Commits · ada366c691150bb9bc1de0a453a37a2eb6745d9f · Roger Ferrer / llvm-epi-0.8

Jul 19, 2011
- Micro-opt: Only emit compact unwind if there is a compact unwind encoding to emit. · ada366c6
  Bill Wendling authored Jul 19, 2011
```
llvm-svn: 135452
```
  ada366c6
- Use the CompactUnwindEncoding from the Frame, if it's defined. · 49bc680b
  Bill Wendling authored Jul 19, 2011
```
llvm-svn: 135451
```
  49bc680b
- Move getInitialFrameState from TargetFrameInfo to MCAsmInfo (suggestions for · 67c033e6
  Evan Cheng authored Jul 18, 2011
```
better location welcome).

llvm-svn: 135438
```
  67c033e6
Jul 18, 2011

Sink getDwarfRegNum, getLLVMRegNum, getSEHRegNum from TargetRegisterInfo down · d60fa58b

Evan Cheng authored Jul 18, 2011

to MCRegisterInfo. Also initialize the mapping at construction time.

This patch eliminate TargetRegisterInfo from TargetAsmInfo. It's another step
towards fixing the layering violation.

llvm-svn: 135424

d60fa58b

Jul 15, 2011
- Move some parts of TargetAsmInfo down to MCAsmInfo. This is not the greatest · a83b37a9
  Evan Cheng authored Jul 15, 2011
```
solution but it is a small step towards removing the horror that is
TargetAsmInfo.

llvm-svn: 135237
```
  a83b37a9
- Remove setting the bit for personality function. It should be set by the linker. · 9b8605f7
  Bill Wendling authored Jul 15, 2011
```
llvm-svn: 135236
```
  9b8605f7
- Encode that we have a personality function. · c998f71d
  Bill Wendling authored Jul 15, 2011
```
llvm-svn: 135227
```
  c998f71d
- * If we have an LSDA, we need to mark it in the encoding. · dafd598f
  Bill Wendling authored Jul 14, 2011
```
* The personality function should be encoded as an absolute pointer to the function.

llvm-svn: 135215
```
  dafd598f
Jul 14, 2011

Unfortunately several files in MC are badly violating layering rule by using · c7ac690b

Evan Cheng authored Jul 14, 2011

TargetAsmInfo, which in turn pulls in TargetRegisterInfo, etc. :-( There are
other cases of violations, but this is probably the worst.

This patch is but one small step towards fixing this. 500 more steps to go. :-(

llvm-svn: 135131

c7ac690b

Jul 13, 2011
- Don't emit the FDE end label if the last thing emitted was a compact unwind and · ee6e776b
  Bill Wendling authored Jul 13, 2011
```
not the FDE

llvm-svn: 135020
```
  ee6e776b
Jul 07, 2011
- Add a target hook to encode the compact unwind information. · b6adf46f
  Bill Wendling authored Jul 07, 2011
```
llvm-svn: 134577
```
  b6adf46f
- Clean up the #includes. · 7ed56250
  Bill Wendling authored Jul 06, 2011
```
llvm-svn: 134557
```
  7ed56250
Jul 01, 2011
- Remove tabs. · c737ac18
  Bill Wendling authored Jun 30, 2011
```
llvm-svn: 134212
```
  c737ac18
- Improve comment: Show the register the DWARF label is added to. · 40cc7497
  Bill Wendling authored Jun 30, 2011
```
llvm-svn: 134209
```
  40cc7497
- Add one more comment to the FDE verbose asm output. · 2fd8d775
  Bill Wendling authored Jun 30, 2011
```
llvm-svn: 134200
```
  2fd8d775
- Add comments to the FDE. · f166ab44
  Bill Wendling authored Jun 30, 2011
```
llvm-svn: 134196
```
  f166ab44
Jun 30, 2011
- Add more comments to the ASM output for the CIE's "moves". · e7fe47e5
  Bill Wendling authored Jun 30, 2011
```
llvm-svn: 134194
```
  e7fe47e5
- Add comments to the ASM output to help understand the compact unwind and CIE tables. · 567a1aee
  Bill Wendling authored Jun 30, 2011
```
llvm-svn: 134191
```
  567a1aee
- * Use the proper size to output the range size. · 75174669
  Bill Wendling authored Jun 30, 2011
```
* Rough in the compact encoding part.

llvm-svn: 134119
```
  75174669
- Stupid error: If the LSDA and Personality functions aren't there, emit 0 instead · 4cdb206e
  Bill Wendling authored Jun 29, 2011
```
of the encoding.

llvm-svn: 134117
```
  4cdb206e
- We don't want to use relocations inside the compact unwind section. Just use the · 8fa4ada5
  Bill Wendling authored Jun 29, 2011
```
symbols instead.

llvm-svn: 134115
```
  8fa4ada5
Jun 23, 2011
- Use a reference. Don't make a useless copy of the vector. · 25fbb923
  Bill Wendling authored Jun 23, 2011
```
llvm-svn: 133707
```
  25fbb923
- Formatting changes. No functionality change. · a4b9d1f9
  Bill Wendling authored Jun 23, 2011
```
llvm-svn: 133706
```
  a4b9d1f9
- Use the presence of the __compact_unwind section to indicate that a target · 9af2fa9d
  Bill Wendling authored Jun 23, 2011
```
supports compact unwind info instead of having a separate flag indicating this.

llvm-svn: 133685
```
  9af2fa9d
- Some skeleton code to emit the compact unwind. If the information is unable to · e8fc92ad
  Bill Wendling authored Jun 23, 2011
```
be emitted in a compact way, we then default to emitting a CIE and FDE.

llvm-svn: 133676
```
  e8fc92ad
Jun 18, 2011
- Directly print to a raw_ostream instead of printing to a buffer first. · f7cb7ef9
  Benjamin Kramer authored Jun 18, 2011
```
llvm-svn: 133352
```
  f7cb7ef9
May 20, 2011
- No reason not to allow defining the CFA as a reg w/ offset zero. · 8aded86e
  Jim Grosbach authored May 20, 2011
```
llvm-svn: 131760
```
  8aded86e
May 11, 2011
- Avoid a gcc warning. · de32fa16
  Rafael Espindola authored May 10, 2011
```
llvm-svn: 131161
```
  de32fa16
May 10, 2011
- On MachO, unlike ELF, there should be no relocation to produce the CIE pointer. · 99f67355
  Rafael Espindola authored May 10, 2011
```
llvm-svn: 131149
```
  99f67355
- The EH symbols are only needed in eh_frame, not debug_frame. · d0d23542
  Rafael Espindola authored May 10, 2011
```
llvm-svn: 131146
```
  d0d23542
- In a debug_frame the cfi offset is to the start of the debug_frame section! · 27390b4a
  Rafael Espindola authored May 10, 2011
```
llvm-svn: 131129
```
  27390b4a
- Add support for producing .deubg_frame sections. · 1ecb12fc
  Rafael Espindola authored May 10, 2011
```
llvm-svn: 131121
```
  1ecb12fc
- Small cleanups. · b6444c04
  Rafael Espindola authored May 10, 2011
```
llvm-svn: 131120
```
  b6444c04
- Remove unused argument. · 0e130d1d
  Rafael Espindola authored May 10, 2011
```
llvm-svn: 131118
```
  0e130d1d
May 09, 2011
- Eliminate an unused line to fix a warning. · 39244582
  NAKAMURA Takumi authored May 08, 2011
```
llvm-svn: 131078
```
  39244582
May 08, 2011

Don't force relaxation of AdvanceLoc instructions on OS X. gdb is happy with · ab39c635

Rafael Espindola authored May 08, 2011

the smaller encoding and this cuts 270336 bytes from a release version of
clang and 1246272 bytes from a debug build.

llvm-svn: 131067

ab39c635

May 07, 2011
- Switch Darwin to the generic CIE/FDE printer. · bcabcdce
  Rafael Espindola authored May 06, 2011
```
llvm-svn: 131031
```
  bcabcdce
May 03, 2011
- MCDwarf: Don't save Twine to local variable, this is almost never safe to do · f6daf94a
  Daniel Dunbar authored May 03, 2011
```
(and should thus never be done).
 - Should fix a crash on win32.

llvm-svn: 130793
```
  f6daf94a
May 01, 2011

GCC uses a different encoding of pointers in the FDE when using · 750cb615
Rafael Espindola authored May 01, 2011
```
-fno-dwarf2-cfi-asm. Implement the same behavior.

llvm-svn: 130637
```
750cb615

Simplify the handling of pcrel relocations on ELF. Now we do the right thing · fd057853

Rafael Espindola authored May 01, 2011

for all symbol differences and can drop the old EmitPCRelSymbolValue
method.

This also make getExprForFDESymbol on ELF equal to the one on MachO, and it
can be made non-virtual.

llvm-svn: 130634

fd057853