Commit Graph

98121 Commits

Author SHA1 Message Date
Eric Christopher
2d6d0fc3f2 Remove special handling for DW_AT_ranges support by constructing the
values with the correct behavior.

llvm-svn: 196441
2013-12-05 00:36:17 +00:00
Logan Chien
558333e1e1 [mc] Fix ELF st_other flag.
ELF_Other_Weakref and ELF_Other_ThumbFunc seems to be LLVM
internal ELF symbol flags.  These should not be emitted to
object file.

This commit defines ELF_STO_Shift for the target-defined
flags for st_other, and increase the value of
ELF_Other_Shift to 16.

llvm-svn: 196440
2013-12-05 00:34:11 +00:00
Michael Ilseman
fb9a99d2cf Use present fast-math flags when applicable in CreateBinOp
We were previously not adding fast-math flags through CreateBinOp()
when it happened to be making a floating point binary operator. This
patch updates it to do so similarly to directly calling CreateF*().

llvm-svn: 196438
2013-12-05 00:32:09 +00:00
Eric Christopher
d051658b5c Fix comment.
llvm-svn: 196437
2013-12-05 00:13:15 +00:00
Cameron McInally
675f9245aa Add AVX512 patterns for v16i32 broadcast and v2i64 zero extend load.
Patch by Aleksey Bader.

llvm-svn: 196435
2013-12-05 00:11:25 +00:00
Eric Christopher
4670039c5b Fix typo.
llvm-svn: 196434
2013-12-04 23:55:09 +00:00
David Blaikie
561838d222 DwarfUnit: Correct comment by generalizing over all units, not just compilation units.
Code review feedback on r196394 by Paul Robinson.

llvm-svn: 196433
2013-12-04 23:39:02 +00:00
Kevin Enderby
218f72b95b Fix a bug in darwin's 32-bit X86 handling of evaluating fixups.
Where it would use a scattered relocation entry but falls back to a
normal relocation entry because the FixupOffset is more than 24-bits.

The bug is in the X86MachObjectWriter::RecordScatteredRelocation() where
it changes reference parameter FixedValue but then returns false to indicate
it did not create a scattered relocation entry.  The fix is simply to save the
original value of the parameter FixedValue at the start of the method and
restore it if we are returning false in that case.

rdar://15526046

llvm-svn: 196432
2013-12-04 23:36:24 +00:00
Eric Christopher
f46aa7d453 Update comment.
llvm-svn: 196431
2013-12-04 23:24:38 +00:00
Eric Christopher
a054e191ad Update comment.
llvm-svn: 196430
2013-12-04 23:24:28 +00:00
Eric Christopher
435de44e9d Remove incorrect comment and pointless cast.
llvm-svn: 196427
2013-12-04 23:05:21 +00:00
Eric Christopher
f8ead1f600 const on its own line is confusing.
llvm-svn: 196426
2013-12-04 22:54:45 +00:00
David Peixotto
b6710ff7c7 Add support for parsing ARM symbol variants on ELF targets
ARM symbol variants are written with parens instead of @ like this:

  .word __GLOBAL_I_a(target1)

This commit adds support for parsing these symbol variants in
expressions. We introduce a new flag to MCAsmInfo that indicates the
parser should use parens to parse the symbol variant. The expression
parser is modified to look for symbol variants using parens instead
of @ when the corresponding MCAsmInfo flag is true.

The MCAsmInfo parens flag is enabled only for ARM on ELF.

By adding this flag to MCAsmInfo, we are able to get rid of
redundant ARM-specific symbol variants and use the generic variants
instead (e.g. VK_GOT instead of VK_ARM_GOT). We use the new
UseParensForSymbolVariant attribute in MCAsmInfo to correctly print
the symbol variants for arm.

To achive this we need to keep a handle to the MCAsmInfo in the
MCSymbolRefExpr class that we can check when printing the symbol
variant.

Updated Tests:
  Changed case of symbol variant to match the generic kind.
  test/CodeGen/ARM/tls-models.ll
  test/CodeGen/ARM/tls1.ll
  test/CodeGen/ARM/tls2.ll
  test/CodeGen/Thumb2/tls1.ll
  test/CodeGen/Thumb2/tls2.ll

PR18080

llvm-svn: 196424
2013-12-04 22:43:20 +00:00
Eric Christopher
1d7adf97f3 Simplify check.
llvm-svn: 196422
2013-12-04 22:29:02 +00:00
Eric Christopher
b46f56f8d6 Reformat slightly.
llvm-svn: 196421
2013-12-04 22:26:43 +00:00
Eric Christopher
5d21b8449e Make RangeSpanList take a symbol for the beginning of the range
rather than magically making the names match.

llvm-svn: 196419
2013-12-04 22:04:50 +00:00
Eric Christopher
bc88095166 Add a FIXME for making the symbol emission functions const.
llvm-svn: 196418
2013-12-04 22:04:46 +00:00
David Blaikie
cd548dcc12 DwarfDebug: Unconditionalize trivial asm comments
While we still have a few (~4) non-trivial comments with string
concatenation, etc that should remain conditionalized, these trivial
literal comments can be simplified.

llvm-svn: 196416
2013-12-04 21:51:05 +00:00
David Blaikie
7ce191027c DwarfDebug: Reduce code duplication for sec offset emission
llvm-svn: 196414
2013-12-04 21:31:26 +00:00
Eric Christopher
522b2d4cf3 Couple of small logical cleanups to use !empty rather than other
checks. No functional change.

llvm-svn: 196412
2013-12-04 21:20:15 +00:00
Yuchen Wu
9559c6af5d llvm-cov: Replace size() with empty() in bool check.
llvm-svn: 196400
2013-12-04 19:18:23 +00:00
Eric Christopher
93cee79b17 Use move and stack allocation for RangeSpanLists. As a result make
a few things more const as well because we're now using const
references to refer to iterators.

llvm-svn: 196398
2013-12-04 19:06:58 +00:00
David Blaikie
399690c42e DebugInfo: Improve test to use llvm-dwarfdump
llvm-svn: 196396
2013-12-04 18:40:29 +00:00
David Blaikie
1c49e697e0 Test fix for r196394
llvm-svn: 196395
2013-12-04 18:34:28 +00:00
David Blaikie
f92edbcbd0 DebugInfo: Remove unused start/end labels for the debug_abbrevs section
Since we always emit only one abbrevation section (shared by all the
compilation units in this module) there's no need for a separate label
at the start of each one (and we weren't using the CU ID anyway, so
there really was only one label). Use the section label instead and drop
the wholely unused debug_abbrev_end label.

llvm-svn: 196394
2013-12-04 18:12:28 +00:00
Cameron McInally
97a9fa294d Fix assembly syntax for AVX512 vector blend instructions.
llvm-svn: 196393
2013-12-04 18:05:36 +00:00
David Blaikie
1ca7b250be DebugInfo: Avoid recreating matching labels in disparate places.
Instead, reuse the same MCSymbol - this should make the code easier to
follow by avoiding hard to trace dependencies between different bits of
code.

llvm-svn: 196392
2013-12-04 17:55:41 +00:00
Michael Liao
7de99b9277 [X86] Check YMM31/ZMM31 as well
- No test case as there's no calling convention preserve YMM31/ZMM31 only

llvm-svn: 196391
2013-12-04 17:44:22 +00:00
Chad Rosier
59cd9a3090 Update the UseFusedMAC definition to directly specify its dependence on having
VFP4.
Patch by Daniel Stewart!

llvm-svn: 196390
2013-12-04 17:16:36 +00:00
Cameron McInally
9c9a78a238 Suppress '(x < y) ? a : 0 -> (x < y) & a' transform on X86 architectures with dedicated mask registers.
Patch by Aleksey Bader.

llvm-svn: 196386
2013-12-04 14:52:33 +00:00
NAKAMURA Takumi
bfddedc1df [CMake] add_lit_target: Let lit.site.cfg free from "--param build_mode" on single configuration builds, like autoconf build.
llvm-svn: 196377
2013-12-04 11:15:17 +00:00
Bill Wendling
bc381d3024 Update email address.
llvm-svn: 196369
2013-12-04 09:42:49 +00:00
Daniel Jasper
ca41e63412 Un-revert r196358: "llvm-cov: Added support for function checksums."
And add the proper fix.

llvm-svn: 196367
2013-12-04 08:57:17 +00:00
Daniel Jasper
a7dd8af910 Revert r196358: "llvm-cov: Added support for function checksums."
This currently breaks clang/test/CodeGen/code-coverage.c. The root cause
is that the newly introduced access to Funcs[j] is out of bounds.

llvm-svn: 196365
2013-12-04 08:23:33 +00:00
Kevin Qin
f5b717aa75 [AArch64 Neon] Add ACLE intrinsic vceqz_f64.
llvm-svn: 196362
2013-12-04 08:02:34 +00:00
Kevin Qin
f93a2e8673 [AArch64 NEON] Add missing compare intrinsics.
llvm-svn: 196360
2013-12-04 07:53:28 +00:00
Yuchen Wu
b1a23c9951 llvm-cov: Added support for function checksums.
The function checksums are hashed from the concatenation of the function
name and line number.

llvm-svn: 196358
2013-12-04 06:00:17 +00:00
Yuchen Wu
37b184215d llvm-cov: Added checks for ident, checksum, name.
Added additional checks for the Identifier, CfgChecksum and Name for
each GCOVFunction. Also added function names in error messages.

llvm-svn: 196356
2013-12-04 05:42:28 +00:00
Yuchen Wu
01a610d829 llvm-cov: Capitalized GCNO and GCDA for consistency.
llvm-svn: 196354
2013-12-04 05:07:36 +00:00
Yuchen Wu
0b08e75d64 llvm-cov: Split GCOVFile's read into GCNO and GCDA.
This splits the file-scope read() function into readGCNO() and
readGCDA(). Also broke file format read into functions that first read
the file type, then check the version.

llvm-svn: 196353
2013-12-04 04:49:23 +00:00
Rafael Espindola
cd5c36b48b Fix a funny typo.
Thanks for j`ey and Sean Silva for noticing it.

llvm-svn: 196344
2013-12-04 02:26:54 +00:00
Sean Silva
c765e40e3a [LangRef] Remove (non-normative) paragraph that doesn't make sense.
It appears to be referring to nonexistent entities. This must be a
carry-over from an older version of the document.

Patch by Mikael Lyngvig!

llvm-svn: 196342
2013-12-04 02:19:25 +00:00
Rafael Espindola
f394b917ee Produce deterministic coff files.
llvm-svn: 196341
2013-12-04 02:02:55 +00:00
Rafael Espindola
94d08ca0e8 Add -mcpu=core2 to all llc invocations in this test.
Should fix the atom buildbot.

llvm-svn: 196340
2013-12-04 01:25:24 +00:00
Juergen Ributzka
8504aa2736 [Stackmap] Specify the triple and cpu to fix the unit test.
llvm-svn: 196339
2013-12-04 01:02:37 +00:00
Juergen Ributzka
f7f5626671 [Stackmap] Emit multi-byte nops for X86.
llvm-svn: 196334
2013-12-04 00:39:08 +00:00
Reed Kotler
45b4f281f2 final patch for very long conditional branches for mips16 constant islands.
this completes the basic port of ARM constant islands to Mips16.
More testing, code review, cleanup is in order but basically everything
seems to be working. A bug in gas is preventing some of the runtime
testing but I hope to resolve this soon.

llvm-svn: 196331
2013-12-03 23:42:51 +00:00
NAKAMURA Takumi
63cd9136f2 check-llvm: Ask llvm-config about assertion mode, instead of llc.
Add --assertion-mode to llvm-config. It emits ON or OFF according to NDEBUG.

llvm-svn: 196329
2013-12-03 23:22:25 +00:00
NAKAMURA Takumi
480ca6a46d utils/llvm-lit: Generate "llvm-lit.py" rather than "llvm-lit" on Win32 hosts to let llvm-lit.py --use-processes work.
llvm-lit needs suffix.py for multiprocess to find a main module.

llvm-svn: 196328
2013-12-03 23:22:18 +00:00
Eric Christopher
907181fbbb Update comment grammar and contents.
llvm-svn: 196323
2013-12-03 22:05:55 +00:00