Commit Graph

95 Commits

Author SHA1 Message Date
Evan Cheng 6ad7da96fe - Change TargetInstrInfo::reMaterialize to pass in TargetRegisterInfo.
- If destination is a physical register and it has a subreg index, use the
  sub-register instead.
This fixes PR5423.

llvm-svn: 88745
2009-11-14 02:55:43 +00:00
Lang Hames 6b7233afc6 Added an API to the SlotIndexes pass to allow new instructions to be inserted into the numbering.
PreAllocSplitting is now using this API to insert code.

llvm-svn: 88725
2009-11-14 00:02:51 +00:00
David Greene 1fbe054450 Add a bool flag to StackObjects telling whether they reference spill
slots.  The AsmPrinter will use this information to determine whether to
print a spill/reload comment.

Remove default argument values.  It's too easy to pass a wrong argument
value when multiple arguments have default values.  Make everything
explicit to trap bugs early.

Update all targets to adhere to the new interfaces..

llvm-svn: 87022
2009-11-12 20:49:22 +00:00
Lang Hames bff25cb044 Fixed an iteration condition in PreAllocSplitting. This should fix some miscompilations casued by PreAllocSplitting.
llvm-svn: 86919
2009-11-12 01:24:08 +00:00
Jeffrey Yasskin b40d3f76a0 Fix DenseMap iterator constness.
This patch forbids implicit conversion of DenseMap::const_iterator to
DenseMap::iterator which was possible because DenseMapIterator inherited
(publicly) from DenseMapConstIterator. Conversion the other way around is now
allowed as one may expect.

The template DenseMapConstIterator is removed and the template parameter
IsConst which specifies whether the iterator is constant is added to
DenseMapIterator.

Actually IsConst parameter is not necessary since the constness can be
determined from KeyT but this is not relevant to the fix and can be addressed
later.

Patch by Victor Zverovich!

llvm-svn: 86636
2009-11-10 01:02:17 +00:00
Evan Cheng ce29331e10 80 col.
llvm-svn: 86521
2009-11-09 06:49:22 +00:00
Lang Hames 05fb9637f6 The Indexes Patch.
This introduces a new pass, SlotIndexes, which is responsible for numbering
instructions for register allocation (and other clients). SlotIndexes numbering
is designed to match the existing scheme, so this patch should not cause any
changes in the generated code.

For consistency, and to avoid naming confusion, LiveIndex has been renamed
SlotIndex.

The processImplicitDefs method of the LiveIntervals analysis has been moved
into its own pass so that it can be run prior to SlotIndexes. This was
necessary to match the existing numbering scheme.

llvm-svn: 85979
2009-11-03 23:52:08 +00:00
Nick Lewycky 02d5f77d26 Remove VISIBILITY_HIDDEN from class/struct found inside anonymous namespaces.
Chris claims we should never have visibility_hidden inside any .cpp file but
that's still not true even after this commit.

llvm-svn: 85042
2009-10-25 06:33:48 +00:00
Lang Hames 16edcd724a Oops. Backing out 84681 - needs to wait for the indexing patch.
llvm-svn: 84682
2009-10-20 21:28:22 +00:00
Lang Hames d8938d1ca8 Added some debugging output to pre-alloc splitting.
llvm-svn: 84681
2009-10-20 21:25:13 +00:00
Evan Cheng 4729191bb2 Distinquish stack slots from other stack objects. They (and fixed objects) get FixedStack PseudoSourceValues.
llvm-svn: 84326
2009-10-17 09:20:14 +00:00
Evan Cheng be27d61cbc Reset kill markers after live interval is reconstructed.
llvm-svn: 83608
2009-10-09 01:17:11 +00:00
Evan Cheng e23984fbbe Remove code that makes no sense.
llvm-svn: 83589
2009-10-08 22:42:35 +00:00
Lang Hames a7780905e7 Oops. Renamed remaining MachineInstrIndex references.
llvm-svn: 83255
2009-10-03 04:31:31 +00:00
Lang Hames 920301ecc5 Renamed MachineInstrIndex to LiveIndex.
llvm-svn: 83254
2009-10-03 04:21:37 +00:00
Lang Hames 6b81770040 Removed static qualifier from a few index related methods. These methods may require a LiveIntervals instance in future.
llvm-svn: 81374
2009-09-09 20:14:17 +00:00
Lang Hames 3fffe62177 Replaces uses of unsigned for indexes in LiveInterval and VNInfo with
a new class, MachineInstrIndex, which hides arithmetic details from
most clients. This is a step towards allowing the register allocator
to update/insert code during allocation.

llvm-svn: 81040
2009-09-04 20:41:11 +00:00
Chris Lattner 1362602eb2 Change Pass::print to take a raw ostream instead of std::ostream,
update all code that this affects.

llvm-svn: 79830
2009-08-23 06:03:38 +00:00
Lang Hames 3b90d973b0 Modified VNInfo. The "copy" member is now a union which holds the copy for a register interval, or the defining register for a stack interval. Access is via getCopy/setCopy and getReg/setReg.
llvm-svn: 78620
2009-08-10 23:43:28 +00:00
Dan Gohman 0402315d41 Use setPreservesAll and setPreservesCFG in CodeGen passes.
llvm-svn: 77754
2009-07-31 23:37:33 +00:00
Evan Cheng 84517443ca Let callers decide the sub-register index on the def operand of rematerialized instructions.
Avoid remat'ing instructions whose def have sub-register indices for now. It's just really really hard to get all the cases right.

llvm-svn: 75900
2009-07-16 09:20:10 +00:00
Torok Edwin 08954aa4e1 Fix assert(0) conversion, as suggested by Chris.
llvm-svn: 75423
2009-07-12 20:07:01 +00:00
Torok Edwin ccb29cd290 Convert more assert(0)+abort() -> LLVM_UNREACHABLE,
and abort()/exit() -> llvm_report_error().

llvm-svn: 75363
2009-07-11 13:10:19 +00:00
Lang Hames dab7b06de9 Improved tracking of value number kills. VN kills are now represented
as an (index,bool) pair. The bool flag records whether the kill is a
PHI kill or not. This code will be used to enable splitting of live
intervals containing PHI-kills.

A slight change to live interval weights introduced an extra spill
into lsr-code-insertion (outside the critical sections). The test 
condition has been updated to reflect this.

llvm-svn: 75097
2009-07-09 03:57:02 +00:00
Lang Hames 16cab1935e VNInfo cleanup.
llvm-svn: 73634
2009-06-17 21:01:20 +00:00
Evan Cheng 210fc62a91 In some rare cases, the register allocator can spill registers but end up not utilizing registers at all. The fundamental problem is linearscan's backtracking can end up freeing more than one allocated registers. However, reloads and restores might be folded into uses / defs and freed registers might not be used at all.
VirtRegMap keeps track of allocations so it knows what's not used. As a horrible hack, the stack coloring can color spill slots with *free* registers. That is, it replace reload and spills with copies from and to the free register. It unfold instructions that load and store the spill slot and replace them with register using variants.

Not yet enabled. This is part 1. More coming.

llvm-svn: 70787
2009-05-03 18:32:42 +00:00
Bob Wilson 51856173c8 Fix pr3954. The register scavenger asserts for inline assembly with
register destinations that are tied to source operands.  The
TargetInstrDescr::findTiedToSrcOperand method silently fails for inline
assembly.  The existing MachineInstr::isRegReDefinedByTwoAddr was very
close to doing what is needed, so this revision makes a few changes to
that method and also renames it to isRegTiedToUseOperand (for consistency
with the very similar isRegTiedToDefOperand and because it handles both
two-address instructions and inline assembly with tied registers).

llvm-svn: 68714
2009-04-09 17:16:43 +00:00
Owen Anderson 4486c1fac0 Remove the "fast" cases for spill and restore point determination, as these were subtlely wrong in obscure cases. Patch the testcase
to account for this change.

llvm-svn: 68093
2009-03-31 08:27:09 +00:00
Owen Anderson 0422efc994 Give the pre-alloc splitter access to the VirtRegMap. It doesn't do anything
useful with it at the moment, but it will in the future.

llvm-svn: 67012
2009-03-14 21:40:05 +00:00
Owen Anderson 380b3838f6 (Hopefully) silence a warning.
llvm-svn: 66158
2009-03-05 08:23:20 +00:00
Owen Anderson ad503987a8 Be more careful about choosing restore points when doing restore folding. This fixes some subtle miscompilations.
llvm-svn: 66147
2009-03-05 07:19:18 +00:00
Owen Anderson 0dedc114de Add a restore folder, which shaves a dozen or so machineinstrs off oggenc. Update a testcase to check this.
llvm-svn: 66029
2009-03-04 08:52:31 +00:00
Owen Anderson f13820148b Fix a crash in the pre-alloc splitter exposed by recent codegen changes.
llvm-svn: 65121
2009-02-20 10:02:23 +00:00
Evan Cheng b5f0ec3eb7 Add TargetInstrInfo::isSafeToMoveRegisterClassDefs. It returns true if it's safe to move an instruction which defines a value in the register class. Replace pre-splitting specific IgnoreRegisterClassBarriers with this new hook.
llvm-svn: 63936
2009-02-06 17:17:30 +00:00
Owen Anderson 7411523ee6 Pre-alloc splitting needs to be more careful to avoid inserting spills/restores
between call frame setup/restore points.  Unfortunately, this regresses 
code size a bit, but at least it's correct now!

llvm-svn: 63837
2009-02-05 05:58:41 +00:00
Owen Anderson 4eda2cbe5f MergeValueInto is too smart: it might choose to do the merge the opposite direction.
Live interval reconstruction needs to account for this, and scour its maps to
prevent dangling references.

llvm-svn: 63558
2009-02-02 22:42:01 +00:00
Evan Cheng 03f862579c Refactor PerformPHIConstruction, no functionality changes.
llvm-svn: 63536
2009-02-02 18:33:18 +00:00
Owen Anderson 19616c08a0 Fix test failures causes by my previous commit.
llvm-svn: 63492
2009-02-01 08:41:54 +00:00
Owen Anderson 66838e9a27 Fix an issue in PHI construction that was exposed by GCC 4.2 producing a different set iteration order for the reg_iterator.
llvm-svn: 63490
2009-02-01 07:06:00 +00:00
Owen Anderson bf77d2eb9d Correct the algorithms for choosing spill and restore points so that we don't try to insert loads/stores between call frame setup and the actual call.
This fixes the last known failure for the pre-alloc-splitter.

llvm-svn: 63339
2009-01-29 22:13:06 +00:00
Owen Anderson c0924a0632 Fix an issue where restores could be inserted after a terminator instruction,
and an iterator invalidation issue.

FreeBench/pifft no longer miscompiles with these fixes!

llvm-svn: 63293
2009-01-29 08:22:06 +00:00
Owen Anderson a590802ed6 Comments are good. :-)
llvm-svn: 63276
2009-01-29 05:41:02 +00:00
Owen Anderson b05ce78c51 Add support for aggressive load-use-store folding. This takes care of the
vast majority of code size regressions introduced by pre-alloc-splitting.

llvm-svn: 63274
2009-01-29 05:28:55 +00:00
Owen Anderson 431053ed2c Make the pre-split-limit option more useful by using a per-function counter.
llvm-svn: 63091
2009-01-27 05:01:15 +00:00
Owen Anderson e84c3f261e Reapply r63025 and r63026, with fixes for the failing testcases.
llvm-svn: 63049
2009-01-26 21:57:31 +00:00
Bill Wendling e7fa3c37c6 Also revert r63206
llvm-svn: 63041
2009-01-26 21:30:17 +00:00
Bill Wendling a39c83d118 Temporarily revert r63025 until the testsuite failures can be fixed.
llvm-svn: 63040
2009-01-26 21:27:03 +00:00
Owen Anderson 5e235d3a7e Get rid of a bunch of dead code now that interval reconstruction is enabled.
llvm-svn: 63026
2009-01-26 19:18:06 +00:00
Owen Anderson feffdec244 Fix an issue where LiveIntervals was trying to be smart about removing kill
markers, and ended up foiling the interval reconstruction.

This allows us to turn on reconstruction in the pre alloc splitter, which
fixes a number of miscompilations.

llvm-svn: 63025
2009-01-26 19:12:06 +00:00
Evan Cheng 72287a599f Silence a bogus compiler warning.
llvm-svn: 63021
2009-01-26 18:33:51 +00:00