Commits · 4dc73fa075db86bc6a07d755d972e9f8ad7336cc · Roger Ferrer / llvm-epi-0.8

Jan 18, 2011
- Add support for mips32 madd and msub instructions. Patch by Akira Hatanaka · 4dc73fa0
  Bruno Cardoso Lopes authored Jan 18, 2011
```
llvm-svn: 123760
```
  4dc73fa0
- add a note · a56c8279
  Chris Lattner authored Jan 18, 2011
```
llvm-svn: 123752
```
  a56c8279
- SPARC backend: Modified LowerCall and LowerFormalArguments so that they use... · c386f8a1
  Venkatraman Govindaraju authored Jan 18, 2011
```
SPARC backend: Modified LowerCall and LowerFormalArguments so that they use CallingConv assignments.

llvm-svn: 123749
```
  c386f8a1
- McARM: Use accessors where appropriate. · 62ea26fb
  Daniel Dunbar authored Jan 18, 2011
```
llvm-svn: 123746
```
  62ea26fb
- McARM: Fill in ASMOperand::dump() for memory operands. · bcd8eb0b
  Daniel Dunbar authored Jan 18, 2011
```
llvm-svn: 123745
```
  bcd8eb0b
- McARM: Make ARMOperand use a union where appropriate. · 510740ee
  Daniel Dunbar authored Jan 18, 2011
```
llvm-svn: 123744
```
  510740ee
- McARM: Unify ParseMemory() successfull return. · f5164f40
  Daniel Dunbar authored Jan 18, 2011
```
llvm-svn: 123740
```
  f5164f40
- McARM: Early exit on failure (NEFC). · 1d5e9549
  Daniel Dunbar authored Jan 18, 2011
```
llvm-svn: 123739
```
  1d5e9549
- McARM: Always keep an offset expression, if used (instead of assuming == 0 if... · 7ed45599
  Daniel Dunbar authored Jan 18, 2011
```
McARM: Always keep an offset expression, if used (instead of assuming == 0 if used but not present), and simplify logic.

Also, clean up various non-sensicalisms in isMemModeRegThumb() and isMemModeImmThumb().

llvm-svn: 123738
```
  7ed45599
- McARM: Add a variety of asserts on the sanity of memory operands. · 5d99420e
  Daniel Dunbar authored Jan 18, 2011
```
llvm-svn: 123737
```
  5d99420e
- McARM: Use a consistent marker for not-set OffsetRegNum. · d8da9e0f
  Daniel Dunbar authored Jan 18, 2011
```
llvm-svn: 123736
```
  d8da9e0f
- McARM: Start marking T2 address operands as such, for the benefit of the parser. · 66e91d4a
  Daniel Dunbar authored Jan 18, 2011
```
llvm-svn: 123722
```
  66e91d4a
- The stub routine that we're calling uses test and so clobbers · 542f8a52
  Eric Christopher authored Jan 18, 2011
```
the flags.

llvm-svn: 123712
```
  542f8a52
- minor change to rafael's recent patches: if something is · ea4e983d
  Chris Lattner authored Jan 18, 2011
```
constant but requires a unique address, we can still put it in a
readonly section, just not a mergable one.

llvm-svn: 123711
```
  ea4e983d
- Remove unused variables found by gcc-4.6's -Wunused-but-set-variable. · 249fcd44
  Jeffrey Yasskin authored Jan 18, 2011
```
llvm-svn: 123707
```
  249fcd44
Jan 17, 2011

Add a missing <cctype> include, from Joerg Sonnenberger! · 69e6206b
Douglas Gregor authored Jan 17, 2011
```
llvm-svn: 123670
```
69e6206b

Split up RotateShift itinerary in SPU. · 06c6d5cd

Kalle Raiskila authored Jan 17, 2011

'rotq*' and 'shlq*' instructions go to the odd pipeline,
wheras the inter-vector equivalents 'rot*', 'shl*' go 
to the even.

llvm-svn: 123622

06c6d5cd

Don't crash SPU BE with memory accesses with big alignmnet. · 7e7b4ac7
Kalle Raiskila authored Jan 17, 2011
```
llvm-svn: 123620
```
7e7b4ac7

Materialize GA addresses with movw + movt pairs for Darwin in PIC mode. e.g. · dfce83c8

Evan Cheng authored Jan 17, 2011

        movw    r0, :lower16:(L_foo$non_lazy_ptr-(LPC0_0+4))
        movt    r0, :upper16:(L_foo$non_lazy_ptr-(LPC0_0+4))
LPC0_0:
        add     r0, pc, r0

It's not yet enabled by default as some tests are failing. I suspect bugs in
down stream tools.

llvm-svn: 123619

dfce83c8

Jan 16, 2011
- Provide instruction sizes for ARMv5 variants of MUL instructions. · 27fc8f64
  Anton Korobeynikov authored Jan 16, 2011
```
This fixes PR8987

llvm-svn: 123598
```
  27fc8f64
- Update README.txt to remove the DAE enhancement. · 6a5171ba
  Anders Carlsson authored Jan 16, 2011
```
llvm-svn: 123597
```
  6a5171ba
- Only put unnamed_addr constants in mergeable sections. Fixes PR8297. · cba4c339
  Rafael Espindola authored Jan 16, 2011
```
llvm-svn: 123585
```
  cba4c339
- fix PR8514, a bug where the "heroic" transformation of shift/and · 35a2e65b
  Chris Lattner authored Jan 16, 2011
```
into and/shift would cause nodes to move around and a dangling pointer
to happen.  The code tried to avoid this with a HandleSDNode, but 
got the details wrong.

llvm-svn: 123578
```
  35a2e65b
- add some commentary · c326ebd1
  Chris Lattner authored Jan 16, 2011
```
llvm-svn: 123572
```
  c326ebd1
- Spill R4 if it's going to be used to restore SP from FP. · 572756ac
  Evan Cheng authored Jan 16, 2011
```
llvm-svn: 123567
```
  572756ac
- Implement AnalyzeBranch in Sparc Backend. · 1b0e2cbf
  Venkatraman Govindaraju authored Jan 16, 2011
```
llvm-svn: 123561
```
  1b0e2cbf
- fix PR8981, a crash trying to form a conditional inc with a floating point compare. · 218092e6
  Chris Lattner authored Jan 16, 2011
```
llvm-svn: 123560
```
  218092e6
- reapply my fix for PR8961 with a tweak to properly handle · 2d186574
  Chris Lattner authored Jan 16, 2011
```
multi-instruction sequences like calls.  Many thanks to Jakob for
finding a testcase.

llvm-svn: 123559
```
  2d186574
- Simplify a README.txt entry significantly to expose the core issue. · ef28abef
  Chandler Carruth authored Jan 16, 2011
```
llvm-svn: 123556
```
  ef28abef
Jan 15, 2011
- 80-col. · cc385c0c
  Eric Christopher authored Jan 15, 2011
```
llvm-svn: 123505
```
  cc385c0c
- Fix a comment. · b7a3c42e
  Bob Wilson authored Jan 15, 2011
```
llvm-svn: 123497
```
  b7a3c42e
- Fix 80-cols. · 2af9551e
  Eric Christopher authored Jan 14, 2011
```
llvm-svn: 123494
```
  2af9551e
Jan 14, 2011

'HiReg' is written but never read. Nuke its · b5241b2b
Ted Kremenek authored Jan 14, 2011
```
declaration and its assignments.

Found by clang static analyzer.

llvm-svn: 123486
```
b5241b2b

Add a possibility to switch between CFI directives- and table-based frame... · 9be547cf

Anton Korobeynikov authored Jan 14, 2011

Add a possibility to switch between CFI directives- and table-based frame description emission. Currently all the backends use table-based stuff.

llvm-svn: 123476

9be547cf

Cleanup · 4d9de6be
Anton Korobeynikov authored Jan 14, 2011
```
llvm-svn: 123475
```
4d9de6be
revert my fastisel patch again which apparently still gives the · e93e4f11
Chris Lattner authored Jan 14, 2011
```
llvm-gcc-i386-linux-selfhost buildbot heartburn...

llvm-svn: 123431
```
e93e4f11
reapply r123414 now that the botz are calmed down and the fix is already in. · 5ca13910
Chris Lattner authored Jan 14, 2011
```
llvm-svn: 123427
```
5ca13910

Completed :lower16: / :upper16: support for movw / movt pairs on Darwin. · d4a5c05c

Evan Cheng authored Jan 14, 2011

- Fixed :upper16: fix up routine. It should be shifting down the top 16 bits first.
- Added support for Thumb2 :lower16: and :upper16: fix up.
- Added :upper16: and :lower16: relocation support to mach-o object writer.

llvm-svn: 123424

d4a5c05c

r123414 broke llvm-gcc bootstrap apparently, revert · 21a64979
Chris Lattner authored Jan 14, 2011
```
llvm-svn: 123422
```
21a64979

fix PR8961 - a fast isel miscompilation where we'd insert a new instruction · 0c34cb42

Chris Lattner authored Jan 14, 2011

after sext's generated for addressing that got folded.  Previously we compiled
test5 into:

_test5:                                 ## @test5
## BB#0:
        movq    -8(%rsp), %rax          ## 8-byte Reload
        movq    (%rdi,%rax), %rdi
        addq    %rdx, %rdi
        movslq  %esi, %rax
        movq    %rax, -8(%rsp)          ## 8-byte Spill
        movq    %rdi, %rax
        ret

which is insane and wrong.  Now we produce:

_test5:                                 ## @test5
## BB#0:
	movslq	%esi, %rax
	movq	(%rdi,%rax), %rax
	addq	%rdx, %rax
	ret

llvm-svn: 123414

0c34cb42