handle conditional move instructions:
-- cpMem<->Reg functions now support CC registers (int and FP) correctly.
-- Scratch registers must be explicitly provided to cpMem<->Reg when
needed, since CC regs need one to be copied to/from memory.
-- CC regs are saved to a scratch register instead of stack.
-- All regs used by a instruction are now recorded in MachineInstr::regsUsed,
since regs used to save values *across* an instruction are not obvious
either from the operands or from the LiveVar sets.
-- An (explicit or implicit) operand may now be both a def and a use.
This is needed for conditional move operations.
So an operand may need spill code both before and after the instruction.
-- class MachineCodeForBasicBlock is now an annotation on BasicBlock.
llvm-svn: 2833
so additional dep. edges have to be added.
This was needed to correctly handle conditional move instructions!
MachineCodeForBasicBlock is now an annotation on BasicBlock.
Renamed "earliestForNode" to "earliestReadyTimeForNode".
llvm-svn: 2826
An (explicit or implicit) operand may now be both a def and a use.
Also add a set of regs used by each instruction.
dump() no longer takes an optional argument, which doesn't work in gdb.
llvm-svn: 2821
-- passing FP arguments to functions with more than 6 arguments
-- passing FP arguments to varargs functions
-- passing FP arguments to functions with no prototypes
-- incorrect coloring for CC registers (both int and FP): interferences
were being completely ignored for int CC and were considered but no
spills were marked for fp CC!
Also some code improvements:
-- better interface to generating machine instr for common cases
(many places still need to be updated to use this interface)
-- annotations on MachineInstr to communicate information from
one codegen phase to another (now used to pass information about
CALL/JMPLCALL operands from selection to register allocation)
-- all sizes and offests in class TargetData are uint64_t instead of uint
llvm-svn: 2642
- Rename runOnMethod to runOnFunction
* Transform getAnalysisUsageInfo into getAnalysisUsage
- Method is now const
- It now takes one AnalysisUsage object to fill in instead of 3 vectors
to fill in
- Pass's now specify which other passes they _preserve_ not which ones
they modify (be conservative!)
- A pass can specify that it preserves all analyses (because it never
modifies the underlying program)
* s/Method/Function/g in other random places as well
llvm-svn: 2333
AddedInstrns sets for the first machine instruction. It is hard
to ensure that the right order is preserved, and sure enough, the
order was broken. Instead, use a separate set for the function entry.
llvm-svn: 2318
AddedInstrns sets for the first machine instruction. It is hard
to ensure that the right order is preserved, and sure enough, the
order was broken. Instead, use a separate set for the function entry.
llvm-svn: 2312
(This can be improved to avoid making the initial pass over the method.)
Also, ensure automatic vars and reg. spills areas are not extended
if their sizes are used for computing some other offset.
llvm-svn: 2310
* Delete the DelaySlotInfo objects created by the SchedulingManager class.
These leaked objects were accounting for 3/4 of the memory leaked by the
backend, so this is a relatively major win.
* Reorganize SchedulingManager::getDelaySlotInfoForInstr so that it has
better code locality (making it easier to read).
llvm-svn: 2197
