OSDN Git Service
Craig Topper [Sun, 4 Mar 2012 03:33:22 +0000 (03:33 +0000)]
Use uint16_t to store registers in callee saved register tables to reduce size of static data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151996
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 4 Mar 2012 02:16:41 +0000 (02:16 +0000)]
Use uint8_t instead of enums to store values in X86 disassembler table. Shaves 150k off the size of X86DisassemblerDecoder.o
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151995
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 3 Mar 2012 20:44:43 +0000 (20:44 +0000)]
Perform the string table optimization for OperandMatchEntries too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151986
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 3 Mar 2012 19:13:26 +0000 (19:13 +0000)]
Shrink the asm matcher tables.
- Shrink the opcode field to 16 bits.
- Shrink the AsmVariantID field to 8 bits.
- Store the mnemonic string in a string table, store a 16 bit index.
- Store a pascal-style length byte in the string instead of a null terminator,
so we can avoid calling strlen on every entry we visit during mnemonic search.
Shrinks X86AsmParser.o from 434k to 201k on x86_64 and eliminates relocs from the table.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151984
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 3 Mar 2012 19:13:20 +0000 (19:13 +0000)]
StringToOffsetTable: Allow uniquing the first element, add an option to skip appending a terminating null.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151983
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 3 Mar 2012 15:04:13 +0000 (15:04 +0000)]
Enable the small vector POD optimization for BitCodeAbbrevOp.
While at it bump the small vector size a bit, it's inside a heap-allocated class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151980
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Sat, 3 Mar 2012 14:24:15 +0000 (14:24 +0000)]
Correctly initialize LineSectionSymbol. Thanks to Duncan Sands for noticing it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151979
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Sat, 3 Mar 2012 13:30:56 +0000 (13:30 +0000)]
Honour --config-prefix also for lit.local.cfg.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151977
91177308-0d34-0410-b5e6-
96231b3b80d8
Francois Pichet [Sat, 3 Mar 2012 09:39:54 +0000 (09:39 +0000)]
Move the NonPOD struct out of the anonymous namespace instead of adding llvm:: everywhere to fix the HashingTest on MSVC .
chandlerc proposed this better solution on IRC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151974
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Sat, 3 Mar 2012 09:36:58 +0000 (09:36 +0000)]
Include cctype for isdigit. Patch by Stephen Hines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151973
91177308-0d34-0410-b5e6-
96231b3b80d8
Francois Pichet [Sat, 3 Mar 2012 07:56:49 +0000 (07:56 +0000)]
Fixes the Hashing tests on MSVC by adding llvm:: prefix to hash_value function call.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151971
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sat, 3 Mar 2012 07:00:58 +0000 (07:00 +0000)]
unittests/ADT/HashingTest.cpp: Temporarily disable a new test introduced in r151891, to appease msvc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151970
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Sat, 3 Mar 2012 04:32:33 +0000 (04:32 +0000)]
llvm/docs/GarbageCollection.html: Prune utf8 BOM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151968
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sat, 3 Mar 2012 00:35:48 +0000 (00:35 +0000)]
hash_state: Don't use initialization target during initialization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151959
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 3 Mar 2012 00:26:30 +0000 (00:26 +0000)]
Fix RA-dependent test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151958
91177308-0d34-0410-b5e6-
96231b3b80d8
David Meyer [Fri, 2 Mar 2012 23:43:51 +0000 (23:43 +0000)]
Fix comments for llvm-readobj, remove extraneous headers
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151957
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 2 Mar 2012 19:19:34 +0000 (19:19 +0000)]
Fix indentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151932
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Fri, 2 Mar 2012 17:46:15 +0000 (17:46 +0000)]
Tidy up. Trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151926
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Fri, 2 Mar 2012 16:24:21 +0000 (16:24 +0000)]
autoconf: Set LLVM_CONFIGTIME to a stable value when using --disable-timestamps.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151921
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 2 Mar 2012 15:34:43 +0000 (15:34 +0000)]
LVI: Recognize the form instcombine canonicalizes range checks into when forming constant ranges.
This could probably be made a lot smarter, but this is a common case and doesn't require LVI to scan a lot
of code. With this change CVP can optimize away the "shift == 0" case in Hashing.h that only gets hit when
"shift" is in a range not containing 0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151919
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 2 Mar 2012 15:34:35 +0000 (15:34 +0000)]
Hashing: microoptimize a truncate on 64 bit away. This currently blocks dead code eliminating the conditional.
The optimizer should handle this eventually, but currently LVI isn't really designed for this kind of stuff.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151918
91177308-0d34-0410-b5e6-
96231b3b80d8
Jia Liu [Fri, 2 Mar 2012 13:41:31 +0000 (13:41 +0000)]
delete dead code, patch by Michael Spencer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151909
91177308-0d34-0410-b5e6-
96231b3b80d8
Jia Liu [Fri, 2 Mar 2012 11:30:51 +0000 (11:30 +0000)]
add llvm.gcroot into GarbageCollection.html, patch bylost lostfreeman@gmail.com.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151908
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 2 Mar 2012 11:16:10 +0000 (11:16 +0000)]
Make the hashing algorithm Endian neutral. This is a bit annoying, but
folks who know something about PPC tell me that the byte swap is crazy
fast and without this the bit mixture would actually be different. It
might not be worse, but I've not measured it and so I'd rather not trust
it. This way, the algorithm is identical on both endianness hosts. I'll
look into any performance issues etc stemming from this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151892
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 2 Mar 2012 10:56:40 +0000 (10:56 +0000)]
Simplify the pair optimization. Rather than using complex type traits,
just ensure that the number of bytes in the pair is the sum of the bytes
in each side of the pair. As long as thats true, there are no extra
bytes that might be padding.
Also add a few tests that previously would have slipped through the
checking. The more accurate checking mechanism catches these and ensures
they are handled conservatively correctly.
Thanks to Duncan for prodding me to do this right and more simply.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151891
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Fri, 2 Mar 2012 10:41:08 +0000 (10:41 +0000)]
ASan: use getTypeAllocSize instead of getTypeStoreSize.
This change replaces getTypeStoreSize with getTypeAllocSize in AddressSanitizer
instrumentation for stack allocations.
One case where old behaviour produced undesired results is an optimization in
InstCombine pass (PromoteCastOfAllocation), which can replace alloca(T) with
alloca(S), where S has the same AllocSize, but a smaller StoreSize. Another
case is memcpy(long double => long double), where ASan will poison bytes 10-15
of a stack-allocated long double (StoreSize 10, AllocSize 16,
sizeof(long double) = 16).
See http://llvm.org/bugs/show_bug.cgi?id=12047 for more context.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151887
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 2 Mar 2012 10:01:29 +0000 (10:01 +0000)]
Add a golden data test that I missed somehow the first time around.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151886
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 2 Mar 2012 10:01:27 +0000 (10:01 +0000)]
Fix bad indenting that was left over from cut/paste of the golden values
for 32-bit builds in here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151885
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 2 Mar 2012 09:46:02 +0000 (09:46 +0000)]
Add a header that was technically missing to see if this gets the
offsetof buildbot errors to go away...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151884
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 2 Mar 2012 09:26:36 +0000 (09:26 +0000)]
We really want to hash pairs of directly-hashable data as directly
hashable data. This matters when we have pair<T*, U*> as a key, which is
quite common in DenseMap, etc. To that end, we need to detect when this
is safe. The requirements on a generic std::pair<T, U> are:
1) Both T and U must satisfy the existing is_hashable_data trait. Note
that this includes the requirement that T and U have no internal
padding bits or other bits not contributing directly to equality.
2) The alignment constraints of std::pair<T, U> do not require padding
between consecutive objects.
3) The alignment constraints of U and the size of T do not conspire to
require padding between the first and second elements.
Grow two somewhat magical traits to detect this by forming a pod
structure and inspecting offset artifacts on it. Hopefully this won't
cause any compilers to panic.
Added and adjusted tests now that pairs, even nested pairs, are treated
as just sequences of data.
Thanks to Jeffrey Yasskin for helping me sort through this and reviewing
the somewhat subtle traits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151883
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 2 Mar 2012 08:32:29 +0000 (08:32 +0000)]
Add support for hashing pairs by delegating to each sub-object. There is
an open question of whether we can do better than this by treating pairs
as boring data containers and directly hashing the two subobjects. This
at least makes the API reasonable.
In order to make this change, I reorganized the header a bit. I lifted
the declarations of the hash_value functions up to the top of the header
with their doxygen comments as these are intended for users to interact
with. They shouldn't have to wade through implementation details. I then
defined them at the very end so that they could be defined in terms of
hash_combine or any other hashing infrastructure.
Added various pair-hashing unittests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151882
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 2 Mar 2012 02:50:46 +0000 (02:50 +0000)]
Prevent obscure and incorrect tail-call optimization.
In this instance we are generating the tail-call during legalizeDAG. The 2nd
floor call can't be a tail call because it clobbers %xmm1, which is defined by
the first floor call. The first floor call can't be a tail-call because it's
not in the tail position. The only reasonable way I could think to fix this
in a target-independent manner was to check for glue logic on the copy reg.
rdar://
10930395
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151877
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 2 Mar 2012 02:11:47 +0000 (02:11 +0000)]
Grammar-o in function name.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151875
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 2 Mar 2012 01:57:55 +0000 (01:57 +0000)]
Grammar.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151874
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 2 Mar 2012 01:57:52 +0000 (01:57 +0000)]
If the linkage name doesn't exist we're supposed to emit a reference
to the string table for the function name, not the function name.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151873
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 2 Mar 2012 01:26:46 +0000 (01:26 +0000)]
Fix an iterator invalidation problem. operator[] on a DenseMap
can insert a new element, invalidating iterators. Use find
instead, and handle the case where the key is not found explicitly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151871
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 2 Mar 2012 01:13:53 +0000 (01:13 +0000)]
Misc micro-optimizations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151869
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 2 Mar 2012 00:48:38 +0000 (00:48 +0000)]
Remove the misguided extension here that reserved two special values in
the hash_code. I'm not sure what I was thinking here, the use cases for
special values are in the *keys*, not in the hashes of those keys.
We can always resurrect this if needed, or clients can accomplish the
same goal themselves. This makes the general case somewhat faster (~5
cycles faster on my machine) and smaller with less branching.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151865
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 2 Mar 2012 00:30:24 +0000 (00:30 +0000)]
Revert "Reorder the sections being output to reduce the number of assembler"
The inline table needs to be constructed ahead of time so that it doesn't try to
create new strings while we're emitting everything.
This reverts commit
a8ff9bccb399183cdd5f1c3cec2bda763664b4b0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151864
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 1 Mar 2012 23:27:13 +0000 (23:27 +0000)]
Neuter the optimization I implemented with r107852 and r108258 which turn some
floating point equality comparisons into integer ones with -ffast-math. The
issue is the optimization causes +0.0 != -0.0.
Now the optimization is only done when one side is known to be 0.0. The other
side's sign bit is masked off for the comparison.
rdar://
10964603
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151861
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 1 Mar 2012 23:20:45 +0000 (23:20 +0000)]
Re-disable the debug output. The comment is there explaining why we want
to keep this around -- updating golden tests is annoying otherwise.
Thanks to Benjamin for pointing this omission out on IRC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151860
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 1 Mar 2012 23:18:44 +0000 (23:18 +0000)]
Switch FoldingSet over to the new hashing infrastructure. We might want
to do more invasive refactoring here to get FoldingSet to use size_t or
even hash_code directly, but for now this is a good first step to remove
Yet Another Hashing Algorithm from LLVM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151859
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 1 Mar 2012 23:06:19 +0000 (23:06 +0000)]
Provide the 32-bit variant of the golden tests. Not sure how I forgot to
do this initially, sorry.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151857
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 1 Mar 2012 22:57:32 +0000 (22:57 +0000)]
Handle regmasks in Thumb1RegisterInfo::saveScavengerRegister().
This function could have r12 live across a function call when compiling
thumb1 code.
The test case for this is not included because it is very long. It must
provoke emergency spilling near a function call. The behavior is
provoked by MultiSource/Applications/JM/lencod, and it triggers an
assertion in the scavenger.
<rdar://problem/
10963642>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151855
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 1 Mar 2012 22:50:31 +0000 (22:50 +0000)]
Reorder the sections being output to reduce the number of assembler
fixups that are being used to determine section offsets. Reduces
the total number of fixups by 50% for a non-trivial testcase.
Part of rdar://
10413936
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151852
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 1 Mar 2012 22:47:09 +0000 (22:47 +0000)]
ARM use the right opcode for FP<->Integer move in fast-isel.
rdar://
10965031
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151850
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 1 Mar 2012 22:42:52 +0000 (22:42 +0000)]
Minimal changes for LLVM to compile under VS11.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151849
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 1 Mar 2012 22:27:29 +0000 (22:27 +0000)]
Changes for migrating to using register mask operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151847
91177308-0d34-0410-b5e6-
96231b3b80d8
David Meyer [Thu, 1 Mar 2012 22:19:54 +0000 (22:19 +0000)]
[Object]
Add ObjectFile::getLoadName() for retrieving the soname/installname of a shared object.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151845
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Thu, 1 Mar 2012 22:13:02 +0000 (22:13 +0000)]
Change ARMInstPrinter::printPredicateOperand() so it will not abort if it
runs into the undefined 15 condition code value.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151844
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 1 Mar 2012 22:12:30 +0000 (22:12 +0000)]
Fix bugs which were introduced when support for base+index floating point loads
and stores was added.
- SelectAddr should return false if Parent is an unaligned f32 load or store.
- Only aligned load and store nodes should be matched to select reg+imm
floating point instructions.
- MIPS does not have support for f64 unaligned load or store instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151843
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 1 Mar 2012 22:10:16 +0000 (22:10 +0000)]
BumpPtrAllocator: Make sure threshold cannot be initialized with a value smaller than the slab size.
This replaces r151834 with a simpler fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151842
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 1 Mar 2012 21:45:51 +0000 (21:45 +0000)]
Fix two warnings in this code that I missed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151839
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Thu, 1 Mar 2012 20:36:32 +0000 (20:36 +0000)]
If BumpPtrAllocator is requested to allocate a size that exceeds the slab size,
increase the slab size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151834
91177308-0d34-0410-b5e6-
96231b3b80d8
Preston Gurd [Thu, 1 Mar 2012 19:57:20 +0000 (19:57 +0000)]
Trivial change to make the test use Use –mcpu=generic,
so that the test will not fail when run on an Intel Atom
processor, due to the Atom scheduler producing an instruction sequence that is
different from that which is normally expected.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151832
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Thu, 1 Mar 2012 19:45:47 +0000 (19:45 +0000)]
Move include/llvm/ADT/SaveAndRestore.h -> include/llvm/Support/SaveAndRestore.h
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151828
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 1 Mar 2012 18:58:59 +0000 (18:58 +0000)]
Add the source file with trivial definitions in it that was missing from
r151822, sorry sorry. =[
We need 'git svn nothave' or some such...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151824
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 1 Mar 2012 18:55:25 +0000 (18:55 +0000)]
Rewrite LLVM's generalized support library for hashing to follow the API
of the proposed standard hashing interfaces (N3333), and to use
a modified and tuned version of the CityHash algorithm.
Some of the highlights of this change:
-- Significantly higher quality hashing algorithm with very well
distributed results, and extremely few collisions. Should be close to
a checksum for up to 64-bit keys. Very little clustering or clumping of
hash codes, to better distribute load on probed hash tables.
-- Built-in support for reserved values.
-- Simplified API that composes cleanly with other C++ idioms and APIs.
-- Better scaling performance as keys grow. This is the fastest
algorithm I've found and measured for moderately sized keys (such as
show up in some of the uniquing and folding use cases)
-- Support for enabling per-execution seeds to prevent table ordering
or other artifacts of hashing algorithms to impact the output of
LLVM. The seeding would make each run different and highlight these
problems during bootstrap.
This implementation was tested extensively using the SMHasher test
suite, and pased with flying colors, doing better than the original
CityHash algorithm even.
I've included a unittest, although it is somewhat minimal at the moment.
I've also added (or refactored into the proper location) type traits
necessary to implement this, and converted users of GeneralHash over.
My only immediate concerns with this implementation is the performance
of hashing small keys. I've already started working to improve this, and
will continue to do so. Currently, the only algorithms faster produce
lower quality results, but it is likely there is a better compromise
than the current one.
Many thanks to Jeffrey Yasskin who did most of the work on the N3333
paper, pair-programmed some of this code, and reviewed much of it. Many
thanks also go to Geoff Pike Pike and Jyrki Alakuijala, the original
authors of CityHash on which this is heavily based, and Austin Appleby
who created MurmurHash and the SMHasher test suite.
Also thanks to Nadav, Tobias, Howard, Jay, Nick, Ahmed, and Duncan for
all of the review comments! If there are further comments or concerns,
please let me know and I'll jump on 'em.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151822
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 1 Mar 2012 18:16:35 +0000 (18:16 +0000)]
Move getSubRegIndex out of generated code into MCRegisterInfo, devirtualize it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151821
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Thu, 1 Mar 2012 17:41:19 +0000 (17:41 +0000)]
Revert r151816 as Jim has the appropriate fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151818
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Thu, 1 Mar 2012 17:31:30 +0000 (17:31 +0000)]
Fix testcases from r151807.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151816
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 1 Mar 2012 17:30:39 +0000 (17:30 +0000)]
Move TargetRegisterInfo::getSubReg() to MCRegisterInfo.
Allows us to de-virtualize the function and provides access to it in
the instruction printer, which is useful for handling composite
physical registers (e.g., ARM register lists).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151815
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 1 Mar 2012 17:30:35 +0000 (17:30 +0000)]
Revert "Emit the SubRegTable with the smallest possible integer type."
This reverts commit 151760.
We want to move getSubReg() from TargetRegisterInfo into MCRegisterInfo,
but to do that, the type of the lookup table needs to be the same for
all targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151814
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 1 Mar 2012 17:30:32 +0000 (17:30 +0000)]
Add missing triple for tests.
Make darwin bots happier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151813
91177308-0d34-0410-b5e6-
96231b3b80d8
Jia Liu [Thu, 1 Mar 2012 15:14:19 +0000 (15:14 +0000)]
rplace Alpha with ARM in docs/WritingAnLLVMBackend.html, patch by chenwj
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151811
91177308-0d34-0410-b5e6-
96231b3b80d8
James Molloy [Thu, 1 Mar 2012 14:32:18 +0000 (14:32 +0000)]
Fix a codegen fault in which log2 or exp2 could be dead-code eliminated even though they could have sideeffects.
Only allow log2/exp2 to be converted to an intrinsic if they are declared "readnone".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151807
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 1 Mar 2012 13:37:55 +0000 (13:37 +0000)]
Make TargetRegisterClasses non-virtual by making the only virtual function a function pointer.
This allows us to make TRC non-polymorphic and value-initializable, eliminating a huge static
initializer and a ton of cruft from the generated code.
Shrinks ARMBaseRegisterInfo.o by ~100k.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151806
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 1 Mar 2012 03:14:13 +0000 (03:14 +0000)]
llvm/test/CMakeLists.txt: Update dependencies to add llvm-readobj to "check".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151795
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 1 Mar 2012 02:16:57 +0000 (02:16 +0000)]
Emit the "is an intrinsic overloaded" table as a bitfield.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151792
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 1 Mar 2012 01:53:15 +0000 (01:53 +0000)]
Pass endian information to constructors. Define separate functions to create
objects for big endian and little endian targets.
Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151788
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 1 Mar 2012 01:43:25 +0000 (01:43 +0000)]
Make InlineSpiller bundle-aware.
Simply treat bundles as instructions. Spill code is inserted between
bundles, never inside a bundle. Rewrite all operands in a bundle at
once.
Don't attempt and memory operand folding inside bundles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151787
91177308-0d34-0410-b5e6-
96231b3b80d8
David Meyer [Thu, 1 Mar 2012 01:36:50 +0000 (01:36 +0000)]
[Object]
* Add begin_dynamic_table() / end_dynamic_table() private interface to ELFObjectFile.
* Add begin_libraries_needed() / end_libraries_needed() interface to ObjectFile, for grabbing the list of needed libraries for a shared object or dynamic executable.
* Implement this new interface completely for ELF, leave stubs for COFF and MachO.
* Add 'llvm-readobj' tool for dumping ObjectFile information.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151785
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 1 Mar 2012 01:26:01 +0000 (01:26 +0000)]
Move getBundleStart() into MachineInstrBundle.h.
This allows the function to be inlined, and makes it suitable for use in
getInstructionIndex().
Also provide a const version. C++ is great for touch typing practice.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151782
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 1 Mar 2012 01:18:32 +0000 (01:18 +0000)]
Emit the intrinsic modref info as a lookup table instead of a huge switch.
Shrinks BasicAliasAnalysis.o from 106k to 56k on i386.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151781
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Thu, 1 Mar 2012 00:41:17 +0000 (00:41 +0000)]
Don't redundantly copy implicit operands when rematerializing.
While we're at it - don't copy vreg implicit operands while rematerializing.
This fixes PR12138.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151779
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Trieu [Thu, 1 Mar 2012 00:29:09 +0000 (00:29 +0000)]
Fix flags for test in MC/MachO/ARM/empty-function-nop.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151778
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Callanan [Thu, 1 Mar 2012 00:15:29 +0000 (00:15 +0000)]
Fixed the 32-bit runtime dynamic loader to allocate
code sections when needed. It just had a conditional
the wrong way around.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151777
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 29 Feb 2012 23:46:50 +0000 (23:46 +0000)]
Implement getSubRegIndex as a linear search on the SubRegTable instead of using a big switch.
- The search bounds are constant, in the worst case (ARM target) it will scan over 30 uint16_ts.
- This method isn't very hot, I had problems finding a testcase where it's called more than a dozen of times (no perf impact).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151773
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Wed, 29 Feb 2012 22:58:34 +0000 (22:58 +0000)]
Added annotations for x86 pc relative loads to llvm's 'C' disassembler.
So with darwin's otool(1) an x86_64 hello world .o file will print:
leaq L_.str(%rip), %rax ## literal pool for: Hello world
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151769
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 29 Feb 2012 22:07:56 +0000 (22:07 +0000)]
Tidy up. 80 columns.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151764
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 29 Feb 2012 21:57:08 +0000 (21:57 +0000)]
Emit the SubRegTable with the smallest possible integer type.
Doesn't help ARM with its massive register set, but halves the size on x86 and all other targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151760
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 29 Feb 2012 21:46:32 +0000 (21:46 +0000)]
Tidy up. Spelling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151758
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 29 Feb 2012 21:43:16 +0000 (21:43 +0000)]
Move the subregister indicies enum into the REGINFO_ENUM section.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151756
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Wed, 29 Feb 2012 21:02:05 +0000 (21:02 +0000)]
BitstreamWriter: Use SmallVector::append instead of multiple push_back calls.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151755
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 29 Feb 2012 20:31:17 +0000 (20:31 +0000)]
Switch TargetRegisterInfo::getSubReg() to use a lookup table.
Instead of nested switch statements, use a lookup table. On ARM, this replaces
a 23k (x86_64 release build) function with a 16k table. Its not unlikely to
be faster, as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151751
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Wed, 29 Feb 2012 20:31:09 +0000 (20:31 +0000)]
BitstreamWriter: Change primary output buffer to be a SmallVector instead of an
std::vector.
- Good for 1-2% speedup on writing PCH for Cocoa.h.
- Clang side API match to follow shortly, there wasn't an easy way to make this
non-breaking.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151750
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Wed, 29 Feb 2012 20:31:05 +0000 (20:31 +0000)]
BitstreamWriter: Isolate access to the underlying buffer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151749
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Wed, 29 Feb 2012 20:31:01 +0000 (20:31 +0000)]
BitcodeWriter: Expose less implementation details -- make BackpatchWord private
and remove getBuffer().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151748
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Wed, 29 Feb 2012 20:30:56 +0000 (20:30 +0000)]
Bitcode: Don't expose WriteBitcodeToStream to clients.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151747
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 29 Feb 2012 19:44:41 +0000 (19:44 +0000)]
Intel Atom instruction itineraries for mov sign extension and mov zero extension.
Patch by Tyler Nowicki!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151743
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Wed, 29 Feb 2012 16:43:11 +0000 (16:43 +0000)]
Add myself to the credits file
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151732
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 29 Feb 2012 13:27:00 +0000 (13:27 +0000)]
LegalizeIntegerTypes: Reorder operations in the "big shift by small amount" optimization, making the lives of later passes easier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151722
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Wed, 29 Feb 2012 11:12:03 +0000 (11:12 +0000)]
Have GVN also do condition propagation when the right-hand side is not
a constant. This fixes PR1768.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151713
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Wed, 29 Feb 2012 08:26:44 +0000 (08:26 +0000)]
Where the alloca'd space actually lives in ram is undefined, and attempting to
pin it down is undefined behaviour.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151710
91177308-0d34-0410-b5e6-
96231b3b80d8
David Meyer [Wed, 29 Feb 2012 02:11:55 +0000 (02:11 +0000)]
[Object] Add symbol attribute flags: ST_ThreadLocal, ST_Common, and ST_Undefined. Implement these completely for ELF.
Rename ST_External to ST_Unknown, and slightly change its semantics. It now only indicates that the symbol's type
is unknown, not that the symbol is undefined. (For that, use ST_Undefined).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151696
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Wed, 29 Feb 2012 01:53:13 +0000 (01:53 +0000)]
Testcase for r151691.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151694
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Wed, 29 Feb 2012 01:46:50 +0000 (01:46 +0000)]
Restrict this transformation to equality conditions.
This transformation is not correct for not-equal conditions:
(trunc x) != C1 & (and x, CA) != C2 -> (and x, CA|CMAX) != C1|C2
Let
C1 == 0
C2 == 0
CA == 0xFF0000
CMAX == 0xFF
and truncating to i8.
The original truth table:
x | A: trunc x != 0 | B: x & 0xFF0000 != 0 | A & B != 0
--------------------------------------------------------------
0x00000 | 0 | 0 | 0
0x00001 | 1 | 0 | 0
0x10000 | 0 | 1 | 0
0x10001 | 1 | 1 | 1
The truth table of the replacement:
x | x & 0xFF00FF != 0
----------------------------
0x00000 | 0
0x00001 | 1
0x10000 | 1
0x10001 | 1
So they are different.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151691
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 29 Feb 2012 01:40:37 +0000 (01:40 +0000)]
Add an analyzeVirtReg() function.
This function does more or less the same as
MI::readsWritesVirtualRegister(), but it supports bundles as well.
It also determines if any constraint requires reading and writing
operands to use the same register. Most clients want to know.
Use the more modern MO.readsReg() instead of trying to sort out undefs
and partial redefines. Stop supporting the extra full <imp-def> operand
as an alternative to <def,undef> sub-register defines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151690
91177308-0d34-0410-b5e6-
96231b3b80d8
Derek Schuff [Wed, 29 Feb 2012 01:09:06 +0000 (01:09 +0000)]
Make MemoryObject accessor members const again
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151687
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Wed, 29 Feb 2012 00:46:46 +0000 (00:46 +0000)]
Support/PathV2: Fix namespace qualifier in make_absolute(), for Win32.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151685
91177308-0d34-0410-b5e6-
96231b3b80d8