Evan Cheng [Mon, 22 Feb 2010 23:37:48 +0000 (23:37 +0000)]
These should not have been committed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96827
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 23:34:12 +0000 (23:34 +0000)]
no need to run llvm-as here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96826
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 22 Feb 2010 23:34:00 +0000 (23:34 +0000)]
Instcombine constant folding can normalize gep with negative index to index with large offset. When instcombine objsize checking transformation sees these geps where the offset seemingly point out of bound, it should just return "i don't know" rather than asserting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96825
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 23:33:44 +0000 (23:33 +0000)]
Change ComplexPattern handling to push the node being matched as
well as the operands produced when the pattern is matched. This
allows CheckSame to work correctly when matching replicated
names involving ComplexPatterns. This fixes a bunch of MSP430
failures, we're down to 13 failures, two of which are
due to a sched bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96824
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Mon, 22 Feb 2010 23:10:38 +0000 (23:10 +0000)]
Updated version of r96634 (which was reverted due to failing 176.gcc and
126.gcc nightly tests. These failures uncovered latent bugs that machine DCE
could remove one half of a stack adjust down/up pair, causing PEI to assert.
This update fixes that, and the tests now pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96822
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 23:07:52 +0000 (23:07 +0000)]
Add a test for canonicalizing ConstantExpr operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96820
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Mon, 22 Feb 2010 22:54:55 +0000 (22:54 +0000)]
Clean up a bit and fix for when SPAdj != 0
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96818
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Mon, 22 Feb 2010 22:47:46 +0000 (22:47 +0000)]
The predicate index isn't fixed, so scan for it to make sure we get the proper
value.
Thumb2 uses the tADJCALLSTACK* instructions, and doesn't need t2 versions, so
remove the FIXME entry.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96817
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 22:43:23 +0000 (22:43 +0000)]
Canonicalize ConstantInts to the right operand of commutative
operators.
The test difference is just due to the multiplication operands
being commuted (and thus requiring a more elaborate match). In
optimized code, that expression would be folded.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96816
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 22:37:11 +0000 (22:37 +0000)]
expand my hack to work with nodes that have flags but no chains and the
isel doesn't know the correct # results. This fixes 8 codegen tests,
down to 22 failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96815
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 22:30:37 +0000 (22:30 +0000)]
add a new CheckMultiOpcode opcode for checking that a node
has one of the list of acceptable opcodes for a complex
pattern. This fixes 4 regtest failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96814
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 22:18:05 +0000 (22:18 +0000)]
When matching patterns that have a complex pattern as their root, make
sure to only run the complex pattern on nodes where the target opts in.
This patch only handles targets with one opcode specified so far, but
fixes 16 failures, only 34 left.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96813
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 22:15:39 +0000 (22:15 +0000)]
add some debug hooks for tracking the behavior of the isel.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96812
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 22:15:05 +0000 (22:15 +0000)]
remove dupes now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96811
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 22:14:47 +0000 (22:14 +0000)]
move #includes earlier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96810
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Mon, 22 Feb 2010 22:08:57 +0000 (22:08 +0000)]
MC/Mach-O: Remove non-sensical comment, and add a missing AddValueSymbols call.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96809
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 22:07:27 +0000 (22:07 +0000)]
Minor formatting cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96808
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 22:05:18 +0000 (22:05 +0000)]
Use Instruction::isCommutative instead of duplicating it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96807
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Mon, 22 Feb 2010 21:50:40 +0000 (21:50 +0000)]
Added SEL, SXTB16, SXTAB16, UXTAB16, SMMULR, SMMLAR, SMMLSR, SMUAD, and SMUSD,
for disassembly only.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96806
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Mon, 22 Feb 2010 21:39:41 +0000 (21:39 +0000)]
Erase deleted instructions from GVN's ValueTable. This fixes assertion
failures from ValueTable::verifyRemoved() when using -debug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96805
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 18:53:26 +0000 (18:53 +0000)]
Actually enable the -enable-unsafe-fp-math tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96796
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Mon, 22 Feb 2010 18:50:54 +0000 (18:50 +0000)]
Added a bunch of instructions for disassembly only:
o signed/unsigned add/subtract
o signed/unsigned halving add/subtract
o unsigned sum of absolute difference [and accumulate]
o signed/unsigned saturate
o signed multiply accumulate/subtract [long] dual
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96795
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Mon, 22 Feb 2010 16:18:09 +0000 (16:18 +0000)]
Mark the return address stack slot as mutable when moving the return address
during a tail call. A parameter might overwrite this stack slot during the tail
call.
The sequence during a tail call is:
1.) load return address to temp reg
2.) move parameters (might involve storing to return address stack slot)
3.) store return address to new location from temp reg
If the stack location is marked immutable CodeGen can colocate load (1) with the
store (3).
This fixes bug 6225.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96783
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Mon, 22 Feb 2010 05:55:32 +0000 (05:55 +0000)]
LLVMC/MultiplePluginPriorities.td: Generally XFAIL this test for now, it is
still failing during (one) llvm-gcc powerpc build, and is also failing on my
x86_64-apple-darwin10.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96781
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 04:11:59 +0000 (04:11 +0000)]
Remove unused variables and parameters.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96780
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 04:10:52 +0000 (04:10 +0000)]
Fix various doxygen warnings.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96779
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 04:09:26 +0000 (04:09 +0000)]
Fix a typo in a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96778
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 04:06:03 +0000 (04:06 +0000)]
Constant-fold certain comparisons with infinity and negative infinity.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96777
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 04:04:24 +0000 (04:04 +0000)]
Rename a variable to avoid a -Wshadow warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96776
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 04:03:39 +0000 (04:03 +0000)]
Remove the logic for reasoning about NaNs from the code that forms
SSE min and max instructions. The real thing this code needs to be
concerned about is negative zero.
Update the sse-minmax.ll test accordingly, and add tests for
-enable-unsafe-fp-math mode as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96775
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 03:59:54 +0000 (03:59 +0000)]
When emitting an instruction which depends on both a post-incremented
induction variable value and a loop-variant value, don't force the
insert position to be at the post-increment position, because it may
not be dominated by the loop-variant value. This fixes a
use-before-def problem noticed on PPC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96774
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 22 Feb 2010 02:07:36 +0000 (02:07 +0000)]
This cast<Instruction> is unnecessary.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96771
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 00:47:38 +0000 (00:47 +0000)]
enhance my hack for flags handling, this allows us to pass
CodeGen/X86/pr2182.ll. Down to 13 x86 failures out of ~1100
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96770
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 22 Feb 2010 00:28:59 +0000 (00:28 +0000)]
fix an incorrect VT: eflags is always i32. The bug was causing us to
create an X86ISD::Cmp node with result type i64 on the
CodeGen/X86/shift-i256.ll testcase and the new isel was assert on it
downstream.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96768
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 23:54:05 +0000 (23:54 +0000)]
fix most of the failures in the x86 suite by handling multiple
result nodes correctly. Note that this includes a horrible hack
in DAGISelHeader which cannot be fixed reasonably without
eliminating (parallel) from input patterns. That, in turn,
can't be done until we support writing multiple result patterns
for the X86and_flag and related multiple-result nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96767
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sun, 21 Feb 2010 21:54:14 +0000 (21:54 +0000)]
MC/X86: Add stub AsmBackend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96763
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sun, 21 Feb 2010 21:53:53 +0000 (21:53 +0000)]
MC: Sketch registry support for target specific assembler backends.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96762
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sun, 21 Feb 2010 21:53:37 +0000 (21:53 +0000)]
Formatting tweaks (trailing whitespace, ordering, comments).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96761
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 20:57:28 +0000 (20:57 +0000)]
fix an odd thinko in an assertion, all arm tests pass now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96758
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 20:53:45 +0000 (20:53 +0000)]
Always emit register class id references as i32 like
DAGISelEmitter does. This fixes 11 arm failures (8
left).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96757
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 20:33:20 +0000 (20:33 +0000)]
add some no-unwinds, other minor cleanups.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96756
91177308-0d34-0410-b5e6-
96231b3b80d8
Anton Korobeynikov [Sun, 21 Feb 2010 20:28:15 +0000 (20:28 +0000)]
It turned out that we failed to emit proper symbol stubs on non-x86/darwin for ages (we emitted a reference to a stub, but no stub was emitted). The code inside x86-32/macho target objfile lowering should actually be the generic one - move it there.
This (I really, really hope) should fix EH issues on ppc/darwin
and arm/darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96755
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 20:15:25 +0000 (20:15 +0000)]
speculatively teach OPC_CheckValueType and OPC_EmitNode to handle
MVT::iPTR.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96753
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 20:10:43 +0000 (20:10 +0000)]
teach OPC_CheckType to handle MVT::iPTR, down to 2 ppc failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96752
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 20:02:15 +0000 (20:02 +0000)]
make this check a bit more generous, it may be outliving its
utility. Down to 6 ppc failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96751
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 19:35:07 +0000 (19:35 +0000)]
Relax an assertion a bit. We allow replacing things like
<4 x i32> with <4 x float> values if they end up the same
register class. This gets us up to 231 passes on the ppc
tests (only 7 fails).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96750
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 19:27:38 +0000 (19:27 +0000)]
add a triple so that this doesn't fail due to linux/ppc register printing
syntax.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96748
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 19:22:06 +0000 (19:22 +0000)]
Sort the patterns before adding them to the FA so that we get the
least cost matches. This gets us from 195 -> 208 passes on the ppc codegen tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96747
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Sun, 21 Feb 2010 19:15:19 +0000 (19:15 +0000)]
Remove a bunch of duplicated code, where there was one version taking a std::ostream
and another taking a raw_ostream, but otherwise identical. Use raw_ostream everywhere.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96746
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 18:53:28 +0000 (18:53 +0000)]
filecheckize and add nouwinds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96745
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sun, 21 Feb 2010 18:51:48 +0000 (18:51 +0000)]
Enable assertion to detect cyclic valno references.
This changes the stack overflow in PR6363 to an assertion failure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96744
91177308-0d34-0410-b5e6-
96231b3b80d8
Anton Korobeynikov [Sun, 21 Feb 2010 12:28:58 +0000 (12:28 +0000)]
IT turns out that during jumpless setcc lowering eq and ne were swapped.
This fixes PR6348
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96734
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 07:55:48 +0000 (07:55 +0000)]
when a match fails and we have to rollback, make sure to keep 'N' in
sync with the top of stack. This fixes a bunch of failures on larger
testcases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96732
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 07:19:06 +0000 (07:19 +0000)]
handle optional in flags that aren't present.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96731
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 07:16:41 +0000 (07:16 +0000)]
emit table indexes before each row so that it is debuggable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96730
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 06:58:27 +0000 (06:58 +0000)]
fix two bugs in OPC_EmitRegister, which makes ppc happier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96728
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 06:44:29 +0000 (06:44 +0000)]
fix a table size miscomputation, target opcodes are 2 bytes.
With this, the matcher actually works reasonably well, but
crashes on larger examples in the scheduler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96727
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 06:30:04 +0000 (06:30 +0000)]
emit to the right streams, to avoid emitting the push
body before the push.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96726
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 06:03:56 +0000 (06:03 +0000)]
oops don't turn this on for everyone yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96725
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 06:03:07 +0000 (06:03 +0000)]
implement the last known missing feature: updating uses of results
of the matched pattern to use the newly created node results. Onto
the "making it actually work" phase!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96724
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 04:53:34 +0000 (04:53 +0000)]
fix and un-xfail X86/vec_ss_load_fold.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96720
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Sun, 21 Feb 2010 04:42:01 +0000 (04:42 +0000)]
Undo r96654. The printing of ARM shift instructions in canonical forms can be
handled in ARMInstPrinter.cpp.
And added PLD/PLDW/PLI (Preload Data/Instruction) for disassembly only.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96719
91177308-0d34-0410-b5e6-
96231b3b80d8
Charles Davis [Sun, 21 Feb 2010 04:26:06 +0000 (04:26 +0000)]
Reduce size of 'StackAlignment' field from 5 to 3 bits. Seriously, who needs a
2GB-aligned stack anyway? 256 bytes is plenty. Requested by Chris.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96718
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 03:24:41 +0000 (03:24 +0000)]
temporarily disable this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96717
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 03:22:59 +0000 (03:22 +0000)]
Lots of improvements to the new dagisel emitter. This gets it to
the point where it is to the 95% feature complete mark, it just
needs result updating to be done (then testing, optimization
etc).
More specificallly, this adds support for chain and flag handling
on the result nodes, support for sdnodexforms, support for variadic
nodes, memrefs, pinned physreg inputs, and probably lots of other
stuff.
In the old DAGISelEmitter, this deletes the dead code related to
OperatorMap, cleans up a variety of dead stuff handling "implicit
remapping" from things like globaladdr -> targetglobaladdr (which
is no longer used because globaladdr always needs to be legalized),
and some minor formatting fixes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96716
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 03:17:59 +0000 (03:17 +0000)]
rename SelectScalarSSELoad -> SelectScalarSSELoadXXX and rewrite
it to follow the mode needed by the new isel. Instead of returning
the input and output chains, it just returns the (currently only one,
which is a silly limitation) node that has input and output chains.
Since we want the old thing to still work, add a new
SelectScalarSSELoad to emulate the old interface. The XXX suffix
and the wrapper will eventually go away.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96715
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 03:15:11 +0000 (03:15 +0000)]
add a hook so that the new isel can run SDNodeXForms.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96714
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 03:13:10 +0000 (03:13 +0000)]
Eliminate some uses of immAllOnes, just use -1, it does
the same thing and is more efficient for the matcher.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96713
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 21 Feb 2010 03:12:16 +0000 (03:12 +0000)]
Eliminate some uses of immAllOnes, just use -1, it does
the same thing and is more efficient for the matcher.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96712
91177308-0d34-0410-b5e6-
96231b3b80d8
Tanya Lattner [Sun, 21 Feb 2010 03:01:05 +0000 (03:01 +0000)]
Test commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96711
91177308-0d34-0410-b5e6-
96231b3b80d8
Tanya Lattner [Sun, 21 Feb 2010 02:59:05 +0000 (02:59 +0000)]
Test commit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96710
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 19 Feb 2010 19:35:48 +0000 (19:35 +0000)]
Rename getSDiv to getExactSDiv to reflect its behavior in cases where
the division would have a remainder.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96693
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 19 Feb 2010 19:32:49 +0000 (19:32 +0000)]
Check for overflow when scaling up an add or an addrec for
scaled reuse.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96692
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Fri, 19 Feb 2010 18:59:53 +0000 (18:59 +0000)]
Revert 96634. It causes assertion failures for 126.gcc and 176.gcc in
the armv6 nightly tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96691
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 19 Feb 2010 18:49:22 +0000 (18:49 +0000)]
Add a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96688
91177308-0d34-0410-b5e6-
96231b3b80d8
Charles Davis [Fri, 19 Feb 2010 18:17:13 +0000 (18:17 +0000)]
Add support for the 'alignstack' attribute to the x86 backend. Fixes PR5254.
Also, FileCheck'ize a test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96686
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 19 Feb 2010 18:12:07 +0000 (18:12 +0000)]
Teach ScalarEvolution how to compute a tripcount for a loop with
true or false as its exit condition. These are usually eliminated by
SimplifyCFG, but the may be left around during a pass which wishes
to preserve the CFG.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96683
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Fri, 19 Feb 2010 17:10:59 +0000 (17:10 +0000)]
Revert Anton's most recent EH patch (r96637), since it breaks a lot of
ARM and Thumb tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96680
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Fri, 19 Feb 2010 11:30:41 +0000 (11:30 +0000)]
Revert commits 96556 and 96640, because commit 96556 breaks the
dragonegg self-host build. I reverted 96640 in order to revert
96556 (96640 goes on top of 96556), but it also looks like with
both of them applied the breakage happens even earlier. The
symptom of the 96556 miscompile is the following crash:
llvm[3]: Compiling AlphaISelLowering.cpp for Release build
cc1plus: /home/duncan/tmp/tmp/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:4982: void llvm::SelectionDAG::ReplaceAllUsesWith(llvm::SDNode*, llvm::SDNode*, llvm::SelectionDAG::DAGUpdateListener*): Assertion `(!From->hasAnyUseOfValue(i) || From->getValueType(i) == To->getValueType(i)) && "Cannot use this version of ReplaceAllUsesWith!"' failed.
Stack dump:
0. Running pass 'X86 DAG->DAG Instruction Selection' on function '@_ZN4llvm19AlphaTargetLowering14LowerOperationENS_7SDValueERNS_12SelectionDAGE'
g++: Internal error: Aborted (program cc1plus)
This occurs when building LLVM using LLVM built by LLVM (via
dragonegg). Probably LLVM has miscompiled itself, though it
may have miscompiled GCC and/or dragonegg itself: at this point
of the self-host build, all of GCC, LLVM and dragonegg were built
using LLVM. Unfortunately this kind of thing is extremely hard
to debug, and while I did rummage around a bit I didn't find any
smoking guns, aka obviously miscompiled code.
Found by bisection.
r96556 | evancheng | 2010-02-18 03:13:50 +0100 (Thu, 18 Feb 2010) | 5 lines
Some dag combiner goodness:
Transform br (xor (x, y)) -> br (x != y)
Transform br (xor (xor (x,y), 1)) -> br (x == y)
Also normalize (and (X, 1) == / != 1 -> (and (X, 1)) != / == 0 to match to "test on x86" and "tst on arm"
r96640 | evancheng | 2010-02-19 01:34:39 +0100 (Fri, 19 Feb 2010) | 16 lines
Transform (xor (setcc), (setcc)) == / != 1 to
(xor (setcc), (setcc)) != / == 1.
e.g. On x86_64
%0 = icmp eq i32 %x, 0
%1 = icmp eq i32 %y, 0
%2 = xor i1 %1, %0
br i1 %2, label %bb, label %return
=>
testl %edi, %edi
sete %al
testl %esi, %esi
sete %cl
cmpb %al, %cl
je LBB1_2
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96672
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Fri, 19 Feb 2010 09:18:53 +0000 (09:18 +0000)]
Correct LastPrimitiveTyID: MetadataType is a primitive type.
This change probably has no functional effect.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96669
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 19 Feb 2010 07:49:56 +0000 (07:49 +0000)]
add emitter support for integer constants and simple physreg references.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96663
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Fri, 19 Feb 2010 07:14:22 +0000 (07:14 +0000)]
recommit 96626, evidence that it broke things appears
to be spurious
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96662
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 19 Feb 2010 07:02:20 +0000 (07:02 +0000)]
fix this to work more determinstically, patch by Thomas Veith!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96661
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Fri, 19 Feb 2010 02:58:33 +0000 (02:58 +0000)]
Test case for r96656.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96657
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Fri, 19 Feb 2010 02:12:06 +0000 (02:12 +0000)]
Added entries for ASR, LSL, LSR, ROR, and RRX so that the disassembler prints
out the canonical form (A8.6.98) instead of the pseudo-instruction as provided
via MOVs.
DBG_ARM_DISASM=YES llvm-mc -triple=arm-unknown-unknown --disassemble
0xc0 0x00 0xa0 0xe1
Opcode=29 Name=ASR Format=ARM_FORMAT_LDMISCFRM
31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
-------------------------------------------------------------------------------------------------
| 1: 1: 1: 0| 0: 0: 0: 1| 1: 0: 1: 0| 0: 0: 0: 0| 0: 0: 0: 0| 0: 0: 0: 0| 1: 1: 0: 0| 0: 0: 0: 0|
-------------------------------------------------------------------------------------------------
asr r0, r0, #1
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96654
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Fri, 19 Feb 2010 01:54:37 +0000 (01:54 +0000)]
Revert 96626, which causes build failure on ppc Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96653
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 19 Feb 2010 00:34:39 +0000 (00:34 +0000)]
Transform (xor (setcc), (setcc)) == / != 1 to
(xor (setcc), (setcc)) != / == 1.
e.g. On x86_64
%0 = icmp eq i32 %x, 0
%1 = icmp eq i32 %y, 0
%2 = xor i1 %1, %0
br i1 %2, label %bb, label %return
=>
testl %edi, %edi
sete %al
testl %esi, %esi
sete %cl
cmpb %al, %cl
je LBB1_2
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96640
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 19 Feb 2010 00:33:13 +0000 (00:33 +0000)]
I confused myself, temporaries will be recorded right along with other inputs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96639
91177308-0d34-0410-b5e6-
96231b3b80d8
Anton Korobeynikov [Fri, 19 Feb 2010 00:29:36 +0000 (00:29 +0000)]
Use the same encoding for EH stuff uniformly on all MachO targets.
This hopefulyl should unbreak EH on PPC/Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96637
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 19 Feb 2010 00:27:40 +0000 (00:27 +0000)]
introduce a new ResultVal and start keeping track of temporary values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96636
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Fri, 19 Feb 2010 00:16:24 +0000 (00:16 +0000)]
Radar
7636153. In the presence of large call frames, it's not sufficient
for ARM to just check if a function has a FP to determine if it's safe
to simplify the stack adjustment pseudo ops prior to eliminating frame
indices. Allow targets to override the default behavior and does so for ARM
and Thumb2.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96634
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 19 Feb 2010 00:05:23 +0000 (00:05 +0000)]
When determining the set of interesting reuse factors, consider
strides in foreign loops. This helps locate reuse opportunities
with existing induction variables in foreign loops and reduces
the need for inserting new ones. This fixes rdar://
7657764.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96629
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Thu, 18 Feb 2010 23:26:33 +0000 (23:26 +0000)]
Indvars needs to explicitly notify ScalarEvolution when it is replacing
a loop exit value, so that if a loop gets deleted, ScalarEvolution
isn't stick holding on to dangling SCEVAddRecExprs for that loop. This
fixes PR6339.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96626
91177308-0d34-0410-b5e6-
96231b3b80d8
Mon P Wang [Thu, 18 Feb 2010 22:33:18 +0000 (22:33 +0000)]
getSplatIndex assumes that the first element of the mask contains the splat index
which is not always true if the mask contains undefs. Modified it to return
the first non undef value.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96621
91177308-0d34-0410-b5e6-
96231b3b80d8
Johnny Chen [Thu, 18 Feb 2010 22:31:18 +0000 (22:31 +0000)]
Added LDRD_PRE/POST & STRD_PRE/POST for disassembly only.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96619
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 18 Feb 2010 22:03:03 +0000 (22:03 +0000)]
add support for referencing registers and immediates,
building the tree to represent them but not emitting
table entries for them yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96617
91177308-0d34-0410-b5e6-
96231b3b80d8
Talin [Thu, 18 Feb 2010 21:43:45 +0000 (21:43 +0000)]
replaceUsesOfWithOnConstant implementation for unions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96616
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Thu, 18 Feb 2010 21:34:02 +0000 (21:34 +0000)]
Hoist this loop-invariant logic out of the loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96614
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 18 Feb 2010 21:33:05 +0000 (21:33 +0000)]
Always normalize spill weights, also for intervals created by spilling.
Moderate the weight given to very small intervals.
The spill weight given to new intervals created when spilling was not
normalized in the same way as the original spill weights calculated by
CalcSpillWeights. That meant that restored registers would tend to hang around
because they had a much higher spill weight that unspilled registers.
This improves the runtime of a few tests by up to 10%, and there are no
significant regressions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96613
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Thu, 18 Feb 2010 21:25:53 +0000 (21:25 +0000)]
Make CodePlacementOpt detect special EH control flow by
checking whether AnalyzeBranch disagrees with the CFG
directly, rather than looking for EH_LABEL instructions.
EH_LABEL instructions aren't always at the end of the
block, due to FP_REG_KILL and other things. This fixes
an infinite loop compiling MultiSource/Benchmarks/Bullet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96611
91177308-0d34-0410-b5e6-
96231b3b80d8