- Jan 16, 2009
-
-
Dan Gohman authored
array instead, since this is what the scheduler actually cares about. And remove a check that is unnecessary, since it can assume that SUnits isn't empty. llvm-svn: 62362
-
Chris Lattner authored
llvm-svn: 62359
-
Devang Patel authored
llvm-svn: 62358
-
Evan Cheng authored
CreateVirtualRegisters does trivial copy coalescing. If a node def is used by a single CopyToReg, it reuses the virtual register assigned to the CopyToReg. This won't work for SDNode that is a clone or is itself cloned. Disable this optimization for those nodes or it can end up with non-SSA machine instructions. llvm-svn: 62356
-
Devang Patel authored
llvm-svn: 62341
-
Devang Patel authored
llvm-svn: 62328
-
Dan Gohman authored
warning. llvm-svn: 62327
-
Mikhail Glushenkov authored
Split Support/Registry.h into two files so that we have less to recompile every time CommandLine.h is changed. llvm-svn: 62312
-
Mikhail Glushenkov authored
llvm-svn: 62307
-
Devang Patel authored
llvm-svn: 62298
-
Devang Patel authored
llvm-svn: 62294
-
Dan Gohman authored
a new toy hazard recognizier heuristic which attempts to direct the scheduler to avoid clumping large groups of loads or stores too densely. llvm-svn: 62291
-
Devang Patel authored
llvm-svn: 62288
-
Devang Patel authored
Validate dbg_* intrinsics before lowering them. llvm-svn: 62286
-
- Jan 15, 2009
-
-
Mon P Wang authored
llvm-svn: 62285
-
Dan Gohman authored
to support MachineInstr-based scheduling in addition to SDNode-based scheduling. llvm-svn: 62284
-
Dan Gohman authored
loops, hoisting instructions all the way out in one step rather than hoisting them one nest level at a time. Also, make a few other code simplifications. This speeds up MachineLICM by several fold. llvm-svn: 62283
-
Rafael Espindola authored
llvm-svn: 62279
-
Devang Patel authored
llvm-svn: 62276
-
Dan Gohman authored
and into the ScheduleDAGInstrs class, so that they don't get destructed and re-constructed for each block. This fixes a compile-time hot spot in the post-pass scheduler. To help facilitate this, tidy and do some minor reorganization in the scheduler constructor functions. llvm-svn: 62275
-
Devang Patel authored
Use light weight DebugInfo object directly. llvm-svn: 62269
-
Dan Gohman authored
previous commit. llvm-svn: 62266
-
Dan Gohman authored
llvm-svn: 62262
-
Dan Gohman authored
llvm-svn: 62260
-
Gabor Greif authored
llvm-svn: 62256
-
- Jan 14, 2009
-
-
Devang Patel authored
llvm-svn: 62209
-
Devang Patel authored
Removoe MachineModuleInfo methods (and related DebugInfoDesc class hierarchy) that were used to handle debug info. llvm-svn: 62199
-
Devang Patel authored
Keep "has debug info" big in MachineModuleInfo to avoid circular dependency between AsmPrinter and CodeGen. llvm-svn: 62191
-
- Jan 13, 2009
-
-
Devang Patel authored
llvm-svn: 62190
-
Devang Patel authored
llvm-svn: 62185
-
Devang Patel authored
llvm-svn: 62184
-
Dan Gohman authored
llvm-svn: 62178
-
Dan Gohman authored
scheduling dependencies. Add assertion checks to help catch this. It appears the Mips target defaults to list-td, and it has a regression test that uses a physreg dependence. Such code was liable to be miscompiled, and now evokes an assertion failure. llvm-svn: 62177
-
Dan Gohman authored
been modified, to avoid trouble in the (unlikely) scenario that D is a reference to an element in one of those arrays. llvm-svn: 62173
-
Duncan Sands authored
via two paths, process it once not twice, d'oh! Analysis, testcase and original patch thanks to Mon Ping Wang. llvm-svn: 62169
-
Duncan Sands authored
was not being cleaned by ExpungeNode. llvm-svn: 62167
-
Duncan Sands authored
llvm-svn: 62166
-
Evan Cheng authored
llvm-svn: 62151
-
Evan Cheng authored
FIX llvm-gcc bootstrap on x86_64 linux. If a virtual register is copied to a physical register, it's not necessarily defined by a copy. We have to watch out it doesn't clobber any sub-register that might be live during its live interval. If the live interval crosses a basic block, then it's not safe to check with the less conservative check (by scanning uses and defs) because it's possible a sub-register might be live out of the block. llvm-svn: 62144
-
Devang Patel authored
Use DebugInfo interface to lower dbg_* intrinsics. llvm-svn: 62127
-