summaryrefslogtreecommitdiffstats
path: root/llvm/lib/CodeGen/LiveIntervalAnalysis.cpp
Commit message (Collapse)AuthorAgeFilesLines
* - Rewrite code that update register live interval that's split.Evan Cheng2008-10-291-3/+24
| | | | | | | - Create and update spill slot live intervals. - Lots of bug fixes. llvm-svn: 58367
* Fix PR2634. Create new virtual registers from spills early so that weDavid Greene2008-10-271-8/+19
| | | | | | | can give it the same stack slot as the spilled interval if it is folded. This prevents the fold/unfold code from pointing to the wrong register. llvm-svn: 58255
* Avoid splitting an interval multiple times; avoid splitting ↵Evan Cheng2008-10-241-0/+9
| | | | | | re-materializable val# (for now). llvm-svn: 58068
* By min, I mean max.Evan Cheng2008-10-181-1/+1
| | | | llvm-svn: 57766
* When creating intervals, leave min(1, numdefs) holes after each instruction.Evan Cheng2008-10-181-5/+13
| | | | llvm-svn: 57765
* Add an option to enable StrongPHIElimination, for ease of testing.Owen Anderson2008-10-071-2/+7
| | | | llvm-svn: 57259
* Switch the MachineOperand accessors back to the short names likeDan Gohman2008-10-031-11/+11
| | | | | | isReg, etc., from isRegister, etc. llvm-svn: 57006
* Fix a simple error in renumbering kill markaers, that took an inordinant ↵Owen Anderson2008-09-301-1/+1
| | | | | | amount of time to track down. llvm-svn: 56889
* Re-apply 56835 along with header file changes.Evan Cheng2008-09-301-5/+15
| | | | llvm-svn: 56848
* Revert commit 56835 since it breaks the build.Duncan Sands2008-09-301-15/+5
| | | | | | | | | | "If a re-materializable instruction has a register operand, the spiller will change the register operand's spill weight to HUGE_VAL to avoid it being spilled. However, if the operand is already in the queue ready to be spilled, avoid re-materializing it". llvm-svn: 56837
* If a re-materializable instruction has a register operand, the spiller will ↵Evan Cheng2008-09-301-5/+15
| | | | | | change the register operand's spill weight to HUGE_VAL to avoid it being spilled. However, if the operand is already in the queue ready to be spilled, avoid re-materializing it. llvm-svn: 56835
* Next round of earlyclobber handling. Approach theDale Johannesen2008-09-241-13/+15
| | | | | | | | | | RA problem by expanding the live interval of an earlyclobber def back one slot. Remove overlap-earlyclobber throughout. Remove earlyclobber bits and their handling from live internals. llvm-svn: 56539
* Fetch the starting index of the block when assigning intervals. This gets ↵Owen Anderson2008-09-211-2/+2
| | | | | | | | live-in indices correct in the presence of things like EH labels. llvm-svn: 56410
* Remove AsmThatEarlyClobber etc. from LiveIntervalAnalysisDale Johannesen2008-09-191-71/+4
| | | | | | | and redo as linked list walk. Logic moved into RA. Per review feedback. llvm-svn: 56326
* Add a bit to mark operands of asm's that conflictDale Johannesen2008-09-171-2/+80
| | | | | | | | | with an earlyclobber operand elsewhere. Propagate this bit and the earlyclobber bit through SDISel. Change linear-scan RA not to allocate regs in a way that conflicts with an earlyclobber. See also comments. llvm-svn: 56290
* Live intervals for live-in registers should begin at the beginning of a ↵Owen Anderson2008-09-151-7/+11
| | | | | | | | | basic block, not at the first instruction. Also, their valno's should have an unknown def. This has no effect currently, but was causing issues when StrongPHIElimination was enabled. llvm-svn: 56231
* Remove isImm(), isReg(), and friends, in favor of Dan Gohman2008-09-131-3/+3
| | | | | | | | | isImmediate(), isRegister(), and friends, to avoid confusion about having two different names with the same meaning. I'm not attached to the longer names, and would be ok with changing to the shorter names if others prefer it. llvm-svn: 56189
* Allow the fast-path spilling code to attempt folding, but still leaving out ↵Owen Anderson2008-08-191-59/+68
| | | | | | remat and splitting. llvm-svn: 55012
* The fast-path still needs to set kill markers and spill/restore points as ↵Owen Anderson2008-08-191-0/+3
| | | | | | | | | appropriate. With this patch, all of MultiSource/Applications and all of SPEC2000/2006 pass with the SimpleSpiller and this fast-path enabled. llvm-svn: 55000
* Add a flag to enable the fast spilling path.Owen Anderson2008-08-191-0/+7
| | | | llvm-svn: 54958
* Fix a few more bugs:Owen Anderson2008-08-181-3/+8
| | | | | | | 1) Assign stack slots to new temporaries. 2) Don't insert an interval into the return vector more than once. llvm-svn: 54956
* Fix several bugs in the new fast-path:Owen Anderson2008-08-181-16/+16
| | | | | | | | 1) Remove an incorrect assertion. 2) Set the stack slot weight properly. 3) Resize the VirtRegMap when needed. llvm-svn: 54949
* Clients of addIntervalForSpills expect the added intervals to be returned ↵Owen Anderson2008-08-181-0/+9
| | | | | | sorted by starting index. llvm-svn: 54939
* Simplify the fast-patch interval spilling by using ↵Owen Anderson2008-08-181-72/+50
| | | | | | MachineRegisterInfo::reg_iterator. llvm-svn: 54930
* Resurrect some ancient code to add spill ranges without attempting folding, ↵Owen Anderson2008-08-181-0/+105
| | | | | | | | | remat, or splitting. This code has been updated to current APIs in so far as it compiles and, in theory, works, but does not take advantage of recent advancements. For instance, it could be improved by using MachineRegisterInfo::use_iterator. llvm-svn: 54924
* Expunge the last uses of std::map from LiveIntervals.Owen Anderson2008-08-131-13/+13
| | | | llvm-svn: 54766
* Move r2iMap_ over to DenseMap from std::map.Owen Anderson2008-08-131-1/+1
| | | | llvm-svn: 54765
* Make the allocation of LiveIntervals explicit, rather than holding them in ↵Owen Anderson2008-08-131-8/+13
| | | | | | | | the r2iMap_ by value. This will prevent references to them from being invalidated if the map is changed. llvm-svn: 54763
* Oops, didn't mean to commit this.Owen Anderson2008-08-061-0/+2
| | | | llvm-svn: 54425
* Only remap each VNInfo once when doing renumbering.Owen Anderson2008-08-061-13/+15
| | | | llvm-svn: 54420
* Value numbers whose def index is a special sentinel value should not be ↵Owen Anderson2008-07-301-11/+13
| | | | | | remapped. llvm-svn: 54218
* More fixes for corner cases when remapping live range indices.Owen Anderson2008-07-301-10/+10
| | | | llvm-svn: 54186
* Don't decrement the BB remap when we don't need to.Owen Anderson2008-07-291-8/+4
| | | | llvm-svn: 54173
* Fold the useful features of alist and alist_node into ilist, andDan Gohman2008-07-281-1/+1
| | | | | | | | | | | | | | | | a new ilist_node class, and remove them. Unlike alist_node, ilist_node doesn't attempt to manage storage itself, so it avoids the associated problems, including being opaque in gdb. Adjust the Recycler class so that it doesn't depend on alist_node. Also, change it to use explicit Size and Align parameters, allowing it to work when the largest-sized node doesn't have the greatest alignment requirement. Change MachineInstr's MachineMemOperand list from a pool-backed alist to a std::list for now. llvm-svn: 54146
* Fix a typo in a comment.Dan Gohman2008-07-281-1/+1
| | | | llvm-svn: 54136
* Revert my previous patch. In retrospect, this is completely the wrong way ↵Owen Anderson2008-07-251-66/+15
| | | | | | to fix this problem. llvm-svn: 54072
* Special cases are needed in renumbering when dealing with renumbering after ↵Owen Anderson2008-07-251-15/+66
| | | | | | | | | a PHI has been removed. The interval previously defined by the PHI needs to be extended to the beginning of its basic block, and the intervals that were inputs need to be trimmed to the end of their basic blocks. llvm-svn: 54070
* Properly remap live ranges whose end indices are the end of the function.Owen Anderson2008-07-251-4/+13
| | | | llvm-svn: 54061
* Make the remapping of interval indices (particularly ending indices) more ↵Owen Anderson2008-07-251-10/+15
| | | | | | | | | robust. This is tricky business, and will probably take a few more iterations to get the last kinks out of it. llvm-svn: 54043
* Disable the new aggressive remat logic introduced in 54000; it causes someDan Gohman2008-07-251-0/+4
| | | | | | | regressions, such as PR2595. Also, there is a significant code-quality issue in SPEC 464.h264ref and a few others. llvm-svn: 54014
* Enable rematerialization of constants using ↵Dan Gohman2008-07-251-15/+89
| | | | | | | | | | AliasAnalysis::pointsToConstantMemory, and knowledge of PseudoSourceValues. This unfortunately isn't sufficient to allow constants to be rematerialized in PIC mode -- the extra indirection is a complication. llvm-svn: 54000
* Enable the insertion of empty indices into LiveInterals, thereby making ↵Owen Anderson2008-07-231-70/+80
| | | | | | renumbering possible. llvm-svn: 53961
* Fix a compile-time regression introduced by my heuristic-changing patch. I ↵Owen Anderson2008-07-231-1/+1
| | | | | | | | | forgot to multiply the instruction count by a constant factor in a few places, which caused the register allocator to require many more iterations. llvm-svn: 53959
* Change the heuristics used in the coalescer, register allocator, and withinOwen Anderson2008-07-221-1/+5
| | | | | | | live intervals itself to use an instruction count approximation that is not affected by inserting empty indices. llvm-svn: 53937
* Fix a memory leak in LiveIntervalAnalysis.Evan Cheng2008-07-191-4/+8
| | | | llvm-svn: 53779
* Re-introduce LeakDetector support for MachineInstrs and MachineBasicBlocks.Dan Gohman2008-07-171-3/+3
| | | | | | | Fix a leak that this turned up in LowerSubregs.cpp. And, comment a leak in LiveIntervalAnalysis.cpp. llvm-svn: 53746
* Typos.Evan Cheng2008-07-121-2/+2
| | | | llvm-svn: 53504
* Fix PR2536: a nasty spiller bug. If a two-address instruction uses a ↵Evan Cheng2008-07-121-0/+23
| | | | | | | | | | | register but the use portion of its live range is not part of its liveinterval, it must be defined by an implicit_def. In that case, do not spill the use. e.g. 8 %reg1024<def> = IMPLICIT_DEF 12 %reg1024<def> = INSERT_SUBREG %reg1024<kill>, %reg1025, 2 The live range [12, 14) are not part of the r1024 live interval since it's defined by an implicit def. It will not conflicts with live interval of r1025. Now suppose both registers are spilled, you can easily see a situation where both registers are reloaded before the INSERT_SUBREG and both target registers that would overlap. llvm-svn: 53503
* - Change the horrible N^2 isRegReDefinedByTwoAddr. Now callers must supply ↵Evan Cheng2008-07-101-4/+6
| | | | | | | | the operand index of def machineoperand and at most one full scan of non-implicit operands is needed. - Change local register allocator to use the new isRegReDefinedByTwoAddr instead of reinventing the wheel. llvm-svn: 53394
* Pool-allocation for MachineInstrs, MachineBasicBlocks, andDan Gohman2008-07-071-2/+2
| | | | | | | | | | | MachineMemOperands. The pools are owned by MachineFunctions. This drastically reduces the number of calls to malloc/free made during the "Emit" phase of scheduling, as well as later phases in CodeGen. Combined with other changes, this speeds up the "instruction selection" phase of CodeGen by 10% in some cases. llvm-svn: 53212
OpenPOWER on IntegriCloud