent)
registers.

At some point, we'll probably want to arrange things so that the generic code
in TargetLowering uses the AsmName fields declared in *RegisterInfo.td in order
to match these inline asm register constraints. If we do that, this change can
be reverted.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187693 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-03 12:25:10 +00:00
Matt Arsenault
6b4dde71cf Minor address space code simplification.
Remove assertion that the verifier should catch.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187692 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-03 01:03:12 +00:00
Bob Wilson
f3289e7d74 Regenerate with changes for -rdynamic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187687 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 22:51:11 +00:00
Bob Wilson
244acf3a4b Link with -rdynamic instead of -Wl,-export-dynamic.
Recent versions of the OS X linker support this but follow the existing
OS X linker convention of using an underscore in the option name, i.e.,
-export_dynamic. Rather than changing our configure scripts to check for
that alternate spelling, it is simpler to just use the compiler's -rdynamic
option and let it deal with translating that to the appropriate linker
option. One potential disadvantage of this approach is that the compiler
will typically ignore -rdynamic on platforms where it is not supported, so
the HAVE_LINK_EXPORT_DYNAMIC in config.h will not necessarily show whether
that option has any effect or not. I don't see any in-tree uses of that
macro, so I'm assuming it is OK.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187686 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 22:51:06 +00:00
Peter Collingbourne
40bacacad3 Add a AttributeSetImpl::dump function.
This is for the benefit of those of us with inferior debuggers which
do not permit member function calls on value types.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187685 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 22:34:30 +00:00
Peter Collingbourne
7bba9c5c0a Make one of the AttributeSet ctors maintain the invariant that the
attribute list is ordered by index.

Differential Revision: http://llvm-reviews.chandlerc.com/D1265

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187682 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 22:29:40 +00:00
Hans Wennborg
6bf104b165 Option parsing: recognize the special -- token
Everything that comes after -- should be treated as a filename. This
enables passing in filenames that would otherwise be conflated with
command-line options.

This is especially important for clang-cl which supports options
starting with /, which are easily conflatable with Unix-style
path names.

Differential Revision: http://llvm-reviews.chandlerc.com/D1274

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187675 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 21:20:27 +00:00
Hal Finkel
e8bc700a87 Fix invalid function pointers in bugpoint ExtractLoops
The ExtractLoops function tries to reduce the failing test case by extracting
one or more loops from the misoptimized piece of the program. In doing this,
ExtractLoops must keep the MiscompiledFunctions vector up-to-date by ensuring
that the pointers refer to functions in the current failing program.
Unfortunately, this is not trivial because:

 - ExtractLoops is iterative, and there are several early exits (and the
   MiscompiledFunctions vector must be consistent with the current program at
every non-fatal exit point).
 - Several of the utility functions used by ExtractLoops (such as
   TestOptimizer, some of which are called through the TestFn callback
parameter, and Linker::LinkModules) delete their inputs upon success.

This change adds several updates of the MiscompiledFunctions vector at
different points. The first is after the initial call to TestMergedProgram
which checks that the loop-extracted program still works. The second is after
the call to TestFn (TestOptimizer, for example). This function will delete its
inputs (which is why the existing ExtractLoops logic cloned the inputs first).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187674 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 21:13:42 +00:00
Joey Gouly
e7bc73b8d1 Add a missing 'return' statement.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187671 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 20:50:01 +00:00
Akira Hatanaka
5e795097b0 [mips] Expand vector truncating stores and extending loads.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187667 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 19:23:33 +00:00
Joey Gouly
787fdb86a7 [ARMv8] Add an assembler warning for the deprecated 'setend' instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187666 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 19:18:12 +00:00
Nadav Rotem
f2855f8886 SLPVectorizer: Fix PR16777. PHInodes may use multiple extracted values that come from different blocks.
Thanks Alexey Samsonov.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187663 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 18:40:24 +00:00
Matt Arsenault
38140c2bbe Teach EmitGEPOffset about address spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187662 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 18:33:34 +00:00
Renato Golin
38ffffeebc Fixes ARM LNT bot from SLP change in O3
This patch fixes the multiple breakages on ARM test-suite after the SLP
vectorizer was introduced by default on O3. The problem was an illegal
vector type on ARMTTI::getCmpSelInstrCost() <3 x i1> which is not simple.

The guard protects this code from breaking (cause of the problems) but
doesn't fix the issue that is generating the odd vector in the first
place, which also needs to be investigated.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187658 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 17:10:04 +00:00
Carlo Kok
fdabd9f391 Bugfix for making the DWARF debug strings and labels to code emitted as secrel32 instead of long opcodes (only for coff). This makes them debuggable with GDB (with fix for 64bits msvc)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187656 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 16:14:15 +00:00
Tim Northover
e57343b636 Fix handling of CHECK-DAG combined with CHECK-NOT
Patch by Daniel Sanders.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187651 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 11:32:50 +00:00
Duncan Sands
1cd2ed8fdf Pacify GCC, which worries about falling off the end of the switch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187649 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 09:37:20 +00:00
Alexey Samsonov
e393b73808 Fix dereferencing end iterator in SimplifyCFG. Patch by Ye Mei.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187646 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 08:06:43 +00:00
NAKAMURA Takumi
ee1d7db74b Revert r187597, "Bugfix for making the DWARF debug strings and labels to code emitted as secrel32 instead of long opcodes (only for coff). This makes them debuggable with GDB."
It broke x86_64-win32 builder in llvm/test/DebugInfo.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187642 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 03:46:05 +00:00
Eric Christopher
9f4eaaf0dc Use @rpath for libraries rather than @executable_path on OSX.
Patch by Benjamin Scarlet!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187641 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 01:51:52 +00:00
Eric Christopher
be9f508aec Temporarily revert "Debug Info Finder|Verifier: handle DbgLoc attached to
instructions." in an attempt to bring back some bots.

This reverts commit r187609.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187638 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 00:49:44 +00:00
Matt Arsenault
7eef3bdd7b Teach InstructionSimplify about pointer address spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187635 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-02 00:10:44 +00:00
Akira Hatanaka
ddbdeefa28 [mips] Make load/store accumulator pseudo instructions codeGenOnly. Also,
remove lines that are setting DecoderNamespace for pseudo atomic instructions.

No intended functionality change.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187632 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-01 23:14:16 +00:00
Matt Arsenault
186f8f9d41 Teach getOrEnforceKnownAlignment about address spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187629 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-01 22:42:18 +00:00
Nadav Rotem
72bc423e7f Move the optlevel check to the frontend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187628 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-01 22:41:58 +00:00
Carlo Kok
cfe855d658 fix for LLVM debug info on llvm-mips-linux where the label name uses % instead of L as a prefix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187623 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-01 22:15:34 +00:00
Bill Wendling
61fc8d670f Use function attributes to indicate that we don't want to realign the stack.
Function attributes are the future! So just query whether we want to realign the
stack directly from the function instead of through a random target options
structure.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187618 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-01 21:42:05 +00:00
Rafael Espindola
8cb1d81250 Expose that the unique file ID has a device and a file component.
The use of sd_dev and st_ino has reached libclang, so expose the two components
in UniqueID so that we can use it in clang.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187616 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-01 21:36:02 +00:00
Daniel Malea
44c8e346c7 Fixed the Intel-syntax X86 disassembler to respect the (existing) option for hexadecimal immediates, to match AT&T syntax. This also brings a new option for C-vs-MASM-style hex.
Patch by Richard Mitton
Reviewed: http://llvm-reviews.chandlerc.com/D1243



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187614 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-01 21:18:16 +00:00
Reed Kotler
b1fa7d4d26 Fix some issues with Mips16 floating when certain intrinsics are present.
This is actually an LLVM bug in the way it generates signatures for these
when soft float is enabled. For example, floor ends up having the signature
of int64(int64). The signature part is not the same as where the actual
parameter types are recorded, and those ARE of course int64(int64) when
soft float is enabled. (Yes, Mips16 hard float uses soft float but with
different runtime rounes but then has to interoperate with Mips32 using
normal floating point). This logic will eventually be moved to the 
Mips16HardFloat pass so it's not worth sorting out these issues in LLVM
since nobody but Mips16 cares about these signatures, as far as I know,
and even I won't eventually either.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187613 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-01 21:17:53 +00:00