1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-21 20:12:56 +02:00
Commit Graph

89147 Commits

Author SHA1 Message Date
Michael Ilseman
27c81fc400 Preserve fast-math flags after reassociation and commutation. Update test cases
llvm-svn: 174571
2013-02-07 01:40:15 +00:00
Michael Ilseman
b0f8194db9 whitespace
llvm-svn: 174569
2013-02-07 01:27:13 +00:00
Michael J. Spencer
053ee2fc5c [Object][ELF] Do the actual devirtualization.
llvm-svn: 174568
2013-02-07 01:24:22 +00:00
Michael J. Spencer
552d05feae [Object][ELF] Devirtualize and simplify dynamic table iteration.
llvm-svn: 174566
2013-02-07 01:17:23 +00:00
Owen Anderson
09e506b2be Fix CMake detection of various cmath functions, and XFAIL the test on platforms that are known to be missing them.
llvm-svn: 174564
2013-02-07 00:54:05 +00:00
Owen Anderson
4d8f6634c5 Conditionalize constant folding of math intrinsics on the availability of an implementation on the host. This is a little bit unfortunate, but until someone decides to implement a full libm for APFloat, we don't have a better way to get this functionality.
llvm-svn: 174561
2013-02-07 00:21:34 +00:00
Owen Anderson
93b6c99e55 Signficantly generalize our ability to constant fold floating point intrinsics, including ones on half types.
llvm-svn: 174555
2013-02-06 22:43:31 +00:00
Eli Bendersky
5d729faceb Fix typo
llvm-svn: 174553
2013-02-06 22:34:46 +00:00
Michael J. Spencer
c795dd0ca7 [Support][ErrorOr] Add support for convertable types.
Thanks to Andrew, David, and Aaron for helping fix this.

llvm-svn: 174552
2013-02-06 22:28:53 +00:00
Eli Bendersky
f78757381b Add a comment to the test that points to the source from which the input object
file was generated.

llvm-svn: 174551
2013-02-06 22:17:40 +00:00
Joe Abbey
6de61a11af Code Custodian (trivial whitespace cleanup)
llvm-svn: 174550
2013-02-06 22:14:06 +00:00
Eric Christopher
19e3875192 Clean up multiple skeleton compile units if we have multiple compile
units coming in.

llvm-svn: 174548
2013-02-06 21:53:56 +00:00
Eric Christopher
9d072d2846 Remove some dead code.
llvm-svn: 174547
2013-02-06 21:53:53 +00:00
Akira Hatanaka
9a19eb22dc [mips] Make NOP a pseudo instruction and expand it to "sll $zero, $zero, 0".
llvm-svn: 174546
2013-02-06 21:50:15 +00:00
Eli Bendersky
294e89257b Add a test for checking the current .debug_frame dumping capability.
The test is a binary placed in test/DebugInfo/Inputs, with a source C
file used for reference/reproducing. The source's first line is a clang
build command for reproducing the binary.

llvm-svn: 174543
2013-02-06 20:55:06 +00:00
Eli Bendersky
1854305220 This is a follow-up on r174446, now taking Atom processors into
account. Atoms use LEA for updating SP in prologs/epilogs, and the
exact LEA opcode depends on the data model.

Also reapplying the test case which was added and then reverted
(because of Atom failures), this time specifying explicitly the CPU in
addition to the triple. The test case now checks all variations (data
mode, cpu Atom vs. Core).

llvm-svn: 174542
2013-02-06 20:43:57 +00:00
Guy Benyei
e5a7b5054c Canonicalize line endings to Linux style also when the --strict-whitespace flag is in use. This flag is supposed to affect horizontal whitespaces only.
llvm-svn: 174541
2013-02-06 20:40:38 +00:00
Bill Wendling
87925691e8 Improve comment.
llvm-svn: 174536
2013-02-06 20:05:44 +00:00
Bill Schmidt
93695a5520 PPC calling convention cleanup.
Most of PPCCallingConv.td is used only by the 32-bit SVR4 ABI.  Rename
things to clarify this.  Also delete some code that's been commented out
for a long time.

llvm-svn: 174526
2013-02-06 17:33:58 +00:00
Tom Stellard
386be85491 R600: Support for indirect addressing v4
Only implemented for R600 so far.  SI is missing implementations of a
few callbacks used by the Indirect Addressing pass and needs code to
handle frame indices.

At the moment R600 only supports array sizes of 16 dwords or less.
Register packing of vector types is currently disabled, which means that a
vec4 is stored in T0_X, T1_X, T2_X, T3_X, rather than T0_XYZW. In order
to correctly pack registers in all cases, we will need to implement an
analysis pass for R600 that determines the correct vector width for each
array.

v2:
  - Add support for i8 zext load from stack.
  - Coding style fixes

v3:
  - Don't reserve registers for indirect addressing when it isn't
    being used.
  - Fix bug caused by LLVM limiting the number of SubRegIndex
    declarations.

v4:
  - Fix 64-bit defines

llvm-svn: 174525
2013-02-06 17:32:29 +00:00
Tim Northover
a6ee94525f Implement external weak (ELF) symbols on AArch64
Weakly defined symbols should evaluate to 0 if they're undefined at
link-time. This is impossible to do with the usual address generation
patterns, so we should use a literal pool entry to materlialise the
address.

llvm-svn: 174518
2013-02-06 16:43:33 +00:00
Eli Bendersky
31f7dcd2e4 Add some comments to new frame entries
llvm-svn: 174515
2013-02-06 16:20:31 +00:00
Tim Northover
4daffb618d Add AArch64 CRC32 instructions
These instructions are a late addition to the architecture, and may
yet end up behind an optional attribute, but for now they're available
at all times.

llvm-svn: 174496
2013-02-06 09:13:13 +00:00
Tim Northover
45bdd458aa Add icache prefetch operations to AArch64
This adds hints to the various "prfm" instructions so that they can
affect the instruction cache as well as the data cache.

llvm-svn: 174495
2013-02-06 09:04:56 +00:00
Bill Wendling
2da9899378 Initial submission for the attribute group feature.
Attribute groups are of the form:

  #0 = attributes { noinline "no-sse" "cpu"="cortex-a8" alignstack=4 }

Target-dependent attributes are represented as strings. Attributes can have
optional values associated with them. E.g., the "cpu" attribute has the value
"cortex-a8".

Target-independent attributes are listed as enums inside the attribute classes.

Multiple attribute groups can be referenced by the same object. In that case,
the attributes are merged together.

llvm-svn: 174493
2013-02-06 06:52:58 +00:00
Craig Topper
4ba9d33da9 Remove extra blank line between closing curly brace and 'else'
llvm-svn: 174492
2013-02-06 06:50:38 +00:00
Craig Topper
d4a8a10013 Remove unused private field to suppress a build warning.
llvm-svn: 174491
2013-02-06 06:48:10 +00:00
Bill Wendling
0835c19d36 Alphabetize the function attributes.
llvm-svn: 174490
2013-02-06 06:22:58 +00:00
Jim Grosbach
258f68e8a5 ARM: Use MCTargetAsmParser::validateTargetOperandClass().
Use the validateTargetOperandClass() hook to match literal '#0' operands in
InstAlias definitions. Previously this required per-instruction C++ munging of the
operand list, but not is handled as a natural part of the matcher. Much better.

No additional tests are required, as the pre-existing tests for these instructions
exercise the new behaviour as being functionally equivalent to the old.

llvm-svn: 174488
2013-02-06 06:00:11 +00:00
Jim Grosbach
6f7278da17 Allow targets to add custom asm operand matching logic.
For example, ARM has several instructions with a literal '#0' immediate in the syntax
that's not represented as an actual operand. The asm matcher is expected a token
operand, but the parser will have created an immediate operand. This is currently
handled by dedicated per-instruction C++ munging of the ParsedAsmOperand list, but
will be better handled by this hook.

llvm-svn: 174487
2013-02-06 06:00:06 +00:00
Eli Bendersky
935df72ee3 Failing builds because a private class member is not being used after
initialization is one of the reasons I consider -werror to be shoddy.

llvm-svn: 174485
2013-02-06 05:37:46 +00:00
Eli Bendersky
6ac7b7b5cf Remove this test in the meantime, since it won't pass on Atom. Atom uses lea
to move the stack pointer in prologs/epilogs. I will fix the test and add it
back later.

llvm-svn: 174484
2013-02-06 03:15:00 +00:00
Eli Bendersky
bd977c5c69 Add virtual desctructor to FrameEntry to avoid error on delete-non-virtual-dtor
llvm-svn: 174483
2013-02-06 03:08:02 +00:00
Evan Cheng
5bd979b50b Tweak check to avoid integer overflow (for insanely large alignments)
llvm-svn: 174482
2013-02-06 02:06:33 +00:00
Bill Wendling
f45bd45085 Add a 'StringRef' version of hasAttribute.
Fix the 'operator==' and 'hasAttributes' queries to take into account
target-dependent attributes.

llvm-svn: 174481
2013-02-06 01:33:42 +00:00
Bill Wendling
555fe288d9 Add methods to merge an AttrBuilder into another builder.
This is useful when parsing an object that references multiple attribute groups.

N.B. If both builders have alignments specified, then they should match!

llvm-svn: 174480
2013-02-06 01:16:00 +00:00
Manman Ren
6edff4edb0 Attempt to recover gdb bot after r174445.
Failure: undefined symbol 'Lline_table_start0'.
Root-cause: we use a symbol subtraction to calculate at_stmt_list, but
the line table entries are not dumped in the assembly.
Fix: use zero instead of a symbol subtraction for Compile Unit 0.

llvm-svn: 174479
2013-02-06 00:59:41 +00:00
Eli Bendersky
7eeb3473bf Fix some formatting & add comments, following Eric's review
llvm-svn: 174473
2013-02-06 00:20:38 +00:00
Bill Wendling
dffa0dbdc3 Add the target-dependent (string) attributes from the AttrBuilder to the AttributeSet.
llvm-svn: 174467
2013-02-05 23:48:36 +00:00
Eli Bendersky
b5ee34671c Add missing file to CMake list
llvm-svn: 174465
2013-02-05 23:37:18 +00:00
Eli Bendersky
6f676af6f1 Test for r174446
llvm-svn: 174464
2013-02-05 23:31:48 +00:00
Eli Bendersky
0802d28eff Initial support for DWARF CFI parsing and dumping in LLVM
llvm-svn: 174463
2013-02-05 23:30:58 +00:00
Eli Bendersky
3576be2063 Be consistent about the field name - AddressSize, not PointerSize. Add
a setter and fix some comments.

llvm-svn: 174462
2013-02-05 23:26:02 +00:00
Bob Wilson
0c5819e04c Preprocess Apple llvmCore headers to reflect NDEBUG setting. <rdar://12568983>
If an Apple llvmCore build is done without assertions, and a client uses
the llvmCore headers with assertions enabled, or vice versa, then things will
break because some of the structure sizes in the API are different.  Use the
unifdef tool to make the headers unconditionally match the way the llvmCore
libraries were built.

llvm-svn: 174460
2013-02-05 22:59:42 +00:00
Andrew Trick
2e729a5aff Revert "[Support][ErrorOr] Add support for convertable types."
This reverts commit a33e1fafac7fedb1b080ef07ddf9ad6ddff3a830.

This unit test crashes on Darwon. It needs to be temporarily reverted
to unblock the test infrastructure.

llvm-svn: 174458
2013-02-05 22:50:20 +00:00
Bill Wendling
baf40a92db Convert to storing the attribute's internals as enums, integers, and strings.
The stuff we're handing are all enums (Attribute::AttrKind), integers and
strings. Don't convert them to Constants, which is an unnecessary step here. The
rest of the changes are mostly mechanical.

llvm-svn: 174456
2013-02-05 22:37:24 +00:00
Eli Bendersky
63b704faa4 Make sure the correct opcodes are used to SUB and ADD the stack
pointer in function prologs/epilogs. The opcodes should depend on the
data model (LP64 vs. ILP32) rather than the architecture bit-ness.

llvm-svn: 174446
2013-02-05 21:53:29 +00:00
Manman Ren
b9bd895a06 Dwarf: support for LTO where a single object file can have multiple line tables
We generate one line table for each compilation unit in the object file.
Reviewed by Eric and Kevin.

rdar://problem/13067005

llvm-svn: 174445
2013-02-05 21:52:47 +00:00
Akira Hatanaka
df9480569b [mips] Do not use function CC_MipsN_VarArg unless the function being analyzed
is a vararg function.

The original code was examining flag OutputArg::IsFixed to determine whether
CC_MipsN_VarArg or CC_MipsN should be called. This is not correct, since this
flag is often set to false when the function being analyzed is a non-variadic
function.

llvm-svn: 174442
2013-02-05 21:18:11 +00:00
Michael J. Spencer
09867f6b42 [objdump,readobj] Document the purpose and goals of each tool.
llvm-svn: 174439
2013-02-05 20:27:22 +00:00