1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-27 05:53:07 +01:00
Commit Graph

31244 Commits

Author SHA1 Message Date
Bill Wendling
696d0388e9 An overhaul of the exception handling code. This is arguably more correct than
what was there before. In "no FP mode", we weren't generating labels and unwind
table entries after each "push" instruction. While more than likely "okay", it's
not technically correct. The major thing was that the ordering of when to define
a new CFA register and at what offset wasn't correct. This would cause the
exception handling to fail in ways most miserable to users.

I also cleaned up some code a bit. There's one function which has a "return" at
the beginning, so it's never used. Should I just remove it? :-)

llvm-svn: 79139
2009-08-15 21:27:32 +00:00
Bill Wendling
962adec4ee Reapply r79127. It was fixed by d0k.
llvm-svn: 79136
2009-08-15 21:21:19 +00:00
Bill Wendling
bfebbb6477 Revert r79127. It was causing compilation errors.
llvm-svn: 79135
2009-08-15 21:14:01 +00:00
Benjamin Kramer
5c54156e82 Don't try to get the context from an erased Instruction.
llvm-svn: 79134
2009-08-15 21:07:49 +00:00
Benjamin Kramer
b5771e5e21 Unbreak build. Evan, please make sure my changes are correct.
llvm-svn: 79133
2009-08-15 20:46:16 +00:00
Nick Lewycky
de61ef6c5e SSI construction should just go ahead and ignore instructions in unreachable
blocks.

llvm-svn: 79132
2009-08-15 20:12:18 +00:00
Evan Cheng
5d841097a9 Change allowsUnalignedMemoryAccesses to take type argument since some targets
support unaligned mem access only for certain types. (Should it be size
instead?)

ARM v7 supports unaligned access for i16 and i32, some v6 variants support it
as well.

llvm-svn: 79127
2009-08-15 19:23:44 +00:00
Jakob Stoklund Olesen
7f4ef2d59a Refine EarlyClobber assert in register scavenger.
It is legal for an inline asm operand to use an earlyclobber register if the
use operand is tied to the earlyclobber operand. The issue is discussed here:

  http://gcc.gnu.org/ml/gcc/1999-04n/msg00431.html

We should perhaps let only the machine code verifier worry about these finer
details. EarlyClobber operands are not really interesting to the scavenger.

This fixes PR4528 for the third time.

llvm-svn: 79122
2009-08-15 18:16:58 +00:00
Chris Lattner
07186e7840 the .eh_frame sections we generate need to be writable (which
is why they are datarel).  This should fix PR4724, and is fallout
from r78890.

llvm-svn: 79111
2009-08-15 16:54:02 +00:00
Nicolas Geoffray
572b4e6cb3 Use the new API for creating an OpaqueType.
llvm-svn: 79107
2009-08-15 15:41:32 +00:00
Chris Lattner
e12f92568f tidy up
llvm-svn: 79101
2009-08-15 15:08:28 +00:00
Nicolas Geoffray
679c83e1bf Update cpp generation with new LLVM API for primitive types.
llvm-svn: 79098
2009-08-15 14:47:42 +00:00
Sanjiv Gupta
2399e49802 Revert a few changes that were done in 78603.
PIC16DebugInfo currently rely on NameStr of composite type descriptors to uniquely
identify debug info for two aggregate type decls with same name.
This implementation will change when we have MDNodes based debug info implemenatation in place

llvm-svn: 79097
2009-08-15 14:36:48 +00:00
Jakob Stoklund Olesen
8f6660c417 Don't setCalleeSavedInfoValid() until spills are interted.
In a naked function, the flag is never set and getPristineRegs() returns an
empty list. That means naked functions are able to clobber callee saved
registers, but that is the whole point of naked functions.

This fixes PR4716.

llvm-svn: 79096
2009-08-15 13:10:46 +00:00
Jakob Stoklund Olesen
db3f862b6f Allow double defs in the machine code verifier after the addPreRegAlloc passes.
llvm-svn: 79095
2009-08-15 13:10:15 +00:00
Richard Osborne
31f04e14f0 Move XCore AsmPrinter to XCore/AsmPrinter directory.
llvm-svn: 79094
2009-08-15 12:53:15 +00:00
Tilmann Scheller
03f517b799 Add support for the PowerPC 64-bit SVR4 ABI.
The Link Register is volatile when using the 32-bit SVR4 ABI.
Make it possible to use the 64-bit SVR4 ABI.
Add non-volatile registers for the 64-bit SVR4 ABI.
Make sure r2 is a reserved register when using the 64-bit SVR4 ABI.
Update PPCFrameInfo for the 64-bit SVR4 ABI.
Add FIXME for 64-bit Darwin PPC.
Insert NOP instruction after direct function calls.
Emit official procedure descriptors.
Create TOC entries for GlobalAddress references.
Spill 64-bit non-volatile registers to the correct slots.
Only custom lower VAARG when using the 32-bit SVR4 ABI.
Use simple VASTART lowering for the 64-bit SVR4 ABI.

llvm-svn: 79091
2009-08-15 11:54:46 +00:00
Jakob Stoklund Olesen
e093ae5810 Update LocalRewriter::DistanceMap when inserting stack loads.
In the included test case, a stack load was not included in DistanceMap. That
caused TransferDeadness to ignore the instruction, leading to a scavenger
assert.

llvm-svn: 79090
2009-08-15 11:03:03 +00:00
Evan Cheng
e65f967fa0 80 col violations.
llvm-svn: 79087
2009-08-15 08:38:52 +00:00
Evan Cheng
9d351a7246 Turn on if-conversion for thumb2.
llvm-svn: 79084
2009-08-15 07:59:10 +00:00
Chris Lattner
aa462f6d51 update for rename.
llvm-svn: 79082
2009-08-15 06:14:07 +00:00
Chris Lattner
8101f115c5 rename PIC16Section.h -> MCSectionPIC16.h for consistency with
the class it defines.

llvm-svn: 79081
2009-08-15 06:13:40 +00:00
Chris Lattner
7021a8b286 cmake likes its explicit list of files to build.
llvm-svn: 79080
2009-08-15 06:10:23 +00:00
Chris Lattner
ee52642cd3 use XCore-specific section with xcore specific cp/dp flags to restore
support for globals going into the appropriate sections with the flags.

This hopefully finishes unbreaking the previous behavior that I broke before.

llvm-svn: 79079
2009-08-15 06:09:35 +00:00
Chris Lattner
2c1a96d901 If ELF subtargets don't want to support 4/8/16-byte mergable sections, allow
them to null out the default section pointers.

llvm-svn: 79078
2009-08-15 06:08:34 +00:00
Chris Lattner
3665e2058e add support for target-specific ELF section flags, add a new MCSectionXCore
class which represents the XCore cp/dp section flags.  No functionality
change yet.

llvm-svn: 79077
2009-08-15 05:56:11 +00:00
Dan Gohman
fcb5e33aa1 Simplify a few more things, eliminating a few more dependencies on
"the current basic block".

llvm-svn: 79069
2009-08-15 02:07:36 +00:00
Dan Gohman
a68aa347d8 Simplify this code to not depend as much on CurMBB.
llvm-svn: 79068
2009-08-15 02:06:22 +00:00
Evan Cheng
cf13e0c732 Do not use frame register to reference fixed stack objects if the function is frameless.
llvm-svn: 79067
2009-08-15 02:05:35 +00:00
Dan Gohman
d0419623b1 Always check to see if raw_fd_ostream's file descriptor is attached to
a terminal, not just when it's STDOUT_FILENO.

llvm-svn: 79066
2009-08-15 02:05:19 +00:00
Dan Gohman
8aea7a0db0 Add support for column computation on unbuffered streams.
llvm-svn: 79065
2009-08-15 02:02:59 +00:00
Dan Gohman
4541bef274 Move FormattedStream's write_impl out of line.
llvm-svn: 79064
2009-08-15 02:01:04 +00:00
Dan Gohman
fe04559261 Remove an unnecessary #include.
llvm-svn: 79063
2009-08-15 01:56:38 +00:00
Chris Lattner
c807a76334 switch DominanceFrontier::splitBlock to use a smallvector for
the pred list instead of a vector, saving a boat load of malloc/free's.

llvm-svn: 79062
2009-08-15 01:39:28 +00:00
Dan Gohman
d69323d37a On x86-64, for a varargs function, don't store the xmm registers to
the register save area if %al is 0. This avoids touching xmm
regsiters when they aren't actually used.

llvm-svn: 79061
2009-08-15 01:38:56 +00:00
Devang Patel
b6add4ace3 Do not completely skip subrange info for a zero sized array.
llvm-svn: 79044
2009-08-14 20:59:16 +00:00
Evan Cheng
4e97ccc132 Leaf functions which do not save CSRs can be frameless even with -disable-fp-elim.
llvm-svn: 79039
2009-08-14 20:48:13 +00:00
Anton Korobeynikov
3a0cde8c91 Allow targets to specify their choice of calling conventions per
libcall. Take advantage of this in the ARM backend to rectify broken
choice of CC when hard float is in effect. PIC16 may want to see if
it could be of use in MakePIC16Libcall, which works unchanged.

Patch by Sandeep!

llvm-svn: 79033
2009-08-14 20:10:52 +00:00
Evan Cheng
67fd47b38b Add Thumb2 lsr hooks.
llvm-svn: 79032
2009-08-14 20:09:37 +00:00
Daniel Dunbar
3c9ab92a24 llvm-mc: Fix bugs where bytes were unintentionally being printed as signed.
- We now print all of 403.gcc cleanly (llvm-mc -> 'as' as diffed to 'as'), minus two 
   'rep;movsl' instructions (which I missed before).

llvm-svn: 79031
2009-08-14 19:59:24 +00:00
Oscar Fuentes
4acb0629da CMake: Corrected variable check.
llvm-svn: 79030
2009-08-14 19:56:04 +00:00
Bruno Cardoso Lopes
e298b64a93 *try* to use a better name to describe how common symbols are marked on the elf object file.
llvm-svn: 79029
2009-08-14 19:45:38 +00:00
Evan Cheng
ebbcd00c17 80 col violation.
llvm-svn: 79026
2009-08-14 19:11:20 +00:00
Daniel Dunbar
b9d9ce9513 llvm-mc: When handling a .set, make sure to print subsequent references to the
symbol as the symbol name itself, not the expression it was defined to. These
have different semantics due to the quirky .set behavior (which absolutizes an
expression that would otherwise be treated as a relocation).

llvm-svn: 79025
2009-08-14 19:10:46 +00:00
Anton Korobeynikov
a52d80928a Cleanup the mess in msp430 target registration and hopefully unbreak the build
llvm-svn: 79024
2009-08-14 19:06:50 +00:00
Evan Cheng
1fa8ad98aa Indentation.
llvm-svn: 79022
2009-08-14 19:01:37 +00:00
Daniel Dunbar
b00271e266 llvm-mc: zerofill shouldn't print quotes around the section,segment.
llvm-svn: 79017
2009-08-14 18:51:45 +00:00
Chris Lattner
be587b554b fix "pc" to be lower case in a target triple, patch by Yonggang Luo
llvm-svn: 79016
2009-08-14 18:48:13 +00:00
Anton Korobeynikov
094a764f5d Hopefully unbreak cmake builds
llvm-svn: 79015
2009-08-14 18:46:49 +00:00
Evan Cheng
7d8caa143b Also shrink immediate branches; also more assembler workarounds.
llvm-svn: 79014
2009-08-14 18:31:44 +00:00