Commit Graph

82519 Commits

Author SHA1 Message Date
Craig Topper
a34cbd451f Mark a static array as const.
llvm-svn: 157368
2012-05-24 04:11:15 +00:00
Craig Topper
2efe750a91 Mark a static table as const. Shrink opcode size in static tables to uint16_t. Simplify loop iterating over one of those tables. No functional change intended.
llvm-svn: 157367
2012-05-24 03:59:11 +00:00
Chad Rosier
b11b7f8b69 Tidy up naming for consistency and other cleanup. No functional change intended.
llvm-svn: 157358
2012-05-23 23:45:10 +00:00
Jakob Stoklund Olesen
c2da6e579e Add a test case for global live range splitting.
llvm-svn: 157357
2012-05-23 23:42:23 +00:00
Jakob Stoklund Olesen
2313afcd00 Add a last resort tryInstructionSplit() to RAGreedy.
Live ranges with a constrained register class may benefit from splitting
around individual uses. It allows the remaining live range to use a
larger register class where it may allocate. This is like spilling to a
different register class.

This is only attempted on constrained register classes.

<rdar://problem/11438902>

llvm-svn: 157354
2012-05-23 22:37:27 +00:00
Bill Wendling
ad5fae97e3 Forgot to reverse conditional.
llvm-svn: 157349
2012-05-23 22:12:50 +00:00
Bill Wendling
9235b4e8b1 Reduce indentation by early detection of 'continue'. No functionality change.
llvm-svn: 157348
2012-05-23 22:09:50 +00:00
Nicolas Geoffray
bf7155d836 llvm-ld does not exist anymore, use llvm-link instead.
llvm-svn: 157342
2012-05-23 20:34:19 +00:00
Kaelyn Uhrain
148e4135bf Fix typo in flag to opt, and also a CHECK-NEXT that doesn't follow a
CHECK. The latter error was hidden by the former, and the test harness
used by e.g. "make check" silently ignored that opt was printing an
error message about an unknown flag instead of running on the test file.

llvm-svn: 157341
2012-05-23 20:21:36 +00:00
Jakob Stoklund Olesen
ce44a7a9ae Correctly deal with identity copies in RegisterCoalescer.
Now that the coalescer keeps live intervals and machine code in sync at
all times, it needs to deal with identity copies differently.

When merging two virtual registers, all identity copies are removed
right away. This means that other identity copies must come from
somewhere else, and they are going to have a value number.

Deal with such copies by merging the value numbers before erasing the
copy instruction. Otherwise, we leave dangling value numbers in the live
interval.

This fixes PR12927.

llvm-svn: 157340
2012-05-23 20:21:06 +00:00
Tim Northover
62cf7e738b Fix how CMake appends -m32 to linker command-lines
llvm-svn: 157337
2012-05-23 18:42:02 +00:00
Chad Rosier
d9dca3aef3 [arm-fast-isel] Add support for non-global callee.
Patch by Jush Lu <jush.msn@gmail.com>.

llvm-svn: 157336
2012-05-23 18:38:57 +00:00
Nuno Lopes
36f35477a1 BoundsChecking: add a couple of simple tests and fix a bug in branch emition
llvm-svn: 157329
2012-05-23 16:24:52 +00:00
Nuno Lopes
68aaef6b21 revert r156383: removal of TYPE_CODE_FUNCTION_OLD
Apparently LLVM only stopped emitting this after LLVM 3.0

llvm-svn: 157325
2012-05-23 15:19:39 +00:00
Patrik Hägglund
51776725b8 Fix the inliner so that the optsize function attribute don't alter the
inline threshold if the global inline threshold is lower (as for -Oz).

Reviewed by Chandler Carruth and Bill Wendling.

llvm-svn: 157323
2012-05-23 13:42:57 +00:00
Patrik Hägglund
6f837093ba Fixed typo in r156905.
llvm-svn: 157320
2012-05-23 12:34:56 +00:00
Patrik Hägglund
822fe63c97 Small fix for the debug output from PBQP (PR12822).
llvm-svn: 157319
2012-05-23 12:12:58 +00:00
Evgeniy Stepanov
3519d90668 Use zero-based shadow by default on Android.
llvm-svn: 157317
2012-05-23 11:52:12 +00:00
Stepan Dyatkovskiy
49092667b1 PR1255(case ranges) related changes in Local Transformations.
llvm-svn: 157315
2012-05-23 08:18:26 +00:00
Craig Topper
49c52dde2b Tidy up spacing.
llvm-svn: 157313
2012-05-23 05:44:51 +00:00
Chris Lattner
9b30a6d7da small refinement to r157218 to save a tiny amount of table size in the common
case.

llvm-svn: 157312
2012-05-23 05:19:18 +00:00
Craig Topper
5161134794 Fix indentation of wrapped line for readability. No functional change.
llvm-svn: 157309
2012-05-23 03:59:53 +00:00
Eric Christopher
809a39bd54 Add support for C++11 enum classes in llvm.
Part of rdar://11496790

llvm-svn: 157303
2012-05-23 00:09:20 +00:00
Nuno Lopes
1a989d236d address some of John Criswell's comments
teach computeAllocSize about realloc, reallocf, and valloc

llvm-svn: 157298
2012-05-22 22:02:19 +00:00
NAKAMURA Takumi
ef6f526307 ARMDisassembler.cpp: Fix utf8 char in comments.
llvm-svn: 157292
2012-05-22 21:47:02 +00:00
Eric Christopher
7b3899918f Untabify and 80-col.
llvm-svn: 157274
2012-05-22 18:45:24 +00:00
Eric Christopher
a087fadf96 Formatting consistency.
llvm-svn: 157273
2012-05-22 18:45:18 +00:00
Jakob Stoklund Olesen
e4801d77e1 Also compute TopoSigs in synthetic register classes.
CodeGenRegisterClass has two constructors. Both need to compute the
TopoSigs BitVector.

llvm-svn: 157271
2012-05-22 18:20:28 +00:00
Nuno Lopes
73d40438e2 hopefully fix the CMake build. sorry for breakage
llvm-svn: 157264
2012-05-22 17:40:46 +00:00
Andrew Trick
bb5824688b LSR fix: add a missing phi check during IV hoisting.
Fixes PR12898: SCEVExpander crash.

llvm-svn: 157263
2012-05-22 17:39:59 +00:00
Nuno Lopes
114b8eaa9c add a new pass to instrument loads and stores for run-time bounds checking
move EmitGEPOffset from InstCombine to Transforms/Utils/Local.h

(a draft of this) patch reviewed by Andrew, thanks.

llvm-svn: 157261
2012-05-22 17:19:09 +00:00
Nuno Lopes
944814b41a revert my previous patches that introduced an additional parameter to the objectsize intrinsic.
After a lot of discussion, we realized it's not the best option for run-time bounds checking

llvm-svn: 157255
2012-05-22 15:25:31 +00:00
Jakob Stoklund Olesen
151b044b75 Only erase virtregs with no uses left.
Also make sure registers aren't erased twice if the dead def mentions
the register twice.

This fixes PR12911.

llvm-svn: 157254
2012-05-22 14:52:12 +00:00
Duncan Sands
39edcc75ac Fix PR12858, a crash due to GVN's PRE not fully removing an instruction from the
leader table.  That's because it wasn't expecting instructions to turn up as
leader for a value number that is not its own, but equality propagation could
create this situation.  One solution is to have the leader table use a WeakVH
but this slows down GVN by about 5%.  Instead just have equality propagation not
add instructions to the leader table, only constants and arguments.  In theory
this might cause GVN to run more (each time it changes something it runs again)
but it doesn't seem to occur enough to cause a slow down.

llvm-svn: 157251
2012-05-22 14:17:53 +00:00
Craig Topper
0fe3f1297b Fix constant used for pshufb mask when lowering v16i8 shuffles. Bug introduced in r157043. Fixes PR12908.
llvm-svn: 157236
2012-05-22 06:09:38 +00:00
Akira Hatanaka
2a600ff5f4 This patch adds a predicate to existing mips32 and mips64 so that those
instruction encodings can be excluded during mips16 processing.

This revision fixes the issue raised by Jim Grosbach.

bool hasStandardEncoding() const { return !inMips16Mode(); }

When micromips is added it will be

bool StandardEncoding() const { return !inMips16Mode()&&  !inMicroMipsMode(); }

No additional testing is needed other than to assure that there is no regression
from this patch.

Patch by Reed Kotler.

llvm-svn: 157234
2012-05-22 03:10:09 +00:00
Nuno Lopes
328f9efec2 fix the quotient returned by sdivrem() for the case when LHS is negative and RHS is positive
based on a patch by Preston Briggs, with some modifications

llvm-svn: 157231
2012-05-22 01:09:48 +00:00
Jim Grosbach
06eae84c89 FileCheck'ize test, and add a bit to test for r157221.
llvm-svn: 157222
2012-05-21 23:50:00 +00:00
Jim Grosbach
6e9ef9c931 ARM: .end_data_region mismatch in Thumb2.
32-bit offset jump tables just use real branch instructions and so aren't
marked as data regions. We were still emitting the .end_data_region
marker though, which assert()ed.

rdar://11499158

llvm-svn: 157221
2012-05-21 23:34:42 +00:00
Pete Cooper
869fc085f9 Added address space qualifier to intrinsic PointerType arguments.
llvm-svn: 157218
2012-05-21 23:21:28 +00:00
Owen Anderson
da2ddf18dc Fix use of an unitialized value in the LegalizeOps expansion for ISD::SUB. No in-tree targets exercise this path.
Patch by Micah Villmow.

llvm-svn: 157215
2012-05-21 22:39:20 +00:00
Jim Grosbach
85edf59851 Thumb2: RSB source register should be rGRP not GPRnopc.
t2RSB defined the operand correctly, but tRSBS didn't.

llvm-svn: 157200
2012-05-21 17:57:17 +00:00
Dan Gohman
992a69b57c Mark an unreachable region of code with llvm_unreachable.
llvm-svn: 157197
2012-05-21 17:41:28 +00:00
Chad Rosier
8d8d0f7479 Typo.
llvm-svn: 157195
2012-05-21 17:13:41 +00:00
Owen Anderson
6187e32939 Make it so that the MArch, MCPU, MAttrs passed to EngineBuilder are actually used.
Patch by Jose Fonseca.

llvm-svn: 157191
2012-05-21 16:57:17 +00:00
Patrik Hägglund
80e0e52143 test commit
llvm-svn: 157184
2012-05-21 13:18:06 +00:00
Stepan Dyatkovskiy
98d41ce6af PR1255 (case ranges: work with ConstantRangesSet instead of ConstantInt) related changes for Execution and Verifier.
llvm-svn: 157183
2012-05-21 10:44:40 +00:00
Craig Topper
57d54b1831 Allow 256-bit shuffles to still be split even if only half of the shuffle comes from two 128-bit pieces.
llvm-svn: 157175
2012-05-21 06:40:16 +00:00
Jakob Stoklund Olesen
c058c41261 Give a small negative bias to giant edge bundles.
This helps compile time when the greedy register allocator splits live
ranges in giant functions. Without the bias, we would try to grow
regions through the giant edge bundles, usually to find out that the
region became too big and expensive.

If a live range has many uses in blocks near the giant bundle, the small
negative bias doesn't make a big difference, and we still consider
regions including the giant edge bundle.

Giant edge bundles are usually connected to landing pads or indirect
branches.

llvm-svn: 157174
2012-05-21 03:11:23 +00:00
Jakob Stoklund Olesen
dadb6e60b9 Clear kill flags on the fly when joining intervals.
With physreg joining out of the way, it is easy to recognize the
instructions that need their kill flags cleared while testing for
interference.

This allows us to skip the final scan of all instructions for an 11%
speedup of the coalescer pass.

llvm-svn: 157169
2012-05-20 21:41:05 +00:00