1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-29 23:12:55 +01:00
Commit Graph

3742 Commits

Author SHA1 Message Date
Dan Gohman
2c5d31ee81 Minor comment cleanups.
llvm-svn: 37321
2007-05-24 14:36:04 +00:00
Dan Gohman
1b1932dda5 Add explicit qualification for namespace MVT members.
llvm-svn: 37320
2007-05-24 14:33:05 +00:00
Evan Cheng
439bf58dc2 Fix a typo that caused combiner to create mal-formed pre-indexed store where value store is the same as the base pointer.
llvm-svn: 37318
2007-05-24 02:35:39 +00:00
Dale Johannesen
d0b7bbe261 Two tail merging improvements:
When considering blocks with more than 2 predecessors, merge the block with
the largest number of matching insns, rather than the first block found.
Considering that 1 matching insn is enough to show a win for candidates that
already end with a branch.

llvm-svn: 37315
2007-05-23 21:07:20 +00:00
Anton Korobeynikov
0f184e86ab Mark all calls as "could throw", when exceptions are enabled. Emit necessary LP info too. This fixes PR1439
llvm-svn: 37311
2007-05-23 11:08:31 +00:00
Chris Lattner
d540c6429f prevent exponential recursion in isNegatibleForFree
llvm-svn: 37310
2007-05-23 07:35:22 +00:00
Evan Cheng
3f44824179 Preliminary iterative if-conversion support.
llvm-svn: 37309
2007-05-23 07:23:16 +00:00
Dale Johannesen
738f94210c name change requested by review of previous patch
llvm-svn: 37289
2007-05-22 18:31:04 +00:00
Owen Anderson
fd5693fcc4 Silence a warning.
llvm-svn: 37288
2007-05-22 18:13:40 +00:00
Dale Johannesen
fe0fe14411 Make tail merging the default, except on powerPC. There was no prior art
for a target-dependent default with a command-line override; this way
should be generally usable.

llvm-svn: 37285
2007-05-22 17:14:46 +00:00
Evan Cheng
d395b14837 If-convert early exit blocks (returns, etc.); bug fixes, etc.
llvm-svn: 37270
2007-05-21 22:22:58 +00:00
Duncan Sands
05d340bbd1 Only emit one entry in the exception action table for each action, even if
it occurs for multiple landing pads.

llvm-svn: 37267
2007-05-21 18:50:28 +00:00
Chris Lattner
27b469b945 same patch as the previous one, but the symmetric case
llvm-svn: 37249
2007-05-19 00:46:51 +00:00
Chris Lattner
2430b7c4aa Disable the (A == (B-A)) -> 2*A == B xform when the sub has multiple uses (in
this case, the xform introduces an extra operation).  This compiles
PowerPC/compare-duplicate.ll into:

_test:
        subf r2, r3, r4
        cmplw cr0, r2, r3
        bne cr0, LBB1_2 ;F

instead of:

_test:
        slwi r2, r3, 1
        subf r3, r3, r4
        cmplw cr0, r4, r2
        bne cr0, LBB1_2 ;F

This is target independent of course.

llvm-svn: 37246
2007-05-19 00:43:44 +00:00
Evan Cheng
ca7bc898aa Clean up.
llvm-svn: 37237
2007-05-18 19:32:08 +00:00
Evan Cheng
ad9ce3c5b2 Change to depth-first traversal.
llvm-svn: 37236
2007-05-18 19:26:33 +00:00
Dale Johannesen
09b2912ab7 Document an inefficiency in tail merging.
llvm-svn: 37235
2007-05-18 18:46:40 +00:00
Dan Gohman
b593ad9fb0 Qualify calls to getTypeForValueType with MVT:: too.
llvm-svn: 37233
2007-05-18 18:41:29 +00:00
Evan Cheng
1ed1741260 Some restructuring in preparation for most aggressive if-conversion.
llvm-svn: 37231
2007-05-18 18:14:37 +00:00
Dan Gohman
d6a33914fb Qualify several calls to functions in the MVT namespace, for consistency.
llvm-svn: 37230
2007-05-18 17:52:13 +00:00
Evan Cheng
234f07808b Watch out for blocks that end with a return.
llvm-svn: 37227
2007-05-18 17:06:53 +00:00
Evan Cheng
40eecb3d06 If true / false blocks fallthrough before ifcvt, add unconditional branches to ifcvt'd block.
llvm-svn: 37200
2007-05-18 01:55:58 +00:00
Dale Johannesen
ca37006ffb Remove some unneeded branches. (spotted by Evan, thanks)
llvm-svn: 37198
2007-05-18 01:28:58 +00:00
Evan Cheng
c6aa69557a Make use of target specific block size limits; bug fixes.
llvm-svn: 37195
2007-05-18 00:20:58 +00:00
Evan Cheng
b4cc6606d5 Move isSuccessor() offline, change it to use std::find.
llvm-svn: 37190
2007-05-17 23:58:53 +00:00
Chris Lattner
910abfbfd1 disable MaskedValueIsZero, ComputeMaskedBits, and SimplifyDemandedBits for
i128 integers.  The 64-bit masks are not wide enough to represent the results.
These should be converted to APInt someday.

llvm-svn: 37169
2007-05-17 18:19:23 +00:00
Chris Lattner
d441344214 add expand support for ADDC/SUBC/ADDE/SUBE so we can codegen 128-bit add/sub on 32-bit (or less) targets
llvm-svn: 37168
2007-05-17 18:15:41 +00:00
Evan Cheng
3163b77d5e Add target hook to specify block size limit for if-conversion.
llvm-svn: 37134
2007-05-16 23:45:53 +00:00
Dale Johannesen
cc99a6fc32 Don't fold bitconvert(load) for preinc/postdec loads. Likewise stores.
llvm-svn: 37130
2007-05-16 22:45:30 +00:00
Evan Cheng
ab88e91c81 isBlockPredicable() always ignore terminal instructions; add comments.
llvm-svn: 37126
2007-05-16 21:54:37 +00:00
Evan Cheng
6f761adecb Rename M_PREDICATED to M_PREDICABLE; Moved isPredicable() to MachineInstr.
llvm-svn: 37121
2007-05-16 20:56:08 +00:00
Duncan Sands
3296999dbb Output exception call-sites in address order, as required by the unwinding
runtime.

llvm-svn: 37104
2007-05-16 12:12:23 +00:00
Chris Lattner
a18b36cf45 Use a ptr set instead of a linear search to unique TokenFactor operands.
This fixes PR1423

llvm-svn: 37102
2007-05-16 06:37:59 +00:00
Evan Cheng
3fb34a4fdd Devang points out that we need an assertion here.
llvm-svn: 37097
2007-05-16 05:11:10 +00:00
Evan Cheng
a781c7788a Bug fix: should check ABI alignment, not pref. alignment.
llvm-svn: 37094
2007-05-16 02:04:50 +00:00
Evan Cheng
aafca4d4a2 Initial commit of (very basic) if converter.
llvm-svn: 37092
2007-05-16 02:00:57 +00:00
Dale Johannesen
78ae2e1acb Remove extra CFG edges before doing these passes; it makes them happier.
llvm-svn: 37089
2007-05-15 21:19:17 +00:00
Lauro Ramos Venancio
251ea5ab65 Fix an infinite recursion in GetNegatedExpression.
llvm-svn: 37086
2007-05-15 17:05:43 +00:00
Duncan Sands
eddbb07561 The index into the actions table is a ULEB128 not a SLEB128.
llvm-svn: 37084
2007-05-15 13:54:14 +00:00
Reid Spencer
b1ac506546 Un-brain-dead-ify the lowering of part set for the reverse case.
llvm-svn: 37071
2007-05-15 02:26:52 +00:00
Chris Lattner
ba648e0d45 Fix some subtle issues handling immediate values. This fixes
test/CodeGen/ARM/2007-05-14-InlineAsmCstCrash.ll

llvm-svn: 37069
2007-05-15 01:33:58 +00:00
Evan Cheng
eea85a767e Added getNumExplicitOperands and findFirstPredOperand.
llvm-svn: 37064
2007-05-15 01:26:09 +00:00
Chris Lattner
eba1b74df3 implement a simple fneg optimization/propagation thing. This compiles:
CodeGen/PowerPC/fneg.ll into:

_t4:
        fmul f0, f3, f4
        fmadd f1, f1, f2, f0
        blr

instead of:

_t4:
        fneg f0, f3
        fmul f0, f0, f4
        fmsub f1, f1, f2, f0
        blr

llvm-svn: 37054
2007-05-14 22:04:50 +00:00
Evan Cheng
9645be8cf5 Only worry about intervening kill if there are more than one live ranges in the interval.
llvm-svn: 37052
2007-05-14 21:23:51 +00:00
Evan Cheng
e9e48bcf28 Fix for PR1406:
v1 =
r2 = move v1
   = op r2<kill>
...
r2 = move v1
   = op r2<kill>

Clear the first r2 kill if v1 and r2 are joined.

llvm-svn: 37050
2007-05-14 21:10:05 +00:00
Evan Cheng
bb876d7323 When marking a register as being implicitly defined, make sure to clear its partial use info as well.
llvm-svn: 37046
2007-05-14 20:39:18 +00:00
Reid Spencer
afe2314bb0 Give names to the final result values of the part_set computations. This
just aids in readability and debugability of the output. No functional change.

llvm-svn: 37037
2007-05-14 17:21:17 +00:00
Anton Korobeynikov
ad0dd14386 Emit function debug frames in one atom. This will prevent us from generating incorrect assembler in case of both
debug information & exception information presented.

llvm-svn: 37019
2007-05-13 17:30:11 +00:00
Anton Korobeynikov
100919126f Emit multiple common EH frames for multiple (including blank) personality
functions. This partly fixes PR1414: now we're restricted only to one
personality function per eh frame, not per module. Further work on
"multiple personalities" topic needs representative example.

llvm-svn: 37018
2007-05-13 15:42:26 +00:00
Anton Korobeynikov
7daaf71b29 More DWARF-related things cleanup:
1. Fix PR1380
2. Apply Duncan's patch from PR1410
3. Insert workaround for "one personality function per module" as noted in PR1414
4. Emit correct debug frames for x86/linux. This partly fixes DebugInfo/2006-11-06-StackTrace.cpp: stack trace is
shown correctly, but arguments for function on top of stack are displayed incorrectly.

llvm-svn: 37015
2007-05-12 22:36:25 +00:00
Reid Spencer
e7a70b7212 Get the size of auto arrays right, regardless of its changing size.
llvm-svn: 37006
2007-05-12 11:07:40 +00:00
Devang Patel
c611adb5b7 Fix http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049516.html
llvm-svn: 36998
2007-05-11 23:14:43 +00:00
Anton Korobeynikov
5754e67b1a Perform correct actions numbers/sizes computation
llvm-svn: 36988
2007-05-11 08:47:35 +00:00
Anton Korobeynikov
2d2a44b71f Fix action No calculation in multiple-invoke-one-LP mode
llvm-svn: 36987
2007-05-11 08:23:57 +00:00
Dale Johannesen
2f28ffc5fd Do not generate branches to entry block. This fixes several test suite
failures on PPC (can happen only when prologue code is null)

llvm-svn: 36979
2007-05-10 23:59:23 +00:00
Anton Korobeynikov
5ba99a95e8 Ooops. Some debugging stuff :)
llvm-svn: 36978
2007-05-10 22:38:46 +00:00
Anton Korobeynikov
0a29b65e0c Allow multiple invokes per landing pad. This (probably) fixes PR1410.
llvm-svn: 36977
2007-05-10 22:34:59 +00:00
Duncan Sands
828ef09f10 Later computations assume we are aligned at this point.
llvm-svn: 36975
2007-05-10 18:40:24 +00:00
Anton Korobeynikov
a25bb3a0ff TypeIds are indexed by j, not i
llvm-svn: 36974
2007-05-10 15:10:34 +00:00
Dale Johannesen
4f9661688c Make tail merging handle many more cases (all it can, I think).
llvm-svn: 36966
2007-05-10 01:01:49 +00:00
Evan Cheng
649f25dad1 Can't fold the bit_convert is the store is a truncating store.
llvm-svn: 36962
2007-05-09 21:49:47 +00:00
Anton Korobeynikov
4423f1a3fd Do not assert, when case range split metric is zero and JTs are not allowed: just emit binary tree in this case. This
fixes PR1403.

llvm-svn: 36959
2007-05-09 20:07:08 +00:00
Bill Wendling
a5f8cb0805 Change names from RA to something unique to get rid of naming conflicts with
certain linkers...

llvm-svn: 36944
2007-05-08 19:02:46 +00:00
Evan Cheng
855b26b72e Eliminate MarkVirtRegAliveInBlock recursion.
llvm-svn: 36943
2007-05-08 19:00:00 +00:00
Evan Cheng
e18b87868d Forgot a check.
llvm-svn: 36910
2007-05-07 21:36:06 +00:00
Evan Cheng
18d994d6d6 Enable a couple of xforms:
- (store (bitconvert v)) -> (store v) if resultant store does not require
higher alignment
- (bitconvert (load v)) -> (load (bitconvert*)v) if resultant load does not
require higher alignment

llvm-svn: 36908
2007-05-07 21:27:48 +00:00
Dale Johannesen
c884abe112 Handle some non-exit blocks in tail merging.
llvm-svn: 36907
2007-05-07 20:57:21 +00:00
Duncan Sands
20a9ed0e20 Parameter attributes on invoke calls were being lost due to the wrong
attribute index being used.  Fix proposed by Anton Korobeynikov, who
asked me to implement and commit it for him.  This is PR1398.

llvm-svn: 36906
2007-05-07 20:49:28 +00:00
Anton Korobeynikov
3765489a61 Detabify
llvm-svn: 36891
2007-05-06 20:14:21 +00:00
Nick Lewycky
c2306ff5b4 Fix typo in comment.
llvm-svn: 36873
2007-05-06 13:37:16 +00:00
Duncan Sands
4459cd664c Use the personality function that was registered with MMI rather than
hardwiring in the C++ one.

llvm-svn: 36789
2007-05-05 20:27:00 +00:00
Chris Lattner
d1534330ee Propagate alignment/volatility in two places.
Implement support for expanding a bitcast from an illegal vector type to
a legal one (e.g. 4xi32 -> 4xf32 in SSE1).  This fixes PR1371 and
CodeGen/X86/2007-05-05-VecCastExpand.ll

llvm-svn: 36787
2007-05-05 19:39:05 +00:00
Duncan Sands
be371cc319 Spelling fix.
llvm-svn: 36781
2007-05-05 16:32:57 +00:00
Anton Korobeynikov
ce48606d7a Emit sections/directives in the proper order. This fixes PR1376. Also,
some small cleanup was made.

llvm-svn: 36780
2007-05-05 09:04:50 +00:00
Duncan Sands
f2323ca89c A bitcast of a global variable may have been constant folded to a GEP -
handle this case too.

llvm-svn: 36745
2007-05-04 17:12:26 +00:00
Evan Cheng
8c8b6ce116 Don't create indexed load / store with zero offset!
llvm-svn: 36716
2007-05-03 23:52:19 +00:00
Chris Lattner
e95894e9f9 Allow i/s to match (gv+c). This fixes CodeGen/PowerPC/2007-05-03-InlineAsm-S-Constraint.ll
and PR1382

llvm-svn: 36672
2007-05-03 16:54:34 +00:00
Devang Patel
cd45427a87 Drop 'const'
llvm-svn: 36662
2007-05-03 01:11:54 +00:00
Anton Korobeynikov
ca71b83e50 Properly set arguments bitwidth of EHSELECT node
llvm-svn: 36654
2007-05-02 22:15:48 +00:00
Devang Patel
8ee9065162 Use 'static const char' instead of 'static const int'.
Due to darwin gcc bug, one version of darwin linker coalesces
static const int, which defauts PassID based pass identification.

llvm-svn: 36652
2007-05-02 21:39:20 +00:00
Lauro Ramos Venancio
57d03e112b Fix build error.
llvm-svn: 36648
2007-05-02 20:37:47 +00:00
Anton Korobeynikov
6e6bfcaf4d Fix couple of bugs connected with eh info:
1. Correct output offsets on Linux
2. Fix "style" of personality function. It shouldn't be indirect.

llvm-svn: 36633
2007-05-01 22:23:12 +00:00
Devang Patel
38a66bc82e Do not use typeinfo to identify pass in pass manager.
llvm-svn: 36632
2007-05-01 21:15:47 +00:00
Evan Cheng
17f967eb52 If call frame is not part of stack frame and no dynamic alloc, eliminateFrameIndex() must adjust SP offset with size of call frames.
llvm-svn: 36625
2007-05-01 09:01:42 +00:00
Evan Cheng
6dc02c2b07 Forgot about chain result; also UNDEF cannot have multiple values.
llvm-svn: 36622
2007-05-01 08:53:39 +00:00
Nate Begeman
767ee95d29 llvm bug #1350, parts 1, 2, and 3.
llvm-svn: 36618
2007-05-01 05:57:02 +00:00
Evan Cheng
248de25fb5 Under normal circumstances, when a frame pointer is not required, we reserve
argument space for call sites in the function immediately on entry to the
current function. This eliminates the need for add/sub sp brackets around call
sites. However, this is not always a good idea. If the "call frame" is large and
the target load / store instructions have small immediate field to encode sp
offset, this can cause poor codegen. In the worst case, this can make it
impossible to scavenge a register if the reserved spill slot is pushed too far
apart from sp / fp.

llvm-svn: 36607
2007-05-01 00:52:08 +00:00
Evan Cheng
fe933cd6ca * Only turn a load to UNDEF if all of its outputs have no uses (indexed loads
produce two results.)
* Do not touch volatile loads.

llvm-svn: 36604
2007-05-01 00:38:21 +00:00
Chris Lattner
8da69f853c Fix PR1228 and CodeGen/Generic/2007-04-30-LandingPadBranchFolding.ll
llvm-svn: 36602
2007-04-30 23:35:00 +00:00
Chris Lattner
c7f45f4572 print isLandingPad() for MBBs
llvm-svn: 36600
2007-04-30 23:12:53 +00:00
Chris Lattner
3b2c717e6a Continue refactoring inline asm code. If there is an earlyclobber output
register, preallocate all input registers and the early clobbered output.

This fixes PR1357 and CodeGen/PowerPC/2007-04-30-InlineAsmEarlyClobber.ll

llvm-svn: 36599
2007-04-30 21:11:17 +00:00
Evan Cheng
5039cf6382 Updates.
llvm-svn: 36594
2007-04-30 18:42:09 +00:00
Chris Lattner
4ac6ba02d1 refactor GetRegistersForValue to take OpInfo as an argument instead of various
pieces of it.  No functionality change.

llvm-svn: 36592
2007-04-30 17:29:31 +00:00
Chris Lattner
e3f7f2afcb refactor some code, no functionality change
llvm-svn: 36590
2007-04-30 17:16:27 +00:00
Chris Lattner
9f10fc6b4c Clean up multi-line asam string printing. Instead of printing:
# InlineAsm Start
        subfc r3,r5,r4
                subfze r4,r3
        # InlineAsm End

print:

        # InlineAsm Start
        subfc r3,r5,r4
        subfze r4,r3
        # InlineAsm End

llvm-svn: 36589
2007-04-30 17:00:18 +00:00
Chris Lattner
087a1aaaab generalize aggregate handling
llvm-svn: 36568
2007-04-29 18:58:03 +00:00
Anton Korobeynikov
cb2004f82c Implement review feedback
llvm-svn: 36564
2007-04-29 18:02:48 +00:00
Chris Lattner
85e1cac6b7 memory operands that have a direct operand should have their stores created
before the copies into physregs are done.  This avoids having flag operands
skip the store, causing cycles in the dag at sched time.  This fixes infinite
loops on these tests:

test/CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll for PR1308
test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll
test/CodeGen/PowerPC/2007-01-31-InlineAsmAddrMode.ll
test/CodeGen/X86/2006-07-12-InlineAsmQConstraint.ll for PR828

llvm-svn: 36547
2007-04-28 21:12:06 +00:00
Chris Lattner
97d396d928 eliminate more redundant constraint type analysis
llvm-svn: 36546
2007-04-28 21:03:16 +00:00
Chris Lattner
4153538017 merge constraint type analysis stuff together.
llvm-svn: 36545
2007-04-28 21:01:43 +00:00