- Nov 13, 2011
-
-
Jakob Stoklund Olesen authored
Nobody cared, StackSlotColoring scans the instructions to find used stack slots. llvm-svn: 144485
-
Jakob Stoklund Olesen authored
Most of this stuff was supporting the old deferred spill code insertion mechanism. Modern spillers just edit machine code in place. llvm-svn: 144484
-
Peter Collingbourne authored
literals of array type, materialise a temporary. llvm-svn: 144483
-
Jakob Stoklund Olesen authored
The information was only used by the register allocator in StackSlotColoring. llvm-svn: 144482
-
Jakob Stoklund Olesen authored
It was off by default. The new register allocators don't have the problems that made it necessary to reallocate registers during stack slot coloring. llvm-svn: 144481
-
Jakob Stoklund Olesen authored
And there was much rejoicing. llvm-svn: 144480
-
Jakob Stoklund Olesen authored
The very complicated VirtRegRewriter is going away. llvm-svn: 144479
-
Jakob Stoklund Olesen authored
This is dead code, all register allocators use InlineSpiller. llvm-svn: 144478
-
Jakob Stoklund Olesen authored
The current register allocators all use the inline spiller. llvm-svn: 144477
-
Jakob Stoklund Olesen authored
It is worth noting that the old spiller would split live ranges around basic blocks. The new spiller doesn't do that. PBQP should do its own live range splitting with SplitEditor::splitSingleBlock() if desired. See RAGreedy::tryBlockSplit(). llvm-svn: 144476
-
- Nov 12, 2011
-
-
Jakob Stoklund Olesen authored
RegAllocGreedy has been the default for six months now. Deleting RegAllocLinearScan makes it possible to also delete VirtRegRewriter and clean up the spiller code. llvm-svn: 144475
-
Jakob Stoklund Olesen authored
Counting the number of occurences of each opcode is not a useful test. llvm-svn: 144474
-
Jakob Stoklund Olesen authored
Or maybe we are just getting lucky. llvm-svn: 144473
-
Jakob Stoklund Olesen authored
llvm-svn: 144472
-
Jakob Stoklund Olesen authored
Filed PR11364 to track the problem. Should the register allocator eliminate dead code? llvm-svn: 144471
-
Jakob Stoklund Olesen authored
This test was committed with a bugfix to RemoveCopyByCommutingDef, but that optimization is no longer triggered by this test. llvm-svn: 144470
-
Jakob Stoklund Olesen authored
This test is for a very specific LocalRewriter bug. LocalRewriter is going away. llvm-svn: 144469
-
Richard Smith authored
or MemberExpr which refers to it. As a side-effect, MemberExprs which refer to static member functions and static data members are now emitted as constant expressions. llvm-svn: 144468
-
Argyrios Kyrtzidis authored
llvm-svn: 144467
-
Argyrios Kyrtzidis authored
and remove them. llvm-svn: 144466
-
Argyrios Kyrtzidis authored
[PCH] When completing an objc forward reference, do not serialize the chain of its categories because it is going to be rewritten (and the chain will be serialized again), otherwise we may form a cycle in its categories list when deserializing. Also introduce ASTMutationListener::CompletedObjCForwardRef to notify that a forward reference was completed; using Decl's isChangedSinceDeserialization/setChangedSinceDeserialization is bug inducing and kinda gross, we should phase it out. Fixes infinite loop in rdar://10418538. llvm-svn: 144465
-
David Blaikie authored
llvm-svn: 144464
-
Jakob Stoklund Olesen authored
I don't think this test does what is was supposed to do, and LocalRewriter is going away anyway. llvm-svn: 144463
-
Jakob Stoklund Olesen authored
llvm-svn: 144462
-
Jakob Stoklund Olesen authored
llvm-svn: 144461
-
Jakob Stoklund Olesen authored
llvm-svn: 144460
-
Jakob Stoklund Olesen authored
This test doesn't expose the issue with RAGreedy. I filed PR11363 to track the missing InlineSpiller feature. llvm-svn: 144459
-
Jakob Stoklund Olesen authored
The test is checking that the output doesn't contains any 'mov ' strings. It does contain movl, though. llvm-svn: 144458
-
Craig Topper authored
Add more AVX2 shift lowering support. Move AVX2 variable shift to use patterns instead of custom lowering code. llvm-svn: 144457
-
Eli Friedman authored
llvm-svn: 144455
-
Nick Lewycky authored
llvm-svn: 144454
-
Akira Hatanaka authored
llvm-svn: 144453
-
Akira Hatanaka authored
llvm-svn: 144452
-
Sean Callanan authored
methods but also class methods for Objective-C. Clang emits Objective-C method names with '\1' at the beginning, and the JIT has pre-existing logic to try prepending a '\1' when searching a module for an instance method (that is, a method whose name begins with '-'). I simply extended it to do the same thing when it encountered a class method (a method whose name begins with '+'). llvm-svn: 144451
-
Johnny Chen authored
llvm-svn: 144450
-
Akira Hatanaka authored
llvm-svn: 144449
-
Akira Hatanaka authored
llvm-svn: 144448
-
Akira Hatanaka authored
llvm-svn: 144447
-
Argyrios Kyrtzidis authored
That stuff can be added later on if we need them. Also add some const goodness. llvm-svn: 144446
-
Daniel Dunbar authored
llvm-svn: 144445
-