Aaron Ballman [Sat, 10 Mar 2012 23:03:01 +0000 (23:03 +0000)]
Fixing a compile warning triggered in MSVC about constant truncation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152518
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Sat, 10 Mar 2012 23:02:54 +0000 (23:02 +0000)]
Make StringRef::getAsInteger work with all integer types. Before this change
it would fail with {,u}int64_t on x86-64 Linux.
This also removes code duplication.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152517
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 10 Mar 2012 22:41:06 +0000 (22:41 +0000)]
Make helper static, so it can be inlined into its sole caller.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152515
91177308-0d34-0410-b5e6-
96231b3b80d8
Gregory Szorc [Sat, 10 Mar 2012 21:44:03 +0000 (21:44 +0000)]
Revert "[llvm.py] Implement interface to enhanced disassembler"
Chris Lattner says the edis interface is going away. It doesn't make
sense to land something that will go away in the near future.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152508
91177308-0d34-0410-b5e6-
96231b3b80d8
Kay Tiong Khoo [Sat, 10 Mar 2012 21:29:49 +0000 (21:29 +0000)]
*fix typo in comment; test of commit access
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152507
91177308-0d34-0410-b5e6-
96231b3b80d8
Gregory Szorc [Sat, 10 Mar 2012 21:05:05 +0000 (21:05 +0000)]
[llvm.py] Implement interface to enhanced disassembler
This requires a C++ change to EDDisassembler's ctor to function properly
(the llvm::InitializeAll* functions aren't being called currently and
there is no way to call them from Python).
Code is partially tested and works well enough for initial commit. There
are probably many small bugs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152506
91177308-0d34-0410-b5e6-
96231b3b80d8
Gregory Szorc [Sat, 10 Mar 2012 21:01:14 +0000 (21:01 +0000)]
[llvm.py] Make LLVMObject.__del__ work if called during __init__
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152505
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Sat, 10 Mar 2012 19:03:51 +0000 (19:03 +0000)]
Move llc + target triple tests into X86
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152502
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 10 Mar 2012 18:20:55 +0000 (18:20 +0000)]
As Duncan pointed out, pointers tend not to be in floating point format...for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152499
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 10 Mar 2012 18:07:46 +0000 (18:07 +0000)]
Don't try to filecheck bitcode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152498
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 10 Mar 2012 17:56:03 +0000 (17:56 +0000)]
Make this transformation slightly less agressive and more correct.
The 'CmpInst::isFalseWhenEqual' function returns 'false' for values other than
simply equality. For instance, it returns 'false' for <= or >=. This isn't the
correct behavior for this transformation, which is checking for strict equality
and non-equality. It was causing the gcc.c-torture/execute/frame-address.c test
to fail because it would completely (and incorrectly) optimize a whole function
into a 'ret i32 0'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152497
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 10 Mar 2012 15:10:06 +0000 (15:10 +0000)]
C files in llvm still have to be C89 compliant, remove C++-style comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152495
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 10 Mar 2012 12:50:44 +0000 (12:50 +0000)]
Microoptimize getVRegDef. def_begin isn't free, don't compute it twice.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152492
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Sat, 10 Mar 2012 08:39:09 +0000 (08:39 +0000)]
Refactor some methods to look through bitcasts and GEPs on pointers into
a common collection of methods on Value, and share their implementation.
We had two variations in two different places already, and I need the
third variation for inline cost estimation.
Reviewed by Duncan Sands on IRC, but further comments here welcome.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152490
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 10 Mar 2012 07:37:27 +0000 (07:37 +0000)]
Fix disasm of iret, sysexit, and sysret when displayed with Intel syntax.
Patch by Kay Tiong Khoo!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152487
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 10 Mar 2012 07:11:55 +0000 (07:11 +0000)]
Implement a more intelligent way of spilling uses across an invoke boundary.
The old way of determine when and where to spill a value that was used inside of
a landing pad resulted in spilling that value everywhere and not just at the
invoke edge.
This algorithm determines which values are used within a landing pad. It then
spills those values before the invoke and reloads them before the uses. This
should prevent excessive spilling in many cases, e.g. inside of loops.
<rdar://problem/
10609139>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152486
91177308-0d34-0410-b5e6-
96231b3b80d8
Gregory Szorc [Sat, 10 Mar 2012 05:50:56 +0000 (05:50 +0000)]
[llvm.py] Define enumerations from Core.h; add OpCode class
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152483
91177308-0d34-0410-b5e6-
96231b3b80d8
Gregory Szorc [Sat, 10 Mar 2012 04:41:24 +0000 (04:41 +0000)]
[llvm.py] Implement interface to object files
It is now possible to load object files and scan over sections, symbols,
and relocations! Includes test code with partial coverage.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152482
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 10 Mar 2012 02:04:38 +0000 (02:04 +0000)]
Fix uses of the C99 PRI format macros not to conflict with C++11 UDLs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152474
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 10 Mar 2012 00:44:11 +0000 (00:44 +0000)]
Report the defining instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152460
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 10 Mar 2012 00:36:06 +0000 (00:36 +0000)]
Add SSA verification to MachineVerifier.
Somehow we never verified SSA dominance before.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152458
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 10 Mar 2012 00:36:04 +0000 (00:36 +0000)]
Use SmallPtrSet instead of DenseSet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152457
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 10 Mar 2012 00:23:58 +0000 (00:23 +0000)]
Give dagcombiner's worklist some inline capacity.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152454
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Sat, 10 Mar 2012 00:03:50 +0000 (00:03 +0000)]
Do not custom lower i64 nodes if i64 is not a legal type. Move lines that set
operation action of nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152452
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 9 Mar 2012 23:46:03 +0000 (23:46 +0000)]
Lower SETCC nodes during legalization. Previously, it was lowered in DAG combine pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152450
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 9 Mar 2012 23:41:44 +0000 (23:41 +0000)]
Assert on SSA errors in LiveVariables.
All uses of a virtual register must be dominated by its def.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152449
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 9 Mar 2012 23:28:30 +0000 (23:28 +0000)]
Remove unused header files.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152447
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 9 Mar 2012 22:34:56 +0000 (22:34 +0000)]
misched: handle scheduler that insert instructions at empty region boundaries.
And add comments, since this is obviously confusing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152445
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Fri, 9 Mar 2012 22:24:10 +0000 (22:24 +0000)]
Add the missing call to Error when a bad X86 scale expression is parsed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152443
91177308-0d34-0410-b5e6-
96231b3b80d8
David Meyer [Fri, 9 Mar 2012 20:59:52 +0000 (20:59 +0000)]
Support reading GNU symbol versions in ELFObjectFile
* Add enums and structures for GNU version information.
* Implement extraction of that information on a per-symbol basis (ELFObjectFile::getSymbolVersion).
* Implement a generic interface, GetELFSymbolVersion(), for getting the symbol version from the ObjectFile (hides the templating).
* Have llvm-readobj print out the version, when available.
* Add a test for the new feature: readobj-elf-versioning.test
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152436
91177308-0d34-0410-b5e6-
96231b3b80d8
David Meyer [Fri, 9 Mar 2012 20:41:57 +0000 (20:41 +0000)]
[Object]
Make Binary::TypeID more granular, to distinguish between ELF 32/64 little/big
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152435
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Fri, 9 Mar 2012 19:21:15 +0000 (19:21 +0000)]
Add statistics on removed switch cases, and fix the phi statistic
to count the number of phis changed, not the number visited.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152425
91177308-0d34-0410-b5e6-
96231b3b80d8
Gregory Szorc [Fri, 9 Mar 2012 18:56:33 +0000 (18:56 +0000)]
[llvm.py] Make ObjectFile destructor work
Previous code had a double free in MemoryBuffer. The tests now pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152422
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 9 Mar 2012 18:50:52 +0000 (18:50 +0000)]
When identifying exit nodes for the reverse-CFG reverse-post-order
traversal, consider nodes for which the only successors are backedges
which the traversal is ignoring to be exit nodes. This fixes a problem
where the bottom-up traversal was failing to visit split blocks along
split loop backedges. This fixes rdar://
10989035.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152421
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Fri, 9 Mar 2012 17:52:49 +0000 (17:52 +0000)]
Fix the x86 disassembler to at least print the lock prefix if it is the first
prefix. Added a FIXME to remind us this still does not work when it is not the
first prefix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152414
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Fri, 9 Mar 2012 16:30:33 +0000 (16:30 +0000)]
[Support] Drop verbose _ATTRIBUTE from LLVM_ATTRIBUTE_{READONLY,READNONE} macro
names.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152413
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Fri, 9 Mar 2012 14:52:44 +0000 (14:52 +0000)]
TableGen/CodeEmitterGen.cpp: Fix an expression of generating bitmask.
~0U might be i32 on 32-bit hosts, then (uint64_t)~0U might not be expected as (i64)0xFFFFFFFF_FFFFFFFF, but as (i64)0x00000000_FFFFFFFF.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152407
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Fri, 9 Mar 2012 14:52:38 +0000 (14:52 +0000)]
test/MC/X86/lit.local.cfg: Fix up to detect 'X86' in targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152406
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Fri, 9 Mar 2012 13:45:18 +0000 (13:45 +0000)]
Eliminate switch cases that can never match, for example removes all
negative switch cases if the branch condition is known to be positive.
Inspired by a recent improvement to GCC's VRP.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152405
91177308-0d34-0410-b5e6-
96231b3b80d8
Anton Korobeynikov [Fri, 9 Mar 2012 10:09:36 +0000 (10:09 +0000)]
Add support for r600 (AMD GPUs HD2XXX - HD6XXX) target triplet.
Patch by Tom Stellard!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152400
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Fri, 9 Mar 2012 09:23:50 +0000 (09:23 +0000)]
Factor out the analysis of addition and subtraction in ComputeMaskedBits. Reuse
it to analyze extractvalue(llvm.[us](add|sub).with.overflow.*) intrinsics!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152398
91177308-0d34-0410-b5e6-
96231b3b80d8
Gregory Szorc [Fri, 9 Mar 2012 09:07:35 +0000 (09:07 +0000)]
[llvm.py] Initial skeleton for Python LLVM bindings
This contains a semi-functional skeleton for the implementation of the
LLVM bindings for Python.
The API for the Object.h interface is roughly designed but not
implemented. MemoryBufferRef is implemented and actually appears to
work!
The ObjectFile unit test fails with a segmentation fault because the
LLVM library isn't being properly initialized. The build system doesn't
know about this code yet, so no alerts should fire.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152397
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 9 Mar 2012 08:02:51 +0000 (08:02 +0000)]
misched: handle scheduling region boundaries nicely.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152393
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 9 Mar 2012 07:45:21 +0000 (07:45 +0000)]
Use uint16_t to store opcodes in static tables in X86 backend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152391
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Charles [Fri, 9 Mar 2012 06:36:45 +0000 (06:36 +0000)]
Fix undefined behavior in the Mips backend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152390
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 9 Mar 2012 04:29:02 +0000 (04:29 +0000)]
misched interface: rename Begin/End to RegionBegin/RegionEnd since they are not private.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152382
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 9 Mar 2012 03:46:42 +0000 (03:46 +0000)]
misched comments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152374
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 9 Mar 2012 03:46:39 +0000 (03:46 +0000)]
revert 152356: verify misched changes using -misched=shuffle.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152373
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 9 Mar 2012 02:49:38 +0000 (02:49 +0000)]
Fix a silly restriction on the fast-path for hash_combine_range. This
caused several clients to select the slow variation. =[ This is extra
annoying because we don't have any realistic way of testing this -- by
design, these two functions *must* compute the same value.
Found while inspecting the output of some benchmarks I'm working on.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152369
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 9 Mar 2012 02:49:36 +0000 (02:49 +0000)]
Undo a previous restriction on the inline cost calculation which Nick
introduced. Specifically, there are cost reductions for all
constant-operand icmp instructions against an alloca, regardless of
whether the alloca will in fact be elligible for SROA. That means we
don't want to abort the icmp reduction computation when we abort the
SROA reduction computation. That in turn frees us from the need to keep
a separate worklist and defer the ICmp calculations.
Use this new-found freedom and some judicious function boundaries to
factor the innards of computing the cost factor of any given instruction
out of the loop over the instructions and into static helper functions.
This greatly simplifies the code, and hopefully makes it more clear what
is happening here.
Reviewed by Eric Christopher. There is some concern that we'd like to
ensure this doesn't get out of hand, and I plan to benchmark the effects
of this change over the next few days along with some further fixes to
the inline cost.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152368
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 9 Mar 2012 02:00:48 +0000 (02:00 +0000)]
Fix a regression from r147481.
Original commit message from r147481:
DAGCombine for transforming 128->256 casts into a vmovaps, rather
then a vxorps + vinsertf128 pair if the original vector came from a load.
Fix:
Unaligned loads need to generate a vmovups.
rdar://
10974078
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152366
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 9 Mar 2012 00:52:20 +0000 (00:52 +0000)]
misched: allow the default scheduler to be one chosen by the target.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152360
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 9 Mar 2012 00:52:17 +0000 (00:52 +0000)]
Added TargetPassConfig::enablePass
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152359
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 9 Mar 2012 00:24:29 +0000 (00:24 +0000)]
Cache MBB->begin. It's possible the scheduler / bundler may change MBB->begin().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152356
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 8 Mar 2012 22:15:23 +0000 (22:15 +0000)]
Silence unused function warning when graphviz is not available.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152346
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 8 Mar 2012 21:22:27 +0000 (21:22 +0000)]
Remove the no longer existent psp triple from a test.
The test fell back to the C backend, making it useless and it started to fail
on configurations that don't build the C backend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152342
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Thu, 8 Mar 2012 14:24:32 +0000 (14:24 +0000)]
Have llvm-mc --version print the list of registered targets like llc does.
Patch by jey.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152315
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Thu, 8 Mar 2012 09:32:21 +0000 (09:32 +0000)]
Revert commit 152300 (ddunbar) since it still seems to be breaking
buildbots. Original commit message:
[ADT] Change the trivial FoldingSetNodeID::Add* methods to be inline, reapplied
with a fix for the longstanding over-read of 32-bit pointer values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152304
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 8 Mar 2012 08:22:45 +0000 (08:22 +0000)]
Use uint16_t to store instruction implicit uses and defs. Reduces static data.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152301
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Thu, 8 Mar 2012 07:42:18 +0000 (07:42 +0000)]
[ADT] Change the trivial FoldingSetNodeID::Add* methods to be inline, reapplied
with a fix for the longstanding over-read of 32-bit pointer values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152300
91177308-0d34-0410-b5e6-
96231b3b80d8
Stepan Dyatkovskiy [Thu, 8 Mar 2012 07:06:20 +0000 (07:06 +0000)]
Taken into account Duncan's comments for r149481 dated by 2nd Feb 2012:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-
20120130/136146.html
Implemented CaseIterator and it solves almost all described issues: we don't need to mix operand/case/successor indexing anymore. Base iterator class is implemented as a template since it may be initialized either from "const SwitchInst*" or from "SwitchInst*".
ConstCaseIt is just a read-only iterator.
CaseIt is read-write iterator; it allows to change case successor and case value.
Usage of iterator allows totally remove resolveXXXX methods. All indexing convertions done automatically inside the iterator's getters.
Main way of iterator usage looks like this:
SwitchInst *SI = ... // intialize it somehow
for (SwitchInst::CaseIt i = SI->caseBegin(), e = SI->caseEnd(); i != e; ++i) {
BasicBlock *BB = i.getCaseSuccessor();
ConstantInt *V = i.getCaseValue();
// Do something.
}
If you want to convert case number to TerminatorInst successor index, just use getSuccessorIndex iterator's method.
If you want initialize iterator from TerminatorInst successor index, use CaseIt::fromSuccessorIndex(...) method.
There are also related changes in llvm-clients: klee and clang.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152297
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 8 Mar 2012 06:55:27 +0000 (06:55 +0000)]
Re-commit r152202 hopefully fixing the MSVC linker error.
Original commit message:
Use uint16_t to store InstrNameIndices in MCInstrInfo. Add asserts to protect all 16-bit string table offsets. Also make sure the string to offset table string is not larger than 65536 characters since larger string literals aren't portable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152296
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Thu, 8 Mar 2012 04:17:15 +0000 (04:17 +0000)]
Revert r152288, "[ADT] Change the trivial FoldingSetNodeID::Add* methods to be
inline.", which is breaking the bots in a way I don't understand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152295
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 8 Mar 2012 03:32:42 +0000 (03:32 +0000)]
Test case for r152280, r152285 and r152290.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152292
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 8 Mar 2012 03:26:37 +0000 (03:26 +0000)]
Invoke setTargetDAGCombine for SELECT.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152290
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Thu, 8 Mar 2012 02:52:00 +0000 (02:52 +0000)]
[ADT] Change the trivial FoldingSetNodeID::Add* methods to be inline.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152288
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 8 Mar 2012 02:14:24 +0000 (02:14 +0000)]
Swap the operands of a select node if the false (the second) operand is 0.
For example, this pattern
(select (setcc lhs, rhs, cc), true, 0)
is transformed to this one:
(select (setcc lhs, rhs, inverse(cc)), 0, true)
This enables MipsDAGToDAGISel::ReplaceUsesWithZeroReg (added in r152280) to
replace 0 with $zero.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152285
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 8 Mar 2012 02:04:19 +0000 (02:04 +0000)]
Rotate two of the functions used to count bonuses for the inline cost
analysis to be methods on the cost analysis's function info object
instead of the code metrics object. These really are just users of the
code metrics, they're building the information for the function's
analysis.
This is the first step of growing the amount of information we collect
about a function in order to cope with pair-wise simplifications due to
allocas.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152283
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 8 Mar 2012 01:59:33 +0000 (01:59 +0000)]
Set minimum function alignment to 3 if target is Mips64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152282
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 8 Mar 2012 01:51:59 +0000 (01:51 +0000)]
This patch eliminates redundant instructions that produce 0.
For example, the first instruction in the code below can be eliminated if the
use of $vr0 is replaced with $zero:
addiu $vr0, $zero, 0
add $vr2, $vr1, $vr0
add $vr2, $vr1, $zero
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152280
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 8 Mar 2012 01:41:12 +0000 (01:41 +0000)]
misched interface: Expose the MachineScheduler pass.
Allow targets to provide their own schedulers (subclass of
ScheduleDAGInstrs) to the misched pass. Select schedulers using
-misched=...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152278
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 8 Mar 2012 00:07:52 +0000 (00:07 +0000)]
ARM don't use MCRelaxAll, as it's not safe on ARM.
The ARM code generator makes aggressive assumptions about the encodings
being selected for branches which MCRelaxAll invalidates.
rdar://
11006355
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152268
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Callanan [Wed, 7 Mar 2012 23:05:25 +0000 (23:05 +0000)]
Improved support in RuntimeDyldMachO for generating
code that will be relocated into another memory space.
Now when relocations are resolved, the address of
the relocation in the host memory (where the JIT is)
is passed separately from the address that the
relocation will be at in the target memory (where
the code will run).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152264
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 23:01:09 +0000 (23:01 +0000)]
Cleanup VLIWPacketizer to use the updated ScheduleDAGInstrs interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152262
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 23:01:06 +0000 (23:01 +0000)]
misched prep: Expose the ScheduleDAGInstrs interface so targets may
implement their own MachineScheduler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152261
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 23:01:02 +0000 (23:01 +0000)]
misched prep: Remove LLVM_LIBRARY_VISIBILITY from ScheduleDAGInstrs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152260
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 23:00:59 +0000 (23:00 +0000)]
misched prep: Comment the ScheduleDAGInstrs interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152259
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 23:00:57 +0000 (23:00 +0000)]
misched prep: Cleanup ScheduleDAGInstrs interface.
ScheduleDAGInstrs will be the main interface for MI-level
schedulers. Make sure it's readable: one page of protected fields, one
page of public methids.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152258
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 23:00:54 +0000 (23:00 +0000)]
misched prep: remove extra "protected"
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152257
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 23:00:52 +0000 (23:00 +0000)]
misched prep: rename InsertPos to End.
ScheduleDAGInstrs knows nothing about how instructions will be moved or inserted.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152256
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 23:00:49 +0000 (23:00 +0000)]
misched preparation: rename core scheduler methods for consistency.
We had half the API with one convention, half with another. Now was a
good time to clean it up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152255
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 7 Mar 2012 22:48:42 +0000 (22:48 +0000)]
Copy the right amount of elements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152254
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 7 Mar 2012 22:33:21 +0000 (22:33 +0000)]
SmallPtrSet: Copy all the elements when swapping, not just numelements.
This fixes a build failure in webkit. Copying all elements shouldn't be
necessary, I'll look out for a better fix soon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152252
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 7 Mar 2012 20:59:26 +0000 (20:59 +0000)]
[fast-isel] ARMEmitCmp generates FMSTAT, which transfers the floating-point
condition flags to CPSR. This allows us to simplify SelectCmp.
Patch by Zonr Chang <zonr.xchg@gmail.com>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152243
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 7 Mar 2012 20:58:59 +0000 (20:58 +0000)]
Use llvm-mc instead of llc. Patch by Jack Carter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152242
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 7 Mar 2012 17:07:20 +0000 (17:07 +0000)]
configure: Don't require a perl interpreter to be present, LLVM's buildsystem doesn't depend on perl anymore.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152234
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 7 Mar 2012 17:06:40 +0000 (17:06 +0000)]
Revert r152202 as it's causing internal buildbot failures.
Original commit message:
Use uint16_t to store InstrNameIndices in MCInstrInfo. Add asserts to protect
all 16-bit string table offsets. Also make sure the string to offset table
string is not larger than 65536 characters since larger string literals aren't
portable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152233
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 7 Mar 2012 16:39:35 +0000 (16:39 +0000)]
Fix infinite loop in nested multiclasses.
Patch by Michael Liao!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152232
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 7 Mar 2012 12:27:35 +0000 (12:27 +0000)]
Try a completely different approach to this type trait to appease older
compilers. It seems that GCC 4.3 (and likely older) simply aren't going
to do SFINAE on non-type template parameters the way Clang and modern
GCCs do...
Now we detect the implicit conversion to an integer type, and then
blacklist classes, pointers, and floating point types. This seems to
work well enough, and I'm hopeful will return the bots to life.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152227
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 7 Mar 2012 11:31:11 +0000 (11:31 +0000)]
Attempt #2 at appeasing GCC 4.3. This compiler really doesn't like these
traits.
With this change, the pattern used here is *extremely* close to the
pattern used elsewhere in the file, so I'm hoping it survives the
build-bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152225
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 7 Mar 2012 10:13:40 +0000 (10:13 +0000)]
Try to clarify this comment some.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152221
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 7 Mar 2012 10:05:35 +0000 (10:05 +0000)]
Switch the is_integral_or_enum trait machinery to use an explicit
template argument and an *implicit* conversion from '0' to a null
pointer. For some bizarre reason, GCC 4.3.2 thinks that the cast to
'(T*)' is invalid inside of an enumerator's value... which it isn't but
whatever. ;] This pattern is used elsewhere in the type_traits header
and so hopefully will survive the wrath of the build bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152220
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 7 Mar 2012 09:54:06 +0000 (09:54 +0000)]
What's better than fixing and simplifying broken hash functions?
Deleting them because they aren't used. =D
Yell if you need these, I'm happy to instead replace them with nice uses
of the new infrastructure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152219
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 7 Mar 2012 09:39:46 +0000 (09:39 +0000)]
Remove another outbreak of customized (and completely broken) hashing.
This one is particularly annoying because the hashing algorithm is
highly specialized, with a strange "equivalence" definition that subsets
the fields involved.
Still, this looks at the exact same set of data as the old code, but
without bitwise or-ing over parts of it and other mixing badness. No
functionality changed here. I've left a substantial fixme about the fact
that there is a cleaner and more principled way to do this, but it
requires making the equality definition actual stable for particular
types...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152218
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Wed, 7 Mar 2012 09:32:32 +0000 (09:32 +0000)]
Add support to the hashing infrastructure for automatically hashing both
integral and enumeration types. This is accomplished with a bit of
template type trait magic. Thanks to Richard Smith for the core idea
here to detect viable types by detecting the set of types which can be
default constructed in a template parameter.
This is used (in conjunction with a system for detecting nullptr_t
should it exist) to provide an is_integral_or_enum type trait that
doesn't need a whitelist or direct compiler support.
With this, the hashing is extended to the more general facility. This
will be used in a subsequent commit to hashing more things, but I wanted
to make sure the type trait magic went through the build bots separately
in case other compilers don't like this formulation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152217
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Wed, 7 Mar 2012 08:49:42 +0000 (08:49 +0000)]
Where the BranchFolding pass removes a branch then adds another better branch,
the DebugLoc information can be maintained throughout by grabbing the DebugLoc
before the RemoveBranch and then passing the result to the InsertBranch.
Patch by Andrew Stanford-Jason!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152212
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 05:46:04 +0000 (05:46 +0000)]
Fix cmake
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152210
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 05:21:54 +0000 (05:21 +0000)]
comment
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152209
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 05:21:52 +0000 (05:21 +0000)]
misched preparation: clarify ScheduleDAG and ScheduleDAGInstrs roles.
ScheduleDAG is responsible for the DAG: SUnits and SDeps. It provides target hooks for latency computation.
ScheduleDAGInstrs extends ScheduleDAG and defines the current scheduling region in terms of MachineInstr iterators. It has access to the target's scheduling itinerary data. ScheduleDAGInstrs provides the logic for building the ScheduleDAG for the sequence of MachineInstrs in the current region. Target's can implement highly custom schedulers by extending this class.
ScheduleDAGPostRATDList provides the driver and diagnostics for current postRA scheduling. It maintains a current Sequence of scheduled machine instructions and logic for splicing them into the block. During scheduling, it uses the ScheduleHazardRecognizer provided by the target.
Specific changes:
- Removed driver code from ScheduleDAG. clearDAG is the only interface needed.
- Added enterRegion/exitRegion hooks to ScheduleDAGInstrs to delimit the scope of each scheduling region and associated DAG. They should be used to setup and cleanup any region-specific state in addition to the DAG itself. This is necessary because we reuse the same ScheduleDAG object for the entire function. The target may extend these hooks to do things at regions boundaries, like bundle terminators. The hooks are called even if we decide not to schedule the region. So all instructions in a block are "covered" by these calls.
- Added ScheduleDAGInstrs::begin()/end() public API.
- Moved Sequence into the driver layer, which is specific to the scheduling algorithm.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152208
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 7 Mar 2012 05:21:47 +0000 (05:21 +0000)]
ScheduleDAGInstrs comments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152207
91177308-0d34-0410-b5e6-
96231b3b80d8