Commits · 6631d7c4be2041b15c5f187d1f144f2de7182cb7 · Roger Ferrer / llvm-epi-0.8

May 09, 2009
- Enable loop bb placement optimization. · 51fa9c71
  Evan Cheng authored May 08, 2009
```
llvm-svn: 71291
```
  51fa9c71
- Avoid warning in release-asserts build. · 90b7cd6a
  Mike Stump authored May 08, 2009
```
llvm-svn: 71275
```
  90b7cd6a
May 08, 2009

Mirror how Fast ISel determines if a region.end intrinsic is the end of an · 88817808

Bill Wendling authored May 08, 2009

inlined function or the end of a function. Before, this was never executing the
"inlined" version of the Record method.

This will become important once the inlined Dwarf writer patch lands.

llvm-svn: 71268

88817808

Compute the offsets of the compile units. We need this so that when we emit a · fa03b008

Bill Wendling authored May 08, 2009

concrete instance of an inlined function, we can get the actual address of the
abstract instance inside of the compile unit.

This isn't currently used, but will be by a future check-in.

llvm-svn: 71263

fa03b008

Minor clean ups. No functionality change. · 2dc580c9
Bill Wendling authored May 08, 2009
```
llvm-svn: 71256
```
2dc580c9
Don't align loop header unless the loop back edge is below the header. · 6e9959e3
Evan Cheng authored May 08, 2009
```
llvm-svn: 71242
```
6e9959e3
Factor out cycle-finder code and make it generic. · 65a58168
Anton Korobeynikov authored May 08, 2009
```
llvm-svn: 71241
```
65a58168
Do not emit bit tests if target does not support natively left shift · c94dbf5b
Anton Korobeynikov authored May 08, 2009
```
llvm-svn: 71240
```
c94dbf5b
Properly expand libcalls for urem / srem. Also make code more straightforward. · e7a9661f
Anton Korobeynikov authored May 08, 2009
```
llvm-svn: 71238
```
e7a9661f
Typo · e2b78115
Anton Korobeynikov authored May 08, 2009
```
llvm-svn: 71237
```
e2b78115
Reverse branch condition only when there is a conditional branch. · 5b9b33bc
Evan Cheng authored May 08, 2009
```
llvm-svn: 71214
```
5b9b33bc
Add explicit braces to disambiguate nested if/else. Removes a warning. · b873ed67
Nick Lewycky authored May 08, 2009
```
llvm-svn: 71211
```
b873ed67

Optimize code placement in loop to eliminate unconditional branches or move... · 2fa28110

Evan Cheng authored May 08, 2009

Optimize code placement in loop to eliminate unconditional branches or move unconditional branch to the outside of the loop. e.g.

///       A:                                                                                                                                                                 
///       ...                                                                                                                                                                
///       <fallthrough to B>                                                                                                                                                 
///                                                                                                                                                                          
///       B:  --> loop header                                                                                                                                                
///       ...                                                                                                                                                                
///       jcc <cond> C, [exit]                                                                                                                                               
///                                                                                                                                                                          
///       C:                                                                                                                                                                 
///       ...                                                                                                                                                                
///       jmp B                                                                                                                                                              
///                                                                                                                                                                          
/// ==>                                                                                                                                                                      
///                                                                                                                                                                          
///       A:                                                                                                                                                                 
///       ...                                                                                                                                                                
///       jmp B                                                                                                                                                              
///                                                                                                                                                                          
///       C:  --> new loop header                                                                                                                                            
///       ...                                                                                                                                                                
///       <fallthough to B>                                                                                                                                                  
///                                                                                                                                                                          
///       B:                                                                                                                                                                 
///       ...                                                                                                                                                                
///       jcc <cond> C, [exit] 

llvm-svn: 71209

2fa28110

Fix pr4100. Do not remove no-op copies when they are dead. The register · e20be418

Bob Wilson authored May 07, 2009

scavenger gets confused about register liveness if it doesn't see them.
I'm not thrilled with this solution, but it only comes up when there are dead
copies in the code, which is something that hopefully doesn't happen much.

Here is what happens in pr4100: As shown in the following excerpt from the
debug output of llc, the source of a move gets reloaded from the stack,
inserting a new load instruction before the move.  Since that source operand
is a kill, the physical register is free to be reused for the destination
of the move.  The move ends up being a no-op, copying R3 to R3, so it is
deleted.  But, it leaves behind the load to reload %reg1028 into R3, and
that load is not updated to show that it's destination operand (R3) is dead.
The scavenger gets confused by that load because it thinks that R3 is live.

Starting RegAlloc of: %reg1025<def,dead> = MOVr %reg1028<kill>, 14, %reg0, %reg0
  Regs have values: 
  Reloading %reg1028 into R3
  Last use of R3[%reg1028], removing it from live set
  Assigning R3 to %reg1025
  Register R3 [%reg1025] is never used, removing it from live set

Alternative solutions might be either marking the load as dead, or zapping
the load along with the no-op copy.  I couldn't see an easy way to do
either of those, though.

llvm-svn: 71196

e20be418

May 07, 2009

Fix a comment (again). · 5352454e
Bob Wilson authored May 07, 2009
```
llvm-svn: 71180
```
5352454e
Fix a comment. · 94c0a9fe
Bob Wilson authored May 07, 2009
```
llvm-svn: 71179
```
94c0a9fe
Revert 71165. It did more than just revert 71158 and it introduced · 4bb6fa23
Dan Gohman authored May 07, 2009
```
several regressions. The problem due to 71158 is now fixed.

llvm-svn: 71176
```
4bb6fa23

Temporarily revert r71158. It was causing a failure during a full bootstrap: · 17f0f654

Bill Wendling authored May 07, 2009

checking for bcopy... no
checking for getc_unlocked... Assertion failed: (0 && "Unknown SCEV kind!"), function operator(), file /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvmCore.roots/llvmCore~obj/src/lib/Analysis/ScalarEvolution.cpp, line 511.
/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvmgcc42.roots/llvmgcc42~obj/src/libdecnumber/decUtility.c:360: internal compiler error: Abort trap
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.
make[4]: *** [decUtility.o] Error 1
make[4]: *** Waiting for unfinished jobs....
Assertion failed: (0 && "Unknown SCEV kind!"), function operator(), file /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvmCore.roots/llvmCore~obj/src/lib/Analysis/ScalarEvolution.cpp, line 511.
/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvmgcc42.roots/llvmgcc42~obj/src/libdecnumber/decNumber.c:5591: internal compiler error: Abort trap
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.
make[4]: *** [decNumber.o] Error 1
make[3]: *** [all-stage2-libdecnumber] Error 2
make[3]: *** Waiting for unfinished jobs....

llvm-svn: 71165

17f0f654

Move the tablegen-produced DebugLoc handling into a AsmWriter::processDebugLoc function. · 58f38113
Argyrios Kyrtzidis authored May 07, 2009
```
No functionality change.

llvm-svn: 71156
```
58f38113
Code refactoring. · 143bae5d
Evan Cheng authored May 07, 2009
```
llvm-svn: 71151
```
143bae5d
Rename "loop aligner" pass to "code placement optimization" pass. · f356a89e
Evan Cheng authored May 07, 2009
```
llvm-svn: 71150
```
f356a89e
Just turn aggressive stack coloring off at -O3. · 5a2fff71
Bill Wendling authored May 07, 2009
```
llvm-svn: 71140
```
5a2fff71
Temporarily revert r71010. It was causing massive failures during self-hosting. · 6144f63b
Bill Wendling authored May 07, 2009
```
llvm-svn: 71138
```
6144f63b
Make DwarfWriter::RecordInlinedFnStart more like the other DwarfWriter's methods: · baf3fee8
Argyrios Kyrtzidis authored May 07, 2009
```
-Have it return a label ID
-Remove the unused Instruction parameter

No functionality change.

llvm-svn: 71132
```
baf3fee8

May 06, 2009
- - Move some debug fields to coincide with how GCC emits them. No functionality · 4d11693f
  Bill Wendling authored May 06, 2009
```
  change.
- Reformatting.

llvm-svn: 71118
```
  4d11693f
- Do not use register as base ptr of pre- and post- inc/dec load / store nodes. · cfc05130
  Evan Cheng authored May 06, 2009
```
llvm-svn: 71098
```
  cfc05130
- CMake: Updated lib/CodeGen/CMakeLists.txt. · 49518d8f
  Oscar Fuentes authored May 06, 2009
```
llvm-svn: 71085
```
  49518d8f
- Add generic expansion of SUB when ADD and XOR · 2338f6c5
  Duncan Sands authored May 06, 2009
```
are legal.  Based on a patch by Micah Villmow.

llvm-svn: 71078
```
  2338f6c5
- Renamed Spiller classes (plus uses and related files) to VirtRegRewriter. · ab68cf24
  Lang Hames authored May 06, 2009
```
llvm-svn: 71057
```
  ab68cf24
May 05, 2009
- If a MachineBasicBlock has multiple ways of reaching another block, · bb2f1075
  Dan Gohman authored May 05, 2009
```
allow it to have multiple CFG edges to that block. This is needed
to allow MachineBasicBlock::isOnlyReachableByFallthrough to work
correctly. This fixes PR4126.

llvm-svn: 71018
```
  bb2f1075
- Enable stack coloring with regs at -O3. · 9bf90021
  Evan Cheng authored May 05, 2009
```
llvm-svn: 71010
```
  9bf90021
- Do not require variable debug info nodes to have a compile unit. · 929b644b
  Chris Lattner authored May 05, 2009
```
For implicit decls like "self" and "_cmd" in ObjC, these decls
should not have a location.

llvm-svn: 70964
```
  929b644b
- Do not substitute if the new register isn't in the register class of the operand being updated. · 6fc03634
  Evan Cheng authored May 05, 2009
```
llvm-svn: 70953
```
  6fc03634
- Move getInstrOperandRegClass from the scheduler to TargetInstrInfo. · 1ff2727c
  Evan Cheng authored May 05, 2009
```
llvm-svn: 70950
```
  1ff2727c
- Do forward and backward substitution to eliminate loads and stores when possible. · dbfb1025
  Evan Cheng authored May 04, 2009
```
llvm-svn: 70937
```
  dbfb1025
- Make DBG_STOPPOINT nodes, and therefore DBG_LABEL labels, get a DebugLoc, so that it · 354b1225
  Chris Lattner authored May 04, 2009
```
shows up in -print-machineinstrs.  This doesn't appear to affect anything, but it was
weird for some DBG_LABELs to have DebugLocs but not all of them.

llvm-svn: 70921
```
  354b1225
May 04, 2009
- Restore a comment. · 71fb950b
  Argyrios Kyrtzidis authored May 04, 2009
```
llvm-svn: 70900
```
  71fb950b
- -Remove the DwarfWriter::RecordSourceLine calls from the instruction selectors. · 9ae29b2d
  Argyrios Kyrtzidis authored May 04, 2009
```
-Depend on DebugLocs for source line info.

(Comes with Regression-Be-Gone(tm))

llvm-svn: 70871
```
  9ae29b2d
- Make sure to color with only allocatable registers for the specific register class. · 3f778056
  Evan Cheng authored May 04, 2009
```
llvm-svn: 70821
```
  3f778056
- The stack slots which share the same stack slot after coloring can, but do not... · 4b6072bc
  Evan Cheng authored May 04, 2009
```
The stack slots which share the same stack slot after coloring can, but do not have to, use the same register. In fact, they each may have different register class requirements.

llvm-svn: 70815
```
  4b6072bc