Commit Graph

43292 Commits

Author SHA1 Message Date
Evan Cheng 994faaf218 Fix comment.
llvm-svn: 60592
2008-12-05 17:00:16 +00:00
Cedric Venet af333378c9 The use of the construct:
for(Type1 B = ...;;) { Type2 B ; ... }
is bad: code is hard to read and VS VS don't like it (it ignore the second declaration of B).
This patch fix the problem in tablegen. Please don't write code like this. 

llvm-svn: 60590
2008-12-05 13:37:30 +00:00
Chris Lattner d2a653af0c Make IsValueFullyAvailableInBlock safe.
llvm-svn: 60588
2008-12-05 07:49:08 +00:00
Chris Lattner 6c425556f2 add a new pop_back_val method which returns the value popped. This is
heretical from a STL standpoint, but is oh-so-useful for things that 
can't throw exceptions when copied, like, well, everything in LLVM.

llvm-svn: 60587
2008-12-05 07:11:05 +00:00
Dan Gohman d24be45d99 Drop the reg argument to isRegReDefinedByTwoAddr, which was redundant.
llvm-svn: 60586
2008-12-05 05:45:42 +00:00
Dan Gohman 67840c9803 Update comments. There is no getArgumentAccesses.
llvm-svn: 60585
2008-12-05 05:35:21 +00:00
Dan Gohman 9bb56443a1 Teach StackSlotColoring to update MachineMemOperands when
changing the stack slots on an instruction, to keep them
consistent with the actual memory addresses.

llvm-svn: 60584
2008-12-05 05:31:14 +00:00
Dan Gohman c1dee225d0 Ignore IMPLICIT_DEF instructions when computing physreg liveness.
While they appear to provide a normal clobbering def, they don't
in the case of the awkward IMPLICIT_DEF+INSERT_SUBREG idiom. It
would be good to change INSERT_SUBREG; until then, this change
allows post-regalloc scheduling to cope in a mildly conservative
way.

llvm-svn: 60583
2008-12-05 05:30:02 +00:00
Evan Cheng 2a03c7e977 Re-did 60519. It turns out Darwin's handling of hidden visibility symbols are a bit more complicate than I expected. Both declarations and weak definitions still need a stub indirection. However, the stubs are in data section and they contain the addresses of the actual symbols.
llvm-svn: 60571
2008-12-05 01:06:39 +00:00
Scott Michel 6ce01ab378 CellSPU: Add new directory under tests/CodeGen/CellSPU to retain tests that
aren't part of the test suite but are generally useful nonetheless, and can
be expanded later to test the backend against the actual Cell SPU system.

There's basically no other good place to put this code, so put it here for
the time being.

- vecoperations.c: Vector shuffles for all supported vector types, tests
  for v16i8 add and multiply.

llvm-svn: 60566
2008-12-05 00:01:00 +00:00
Ted Kremenek 123a35a81c Have raw_fd_ostream keep track of the position in the file to make tell() go faster by not requiring a flush().
llvm-svn: 60560
2008-12-04 22:51:11 +00:00
Devang Patel 8c84d28250 Enable LoopIndexSplit pass.
llvm-svn: 60555
2008-12-04 21:40:31 +00:00
Devang Patel c56423b500 Rewrite code that 1) filters loops and 2) calculates new loop bounds.
This fixes many bugs. I will add more test cases in a separate check-in.

Some day, the code that manipulates CFG and updates dom. info could use refactoring help.

llvm-svn: 60554
2008-12-04 21:38:42 +00:00
Owen Anderson 0bcbe8f6a8 Factor out some common code.
llvm-svn: 60553
2008-12-04 21:20:30 +00:00
Scott Michel ea3c49d43d CellSPU: Fix bug 3055
- Add v4f32, v2f64 to LowerVECTOR_SHUFFLE
- Look for vector rotate in shuffle elements, generate a vector rotate
  instead of a full-blown shuffle when opportunity presents itself.
- Generate larger test harness and fix a few interesting but obscure bugs.

llvm-svn: 60552
2008-12-04 21:01:44 +00:00
Duncan Sands 471a654711 When allocating a stack temporary, use the correct
number of bytes for types such as i1 which are not
a multiple of 8 bits in length.

llvm-svn: 60543
2008-12-04 18:08:40 +00:00
Scott Michel 187250bd94 Missing closing brace and reverse conditional condition on NDEBUG
llvm-svn: 60541
2008-12-04 17:16:59 +00:00
Chris Lattner 8f723670ce Start simplifying a switch that has a successor that is a switch.
llvm-svn: 60534
2008-12-04 06:31:07 +00:00
Chris Lattner 5cee4626b8 This code is apparently quite confused. In the meantime,
get it building when NDEBUG is set.

llvm-svn: 60532
2008-12-04 06:14:27 +00:00
Bill Wendling 6949f6135b Temporarily revert r60519. It was causing a bootstrap failure:
/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.obj/./gcc/xgcc -B/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.obj/./gcc/ -B/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.install/i386-apple-darwin9.5.0/bin/ -B/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.install/i386-apple-darwin9.5.0/lib/ -isystem /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.install/i386-apple-darwin9.5.0/include -isystem /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.install/i386-apple-darwin9.5.0/sys-include -DHAVE_CONFIG_H -I. -I../../../llvm-gcc.src/libgomp -I. -I../../../llvm-gcc.src/libgomp/config/posix -I../../../llvm-gcc.src/libgomp -Wall -pthread -Werror -O2 -g -O2 -MT barrier.lo -MD -MP -MF .deps/barrier.Tpo -c ../../../llvm-gcc.src/libgomp/barrier.c  -fno-common -DPIC -o .libs/barrier.o
checking for sys/file.h... /var/folders/zG/zGE-ZJOGFiGjv0B5cs5oYE+++TM/-Tmp-//cc34Jg5P.s:13:non-relocatable subtraction expression, "_gomp_tls_key" minus "L1$pb"
/var/folders/zG/zGE-ZJOGFiGjv0B5cs5oYE+++TM/-Tmp-//cc34Jg5P.s:13:symbol: "_gomp_tls_key" can't be undefined in a subtraction expression
make[4]: *** [barrier.lo] Error 1
make[4]: *** Waiting for unfinished jobs....
/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.obj/./gcc/xgcc -B/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.obj/./gcc/ -B/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.install/i386-apple-darwin9.5.0/bin/ -B/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.install/i386-apple-darwin9.5.0/lib/ -isystem /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.install/i386-apple-darwin9.5.0/include -isystem /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm-gcc.install/i386-apple-darwin9.5.0/sys-include -DHAVE_CONFIG_H -I. -I../../../llvm-gcc.src/libgomp -I. -I../../../llvm-gcc.src/libgomp/config/posix -I../../../llvm-gcc.src/libgomp -Wall -pthread -Werror -O2 -g -O2 -MT alloc.lo -MD -MP -MF .deps/alloc.Tpo -c ../../../llvm-gcc.src/libgomp/alloc.c -o alloc.o >/dev/null 2>&1
yes
checking for sys/param.h... make[3]: *** [all-recursive] Error 1
make[2]: *** [all] Error 2
make[1]: *** [all-target-libgomp] Error 2
make[1]: *** Waiting for unfinished jobs....

llvm-svn: 60527
2008-12-04 04:07:00 +00:00
Scott Michel 40f54d2257 CellSPU:
- First patch from Nehal Desai, a new contributor at Aerospace. Nehal's patch
  fixes sign/zero/any-extending loads for integers and floating point. Example
  code, compiled w/o debugging or optimization where he first noticed the bug:

  int main(void) {
    float a = 99.0;
    printf("%d\n", a);
    return 0;
  }

  Verified that this code actually works on a Cell SPU.

Changes by Scott Michel:
- Fix bug in the value type list constructed by SPUISD::LDRESULT to include
  both the load result's result and chain, not just the chain alone.
- Simplify LowerLOAD and remove extraneous and unnecessary chains.
- Remove unused SPUISD pseudo instructions.

llvm-svn: 60526
2008-12-04 03:02:42 +00:00
Dan Gohman 44f57df254 Use register names instead of numbers in debug output.
llvm-svn: 60525
2008-12-04 02:15:26 +00:00
Dan Gohman 30cad9c192 Make debug output more informative.
llvm-svn: 60524
2008-12-04 02:14:57 +00:00
Evan Cheng 011c4fa8a1 Visibility hidden GVs do not require extra load of symbol address from the GOT or non-lazy-ptr.
llvm-svn: 60519
2008-12-04 01:56:50 +00:00
Dan Gohman 3aab10b932 Add minimal support for disambiguating memory references. Currently
the main thing this covers is spills to distinct spill slots.

llvm-svn: 60517
2008-12-04 01:35:46 +00:00
Chris Lattner 75c2661d24 add a debugging option to help track down j-t problems.
llvm-svn: 60514
2008-12-04 00:07:59 +00:00
Dan Gohman 84efaf6a63 Rewrite the liveness bookkeeping code to fix a bunch of
issues with subreg operands and tied operands.

llvm-svn: 60510
2008-12-03 23:07:27 +00:00
Dale Johannesen 941c37c2f6 Make the debugging dump be a full line.
llvm-svn: 60509
2008-12-03 22:45:31 +00:00
Dale Johannesen 4e9e6ea604 Remove an unused field.
llvm-svn: 60508
2008-12-03 22:43:56 +00:00
Dan Gohman ee1cd1781a Have PseudoSourceValue override Value::dump, so that it works
on PseudoSourceValue values. This also fixes a FIXME in
lib/VMCode/AsmWriter.cpp.

llvm-svn: 60507
2008-12-03 21:37:21 +00:00
Dale Johannesen f7a588b909 Fix a misspelled function name.
llvm-svn: 60506
2008-12-03 20:56:12 +00:00
Chris Lattner dc3f6f2c12 Factor some code into a new FoldSingleEntryPHINodes method.
llvm-svn: 60501
2008-12-03 19:44:02 +00:00
Dan Gohman 0c91a5f56c Fix an inconsistency in a comment.
llvm-svn: 60500
2008-12-03 19:38:38 +00:00
Evan Cheng 1339e72d97 Use mmx (punpckldq VR64, (mmx_v_set0)) to clear high 32-bits of a VR64 register.
llvm-svn: 60499
2008-12-03 19:38:05 +00:00
Dan Gohman 434a3ca8e9 Don't charge the full latency for anti and output dependencies. This is
an area where eventually it would be good to use target-dependent
information.

llvm-svn: 60498
2008-12-03 19:37:34 +00:00
Dale Johannesen 5d84f685a8 A step towards geting linux ppc to work (see PR 3099)
llvm-svn: 60497
2008-12-03 19:33:10 +00:00
Dan Gohman 444baea236 When looking for anti-dependences on the critical path, don't bother
examining non-anti-dependence edges.

llvm-svn: 60496
2008-12-03 19:32:26 +00:00
Dan Gohman 1a32dda4aa Add a comment about callee-saved registers.
llvm-svn: 60495
2008-12-03 19:30:13 +00:00
Dale Johannesen d49ceff6ba Fix a really wrong comment.
llvm-svn: 60494
2008-12-03 19:25:46 +00:00
Chris Lattner 8820ca5560 fix a really incorrect comment.
llvm-svn: 60492
2008-12-03 19:18:54 +00:00
Dan Gohman 3f86b51333 Split foldMemoryOperand into public non-virtual and protected virtual
parts, and add target-independent code to add/preserve
MachineMemOperands.

llvm-svn: 60488
2008-12-03 18:43:12 +00:00
Dan Gohman 69cc2cbbff Rename isSimpleLoad to canFoldAsLoad, to better reflect its meaning.
llvm-svn: 60487
2008-12-03 18:15:48 +00:00
Dan Gohman 78407ac8e5 Extend X86's addFrameReference to add a MachineMemOperand for
the frame reference. This will help post-RA scheduling determine
that spills to distinct stack slots are independent.

llvm-svn: 60486
2008-12-03 18:11:40 +00:00
Rafael Espindola 74dc32d422 Fix some tests. The grep for "il" was matching "file".
llvm-svn: 60485
2008-12-03 17:14:56 +00:00
Dan Gohman 810daf7e93 Update a comment.
llvm-svn: 60484
2008-12-03 17:10:41 +00:00
Duncan Sands f52e518d05 Only check that the result of the mapping was not
a new node if the node was actually remapped.

llvm-svn: 60482
2008-12-03 12:36:16 +00:00
Rafael Espindola cda011b5ad Fix bug 3140.
Print a single parameter .file directive if we have an ELF target.

llvm-svn: 60480
2008-12-03 11:01:37 +00:00
Richard Osborne feece7edab Add support for ISD::TRAP to the XCore backend
llvm-svn: 60479
2008-12-03 10:59:16 +00:00
Evan Cheng 501089f6f4 Refactor code. No functionality change.
llvm-svn: 60478
2008-12-03 08:38:43 +00:00
Bill Wendling f8d1ef9842 CC should only be a ConstantSDNode at this point. Just use 'cast' instead of 'dyn_cast'.
llvm-svn: 60477
2008-12-03 08:32:02 +00:00
Evan Cheng b5a97ff651 Fix test.
llvm-svn: 60476
2008-12-03 08:20:45 +00:00
Chris Lattner 350fc5721d testcase for br undef folding.
llvm-svn: 60471
2008-12-03 07:48:27 +00:00
Chris Lattner 595c7279bd Teach jump threading some more simple tricks:
1) have it fold "br undef", which does occur with
   surprising frequency as jump threading iterates.
2) teach j-t to delete dead blocks.  This removes the successor
   edges, reducing the in-edges of other blocks, allowing 
   recursive simplification.
3) Fold things like:
     br COND, BBX, BBY
  BBX:
     br COND, BBZ, BBW

   which also happens because jump threading iterates.

llvm-svn: 60470
2008-12-03 07:48:08 +00:00
Chris Lattner 37e0136fef third time is the charm.
llvm-svn: 60469
2008-12-03 07:45:15 +00:00
Chris Lattner c04a1ffa9a fix assertion.
llvm-svn: 60468
2008-12-03 07:43:05 +00:00
Chris Lattner 50532410d1 don't spew tons of stuff to the output. This testcase is *not* for
loop deletion (it is for a ton of passes), which is very bad.

llvm-svn: 60465
2008-12-03 06:41:50 +00:00
Chris Lattner 7eb270ed03 Rename DeleteBlockIfDead to DeleteDeadBlock and make it
unconditionally delete the block.  All likely clients will
do the checking anyway.

llvm-svn: 60464
2008-12-03 06:40:52 +00:00
Chris Lattner bcc904a67c Factor some code out of SimplifyCFG, forming a new
DeleteBlockIfDead method.

llvm-svn: 60463
2008-12-03 06:37:44 +00:00
Dan Gohman cc78cdf275 Mark x86's V_SET0 and V_SETALLONES with isSimpleLoad, and teach X86's
foldMemoryOperand how to "fold" them, by converting them into constant-pool
loads. When they aren't folded, they use xorps/cmpeqd, but for example when
register pressure is high, they may now be folded as memory operands, which
reduces register pressure.

Also, mark V_SET0 isAsCheapAsAMove so that two-address-elimination will
remat it instead of copying zeros around (V_SETALLONES was already marked).

llvm-svn: 60461
2008-12-03 05:21:24 +00:00
Bill Wendling e3402692d8 Change label to 'carry' for unsigned adds.
llvm-svn: 60460
2008-12-03 02:43:12 +00:00
Dan Gohman ae3ba45eb2 Add a sanity-check to tablegen to catch the case where isSimpleLoad
is set but mayLoad is not set. Fix all the problems this turned up.

Change code to not use isSimpleLoad instead of mayLoad unless it
really wants isSimpleLoad.

llvm-svn: 60459
2008-12-03 02:30:17 +00:00
Dan Gohman ac5392c596 Fix a missing #include.
llvm-svn: 60458
2008-12-03 02:10:00 +00:00
Dan Gohman a3d9eb0131 Add an explicit keyword.
llvm-svn: 60457
2008-12-03 01:55:47 +00:00
Dan Gohman 9810e76b30 Replace a #include with a forward-declaration.
llvm-svn: 60456
2008-12-03 01:53:18 +00:00
Dan Gohman 0c8df671ac Fix this comment to reflect that it applies to types other
than just i32.

llvm-svn: 60455
2008-12-03 01:39:44 +00:00
Dan Gohman 5d3d1f69e1 Fix byval arguments in the fastcc calling convention. The fastcc convention
delegates to the regular x86-32 convention which handles byval, but only
after it handles a few cases, and it's necessary to handle byval before
handling those cases. This fixes PR3122 (and rdar://6400815), llvm-gcc
miscompiling LLVM.

llvm-svn: 60453
2008-12-03 01:28:04 +00:00
Dan Gohman 971c88f3b2 Add nounwind attributes to this test.
llvm-svn: 60451
2008-12-03 01:10:18 +00:00
Dale Johannesen b43a689520 testcases for recent dag combiner changes
llvm-svn: 60449
2008-12-03 00:52:41 +00:00
Chris Lattner 0a12e95362 Fix isIntN to work with APInts > 64 bits. This method is only
used by clang apparently.

llvm-svn: 60446
2008-12-02 23:33:29 +00:00
Evan Cheng e62150cae4 Remove a (what appears to be) overly strict assertion. Here is what happened:
1. ppcf128 select is expanded to f64 select's.
2. f64 select operand 0 is an i1 truncate, it's promoted to i32 zero_extend.
3. f64 select is updated. It's changed back to a "NewNode" and being re-analyzed.
4. f64 select operands are being processed. Operand 0 is a "NewNode". It's being expunged out of ReplacedValues map.
5. ExpungeNode tries to remap f64 select and notice it's a "NewNode" and assert.
Duncan, please take a look. Thanks.

llvm-svn: 60443
2008-12-02 21:57:09 +00:00
Dale Johannesen 4d2ecb8f68 Minor rewrite per review feedback.
llvm-svn: 60442
2008-12-02 21:17:11 +00:00
Scott Michel 9b0b28e021 Non-functional change: make custom lowering for truncate stylistically
consistent with the way it's generally done in other places.

llvm-svn: 60439
2008-12-02 19:55:08 +00:00
Scott Michel 7364025ff8 CellSPU:
- Incorporate Tilmann Scheller's ISD::TRUNCATE custom lowering patch
- Update SPU calling convention info, even if it's not used yet (but can be
  at some point or another)
- Ensure that any-extended f32 loads are custom lowered, especially when
  they're promoted for use in printf.

llvm-svn: 60438
2008-12-02 19:53:53 +00:00
Dan Gohman 7e9daef644 Fix a typo in a comment.
llvm-svn: 60434
2008-12-02 19:27:20 +00:00
Owen Anderson ecd5b5f2b2 Add support for folding spills into preceding defs when doing pre-alloc splitting.
llvm-svn: 60433
2008-12-02 18:53:47 +00:00
Dale Johannesen 54bdec238a One more transformation.
llvm-svn: 60432
2008-12-02 18:40:40 +00:00
Dale Johannesen 70060013d2 Make the code do what the comment says it does.
llvm-svn: 60431
2008-12-02 18:40:09 +00:00
Chris Lattner 027d726f10 Comment typeo fix, thanks Duncan!
llvm-svn: 60429
2008-12-02 18:33:11 +00:00
Tilmann Scheller 318ccb0e62 make it possible to custom lower TRUNCATE (needed for the CellSPU target)
llvm-svn: 60409
2008-12-02 12:12:25 +00:00
Chris Lattner 1db9bbe802 Implement PRE of loads in the GVN pass with a pretty cheap and
straight-forward implementation.  This does not require any extra
alias analysis queries beyond what we already do for non-local loads.

Some programs really really like load PRE.  For example, SPASS triggers
this ~1000 times, ~300 times in 255.vortex, and ~1500 times on 403.gcc.

The biggest limitation to the implementation is that it does not split
critical edges.  This is a huge killer on many programs and should be
addressed after the initial patch is enabled by default.

The implementation of this should incidentally speed up rejection of 
non-local loads because it avoids creating the repl densemap in cases 
when it won't be used for fully redundant loads.

This is currently disabled by default.
Before I turn this on, I need to fix a couple of miscompilations in
the testsuite, look at compile time performance numbers, and look at
perf impact.  This is pretty close to ready though.

llvm-svn: 60408
2008-12-02 08:16:11 +00:00
Nick Lewycky 4d9966dd2d Add a new SCEV representing signed division.
llvm-svn: 60407
2008-12-02 08:05:48 +00:00
Mon P Wang 6e1c6ad127 Removed some unnecessary code in widening.
llvm-svn: 60406
2008-12-02 07:35:08 +00:00
Chris Lattner 9c1b5027e7 add a little helper function that does PHI translation.
llvm-svn: 60405
2008-12-02 07:16:45 +00:00
Chris Lattner 0cdc0bbb8a add a note
llvm-svn: 60404
2008-12-02 06:32:34 +00:00
Bill Wendling 87beb9b909 Remove some errors that crept in. No functionality change.
llvm-svn: 60403
2008-12-02 06:24:20 +00:00
Bill Wendling 790b4bf9a9 Merge two if-statements into one.
llvm-svn: 60402
2008-12-02 06:22:04 +00:00
Bill Wendling 5635295266 More styalistic changes. No functionality change.
llvm-svn: 60401
2008-12-02 06:18:11 +00:00
Chris Lattner 734e72fb51 add densemap range insertion method.
llvm-svn: 60400
2008-12-02 06:08:04 +00:00
Bill Wendling 85de4b35ca - Remove the buggy -X/C -> X/-C transform. This isn't valid when X isn't a
constant. If X is a constant, then this is folded elsewhere.

- Added a note to Target/README.txt to indicate that we'd like to implement
  this when we're able.

llvm-svn: 60399
2008-12-02 05:12:47 +00:00
Bill Wendling 5369db5917 Improve comment.
llvm-svn: 60398
2008-12-02 05:09:00 +00:00
Bill Wendling 21716dff5e - Reduce nesting.
- No need to do a swap on a canonicalized pattern.

No functionality change.

llvm-svn: 60397
2008-12-02 05:06:43 +00:00
Chris Lattner ead1a61b47 some random comment improvements.
llvm-svn: 60395
2008-12-02 04:52:26 +00:00
Owen Anderson 35bd70c07a Add a test for my previous PRE fix.
llvm-svn: 60394
2008-12-02 04:25:42 +00:00
Owen Anderson d930420ccf Fix an issue that Chris noticed, where local PRE was not properly instantiating
a new value numbering set after splitting a critical edge.  This increases
the number of instances of PRE on 403.gcc from ~60 to ~570.

llvm-svn: 60393
2008-12-02 04:09:22 +00:00
Evan Cheng 1718fd4375 Fix PR3124: overly strict assert.
llvm-svn: 60392
2008-12-02 02:15:36 +00:00
Dale Johannesen 8c76670b5a Add a few more transformations.
llvm-svn: 60391
2008-12-02 01:30:54 +00:00
Bill Wendling 30e9dc81c8 Second stab at target-dependent lowering of everyone's favorite nodes: [SU]ADDO
- LowerXADDO lowers [SU]ADDO into an ADD with an implicit EFLAGS define. The
  EFLAGS are fed into a SETCC node which has the conditional COND_O or COND_C,
  depending on the type of ADDO requested.

- LowerBRCOND now recognizes if it's coming from a SETCC node with COND_O or
  COND_C set.

llvm-svn: 60388
2008-12-02 01:06:39 +00:00
Bill Wendling 122c515809 Reapply r60382. This time, don't mark "ADC" nodes with "implicit EFLAGS".
llvm-svn: 60385
2008-12-02 00:07:05 +00:00
Bill Wendling 351b6659ad Temporarily revert r60382. It caused CodeGen/X86/i2k.ll and others to fail.
llvm-svn: 60383
2008-12-01 23:44:08 +00:00
Bill Wendling a435b1aebc - Have "ADD" instructions return an implicit EFLAGS.
- Add support for seto, setno, setc, and setnc instructions.

llvm-svn: 60382
2008-12-01 23:30:42 +00:00
Bill Wendling 2d59863d06 Expand getVTList, getNodeValueTypes, and SelectNodeTo to handle more value types.
llvm-svn: 60381
2008-12-01 23:28:22 +00:00
Chris Lattner b2f131a4ab Add rdar reference, make this actually fail when the patch isn't applied.
llvm-svn: 60376
2008-12-01 22:35:31 +00:00
Dale Johannesen 069a4eee55 Consider only references to an IV within the loop when
figuring out the base of the IV.  This produces better
code in the example.  (Addresses use (IV) instead of 
(BASE,IV) - a significant improvement on low-register
machines like x86).

llvm-svn: 60374
2008-12-01 22:00:01 +00:00
Chris Lattner fd2a76170c reenable array_pod_sort, this time hopefully happy on 64-bit
and big endian systems.

llvm-svn: 60371
2008-12-01 21:11:25 +00:00
Bill Wendling 6f71bce4cf Don't rebuild RHSNeg. Just use the one that's already there.
llvm-svn: 60370
2008-12-01 21:06:30 +00:00
Bill Wendling 84f6f2539f Document what this check is doing. Also, no need to cast to ConstantInt.
llvm-svn: 60369
2008-12-01 21:03:43 +00:00
Bill Wendling e6c87a4952 Use a simple comparison. Overflow on integer negation can only occur when the
integer is "minint".

llvm-svn: 60366
2008-12-01 19:46:27 +00:00
Chris Lattner e74e210a3f don't #include <algorithm> into the llvm namespace.
llvm-svn: 60365
2008-12-01 19:45:45 +00:00
Scott Michel 08a4e2045d CellSPU:
- Fix v2[if]64 vector insertion code before IBM files a bug report.
- Ensure that zero (0) offsets relative to $sp don't trip an assert
  (add $sp, 0 gets legalized to $sp alone, tripping an assert)
- Shuffle masks passed to SPUISD::SHUFB are now v16i8 or v4i32

llvm-svn: 60358
2008-12-01 17:56:02 +00:00
Chris Lattner 001181731b switch to std::sort until I have time to sort this out.
llvm-svn: 60354
2008-12-01 17:00:08 +00:00
Chris Lattner 5fb10b961b cleanups suggested by duncan, thanks!
llvm-svn: 60353
2008-12-01 16:55:19 +00:00
Chris Lattner 2bc97759b3 define array_pod_sort in terms of operator< instead of my brain
damaged approximation.  This should fix it on big endian platforms
and on 64-bit.

llvm-svn: 60352
2008-12-01 16:50:01 +00:00
Duncan Sands 3d960941b1 There are no longer any places that require a
MERGE_VALUES node with only one operand, so get
rid of special code that only existed to handle
that possibility.

llvm-svn: 60349
2008-12-01 11:41:29 +00:00
Duncan Sands 6ed40141f7 Change the interface to the type legalization method
ReplaceNodeResults: rather than returning a node which
must have the same number of results as the original
node (which means mucking around with MERGE_VALUES,
and which is also easy to get wrong since SelectionDAG
folding may mean you don't get the node you expect),
return the results in a vector.

llvm-svn: 60348
2008-12-01 11:39:25 +00:00
Bill Wendling 47f733e4ea Generalize the FoldOrWithConstant method to fold for any two constants which
don't have overlapping bits.

llvm-svn: 60344
2008-12-01 08:32:40 +00:00
Bill Wendling 22e761b302 Reduce copy-and-paste code by splitting out the code into its own function.
llvm-svn: 60343
2008-12-01 08:23:25 +00:00
Bill Wendling 582fe6b0ca Use m_Specific() instead of double matching.
llvm-svn: 60341
2008-12-01 08:09:47 +00:00
Bill Wendling 4eecfb655b Move pattern check outside of the if-then statement. This prevents us from fiddling with constants unless we have to.
llvm-svn: 60340
2008-12-01 07:47:02 +00:00
Chris Lattner 6f5bf6a718 Rename some variables, only increment BI once at the start of the loop instead of throughout it.
llvm-svn: 60339
2008-12-01 07:35:54 +00:00
Chris Lattner f00aae4968 pull the predMap densemap out of the inner loop of performPRE, so
that it isn't reallocated all the time.  This is a tiny speedup for
GVN: 3.90->3.88s

llvm-svn: 60338
2008-12-01 07:29:03 +00:00
Chris Lattner 2b07d3ccde switch a couple more calls to use array_pod_sort.
llvm-svn: 60337
2008-12-01 06:52:57 +00:00
Chris Lattner a29f0e19ff don't assume iterators implicitly convert to pointers.
llvm-svn: 60336
2008-12-01 06:50:46 +00:00
Chris Lattner 2c2dd15a85 Introduce a new array_pod_sort function and switch LSR to use it
instead of std::sort.  This shrinks the release-asserts LSR.o file
by 1100 bytes of code on my system.

We should start using array_pod_sort where possible.

llvm-svn: 60335
2008-12-01 06:49:59 +00:00
Chris Lattner 2aebea5735 Eliminate use of setvector for the DeadInsts set, just use a smallvector.
This is a lot cheaper and conceptually simpler.

llvm-svn: 60332
2008-12-01 06:27:41 +00:00
Chris Lattner 4da78e3774 DeleteTriviallyDeadInstructions is always passed the
DeadInsts ivar, just use it directly.

llvm-svn: 60330
2008-12-01 06:14:28 +00:00
Chris Lattner a68a5a4784 simplify DeleteTriviallyDeadInstructions again, unlike my previous
buggy rewrite, this notifies ScalarEvolution of a pending instruction
about to be removed and then erases it, instead of erasing it then 
notifying.

llvm-svn: 60329
2008-12-01 06:11:32 +00:00
Chris Lattner 9e6b243428 simplify these patterns using m_Specific. No need to grep for
xor in testcase (or is a substring).

llvm-svn: 60328
2008-12-01 05:16:26 +00:00
Chris Lattner 88a1f0213d Teach jump threading to clean up after itself, DCE and constfolding the
new instructions it simplifies.  Because we're threading jumps on edges
with constants coming in from PHI's, we inherently are exposing a lot more
constants to the new block.  Folding them and deleting dead conditions
allows the cost model in jump threading to be more accurate as it iterates.

llvm-svn: 60327
2008-12-01 04:48:07 +00:00
Chris Lattner 856684d360 The PreVerifier pass preserves everything. In practice, this
prevents the passmgr from adding yet-another domtree invocation
for Verifier if there is already one live.

llvm-svn: 60326
2008-12-01 03:58:38 +00:00
Chris Lattner 084b3a47d3 Change instcombine to use FoldPHIArgGEPIntoPHI to fold two operand PHIs
instead of using FoldPHIArgBinOpIntoPHI.  In addition to being more
obvious, this also fixes a problem where instcombine wouldn't merge two
phis that had different variable indices.  This prevented instcombine
from factoring big chunks of code in 403.gcc.  For example:

 insn_cuid.exit:                
-       %tmp336 = load i32** @uid_cuid, align 4      
-       %tmp337 = getelementptr %struct.rtx_def* %insn_addr.0.ph.i, i32 0, i32 3    
-       %tmp338 = bitcast [1 x %struct.rtunion]* %tmp337 to i32*               
-       %tmp339 = load i32* %tmp338, align 4           
-       %tmp340 = getelementptr i32* %tmp336, i32 %tmp339     
        br label %bb62
 
 bb61:       
-       %tmp341 = load i32** @uid_cuid, align 4     
-       %tmp342 = getelementptr %struct.rtx_def* %insn, i32 0, i32 3        
-       %tmp343 = bitcast [1 x %struct.rtunion]* %tmp342 to i32*           
-       %tmp344 = load i32* %tmp343, align 4        
-       %tmp345 = getelementptr i32* %tmp341, i32 %tmp344          
        br label %bb62
 
 bb62:      
-       %iftmp.62.0.in = phi i32* [ %tmp345, %bb61 ], [ %tmp340, %insn_cuid.exit ]         
+       %insn.pn2 = phi %struct.rtx_def* [ %insn, %bb61 ], [ %insn_addr.0.ph.i, %insn_cuid.exit ]         
+       %tmp344.pn.in.in = getelementptr %struct.rtx_def* %insn.pn2, i32 0, i32 3     
+       %tmp344.pn.in = bitcast [1 x %struct.rtunion]* %tmp344.pn.in.in to i32*  
+       %tmp341.pn = load i32** @uid_cuid     
+       %tmp344.pn = load i32* %tmp344.pn.in 
+       %iftmp.62.0.in = getelementptr i32* %tmp341.pn, i32 %tmp344.pn   
        %iftmp.62.0 = load i32* %iftmp.62.0.in     

llvm-svn: 60325
2008-12-01 03:42:51 +00:00
Chris Lattner 9d02a70a7d Teach inst combine to merge GEPs through PHIs. This is really
important because it is sinking the loads using the GEPs, but
not the GEPs themselves.  This triggers 647 times on 403.gcc
and makes the .s file much much nicer.  For example before:

        je      LBB1_87 ## bb78
LBB1_62:        ## bb77
        leal    84(%esi), %eax
LBB1_63:        ## bb79
        movl    (%eax), %eax
...
LBB1_87:        ## bb78
        movl    $0, 4(%esp)
        movl    %esi, (%esp)
        call    L_make_decl_rtl$stub
        jmp     LBB1_62 ## bb77


after:

        jne     LBB1_63 ## bb79
LBB1_62:        ## bb78
        movl    $0, 4(%esp)
        movl    %esi, (%esp)
        call    L_make_decl_rtl$stub
LBB1_63:        ## bb79
        movl    84(%esi), %eax

The input code was (and the GEPs are merged and
the PHI is now eliminated by instcombine):

        br i1 %tmp233, label %bb78, label %bb77
bb77:           
        %tmp234 = getelementptr %struct.tree_node* %t_addr.3, i32 0, i32 0, i32 22              
        br label %bb79
bb78:           
        call void @make_decl_rtl(%struct.tree_node* %t_addr.3, i8* null) nounwind
        %tmp235 = getelementptr %struct.tree_node* %t_addr.3, i32 0, i32 0, i32 22              
        br label %bb79
bb79:           
        %iftmp.12.0.in = phi %struct.rtx_def** [ %tmp235, %bb78 ], [ %tmp234, %bb77 ]           
        %iftmp.12.0 = load %struct.rtx_def** %iftmp.12.0.in             

llvm-svn: 60322
2008-12-01 02:34:36 +00:00
Chris Lattner 8facc59e72 testcase for my previous commit.
llvm-svn: 60315
2008-12-01 01:42:03 +00:00
Chris Lattner 9ce8995d24 Make GVN be more intelligent about redundant load
elimination: when finding dependent load/stores, realize that
they are the same if aliasing claims must alias instead of relying
on the pointers to be exactly equal.  This makes load elimination
more aggressive.  For example, on 403.gcc, we had:

<     68 gvn    - Number of instructions PRE'd
< 152718 gvn    - Number of instructions deleted
<  49699 gvn    - Number of loads deleted
<   6153 memdep - Number of dirty cached non-local responses
< 169336 memdep - Number of fully cached non-local responses
< 162428 memdep - Number of uncached non-local responses

now we have:

>     64 gvn    - Number of instructions PRE'd
> 153623 gvn    - Number of instructions deleted
>  49856 gvn    - Number of loads deleted
>   5022 memdep - Number of dirty cached non-local responses
> 159030 memdep - Number of fully cached non-local responses
> 162443 memdep - Number of uncached non-local responses

That's an extra 157 loads deleted and extra 905 other instructions nuked.

This slows down GVN very slightly, from 3.91 to 3.96s.

llvm-svn: 60314
2008-12-01 01:31:36 +00:00
Chris Lattner 7e61dafc95 Reimplement the non-local dependency data structure in terms of a sorted
vector instead of a densemap.  This shrinks the memory usage of this thing
substantially (the high water mark) as well as making operations like
scanning it faster.  This speeds up memdep slightly, gvn goes from
3.9376 to 3.9118s on 403.gcc

This also splits out the statistics for the cached non-local case to
differentiate between the dirty and clean cached case.  Here's the stats
for 403.gcc:

  6153 memdep - Number of dirty cached non-local responses
169336 memdep - Number of fully cached non-local responses
162428 memdep - Number of uncached non-local responses

yay for caching :)

llvm-svn: 60313
2008-12-01 01:15:42 +00:00
Bill Wendling 5b902c5b1e Implement ((A|B)&1)|(B&-2) -> (A&1) | B transformation. This also takes care of
permutations of this pattern.

llvm-svn: 60312
2008-12-01 01:07:11 +00:00
Eli Friedman 6f0730ff11 Fix bogus assertion using getSExtValue for legitimate values, like -1 in
an 128-bit-wide integer.  No testcase; the issue I ran into depends on 
local changes.

llvm-svn: 60311
2008-12-01 00:43:48 +00:00
Chris Lattner 8541edec44 Cache analyses in ivars and add some useful DEBUG output.
This speeds up GVN from 4.0386s to 3.9376s.

llvm-svn: 60310
2008-12-01 00:40:32 +00:00
Chris Lattner 80c7d81e81 improve indentation, do cheap checks before expensive ones,
remove some fixme's.  This speeds up GVN very slightly on 403.gcc 
(4.06->4.03s)

llvm-svn: 60309
2008-11-30 23:39:23 +00:00
Chris Lattner 47e81d0e90 Eliminate the DepResultTy abstraction. It is now completely
redundant with MemDepResult, and MemDepResult has a nicer interface.

llvm-svn: 60308
2008-11-30 23:17:19 +00:00
Eli Friedman 11c15a5de7 Minor cleanup: use getTrue and getFalse where appropriate. No
functional change.

llvm-svn: 60307
2008-11-30 22:48:49 +00:00
Eli Friedman 55e4becba9 Some minor cleanups to instcombine; no functionality change.
Note that the FoldOpIntoPhi call is dead because it's impossible for the 
first operand of a subtraction to be both a ConstantInt and a PHINode.

llvm-svn: 60306
2008-11-30 21:09:11 +00:00
Chris Lattner 13cae612b9 Cache TargetData/AliasAnalysis in the pass instead of calling
getAnalysis<>.  getAnalysis<> is apparently extremely expensive.
Doing this speeds up GVN on 403.gcc by 16%!

llvm-svn: 60304
2008-11-30 19:24:31 +00:00
Chris Lattner 15b598618e add the rest of the comparison routines.
llvm-svn: 60303
2008-11-30 19:10:41 +00:00
Bill Wendling de89bc275c Add instruction combining for ((A&~B)|(~A&B)) -> A^B and all permutations.
llvm-svn: 60291
2008-11-30 13:52:49 +00:00
Bill Wendling 9eef421e12 Implement (A&((~A)|B)) -> A&B transformation in the instruction combiner. This
takes care of all permutations of this pattern.

llvm-svn: 60290
2008-11-30 13:08:13 +00:00
Bill Wendling 2fe3229824 Forgot one remaining call to getSExtValue().
llvm-svn: 60289
2008-11-30 12:41:09 +00:00
Bill Wendling 2d2e7861b5 getSExtValue() doesn't work for ConstantInts with bitwidth > 64 bits. Use all
APInt calls instead.

This fixes PR3144.

llvm-svn: 60288
2008-11-30 12:38:24 +00:00
Eli Friedman 09bc610945 Optimize memmove and memset into the LLVM builtins. Note that these
only show up in code from front-ends besides llvm-gcc, like clang.

llvm-svn: 60287
2008-11-30 08:32:11 +00:00
Eli Friedman e9ef170d4a A couple small cleanups, plus a new potential optimization.
llvm-svn: 60286
2008-11-30 07:52:27 +00:00
Eli Friedman e16c0ff1d3 Moving potential optimizations out of PR2330 into lib/Target/README.txt.
Hopefully this isn't too much stuff to dump into this file.

llvm-svn: 60285
2008-11-30 07:36:04 +00:00
Eli Friedman c8228d263b Followup to r60283: optimize arbitrary width signed divisions as well
as unsigned divisions.  Same caveats as before.

llvm-svn: 60284
2008-11-30 06:35:39 +00:00
Eli Friedman 1b7fc154a5 Fix for PR2164: allow transforming arbitrary-width unsigned divides into
multiplies.

Some more cleverness would be nice, though. It would be nice if we 
could do this transformation on illegal types.  Also, we would 
prefer a narrower constant when possible so that we can use a narrower
multiply, which can be cheaper.

llvm-svn: 60283
2008-11-30 06:02:26 +00:00
Bill Wendling 7abf352f44 Don't make TwoToExp signed by default.
llvm-svn: 60279
2008-11-30 05:29:33 +00:00
Bill Wendling af200e9237 From Hacker's Delight:
"For signed integers, the determination of overflow of x*y is not so simple. If
x and y have the same sign, then overflow occurs iff xy > 2**31 - 1. If they
have opposite signs, then overflow occurs iff xy < -2**31."

In this case, x == -1.

llvm-svn: 60278
2008-11-30 05:01:05 +00:00
Eli Friedman bd0f57821a APIntify a test which is potentially unsafe otherwise, and fix the
nearby FIXME.

I'm not sure what the right way to fix the Cell test was; if the 
approach I used isn't okay, please let me know.

llvm-svn: 60277
2008-11-30 04:59:26 +00:00
Bill Wendling 361c0e5f9c Strengthen check for div inst-combining.
llvm-svn: 60276
2008-11-30 04:33:53 +00:00
Bill Wendling 70635adea3 Instcombine was illegally transforming -X/C into X/-C when either X or C
overflowed on negation. This commit checks to make sure that neithe C nor X
overflows. This requires that the RHS of X (a subtract instruction) be a
constant integer.

llvm-svn: 60275
2008-11-30 03:42:12 +00:00
Chris Lattner 441042796d Two changes: Make getDependency remove QueryInst for a dirty record's
ReverseLocalDeps when we update it.  This fixes a regression test
failure from my last commit.

Second, for each non-local cached information structure, keep a bit that
indicates whether it is dirty or not.  This saves us a scan over the whole
thing in the common case when it isn't dirty.

llvm-svn: 60274
2008-11-30 02:52:26 +00:00
Eli Friedman 9ccf574ea0 Fix a link issue I ran into trying compiling LLVM on MinGW with CMake.
Hopefully this doesn't break anyone else's build... it shouldn't unless 
the MinGW variable means something other than compiling with MinGW.

llvm-svn: 60273
2008-11-30 02:42:05 +00:00
Chris Lattner fc678e2af5 introduce a typedef, no functionality change.
llvm-svn: 60272
2008-11-30 02:30:50 +00:00
Chris Lattner 1b810bd5e6 Change NonLocalDeps to be a densemap of pointers to densemap
instead of containing them by value.  This increases the density
(!) of NonLocalDeps as well as making the reallocation case 
faster.  This speeds up gvn on 403.gcc by 2% and makes room for
future improvements.

I'm not super thrilled with having to explicitly manage the new/delete
of the map, but it is necesary for the next change.

llvm-svn: 60271
2008-11-30 02:28:25 +00:00
Chris Lattner ff862c4e88 calls never depend on allocations.
llvm-svn: 60268
2008-11-30 01:44:00 +00:00
Chris Lattner 3ff6d01586 Fix a fixme by making memdep's handling of allocations more logical.
If we see that a load depends on the allocation of its memory with no
intervening stores, we now return a 'None' depedency instead of "Normal".
This tweaks GVN to do its optimization with the new result.

llvm-svn: 60267
2008-11-30 01:39:32 +00:00
Chris Lattner 60444f8aa5 implement a fixme by introducing a new getDependencyFromInternal
method that returns its result as a DepResultTy instead of as a
MemDepResult.  This reduces conversion back and forth.

llvm-svn: 60266
2008-11-30 01:26:32 +00:00
Chris Lattner 2059753e66 Move the getNonLocalDependency method to a more logical place in
the file, no functionality change.

llvm-svn: 60265
2008-11-30 01:18:27 +00:00
Chris Lattner 3d5d5f2c6d REmove an old fixme, resolve another fixme by adding liberal
comments about what this class does.

llvm-svn: 60264
2008-11-30 01:17:08 +00:00
Chris Lattner ada1f87988 remove a bit of incorrect code that tried to be tricky about speeding up
dependencies.  The basic situation was this: consider if we had:

  store1
  ...
  store2
  ...
  store3

Where memdep thinks that store3 depends on store2 and store2 depends 
on store1.  The problem happens when we delete store2: The code in 
question was updating dep info for store3 to be store1.  This is a
spiffy optimization, but is not safe at all, because aliasing isn't
transitive.  This bug isn't exposed today with DSE because DSE will only
zap store2 if it is identifical to store 3, and in this case, it is 
safe to update it to depend on store1.  However, memcpyopt is not so
fortunate, which is presumably why the "dropInstruction" code used to
exist.

Since this doesn't actually provide a speedup in practice, just rip the
code out.

llvm-svn: 60263
2008-11-30 01:09:30 +00:00
Chris Lattner 424d2d8d86 fix indentation. std::pair is "isPod" if the first/second are both isPod.
llvm-svn: 60262
2008-11-30 00:50:20 +00:00
Nick Lewycky 7450a7cbf5 Remove warning about declaration does not declare anything. This class was
already declared in the other headers.

llvm-svn: 60261
2008-11-30 00:36:34 +00:00
Chris Lattner 63bd586d35 Eliminate the dropInstruction method, which is not needed any more.
Fix a subtle iterator invalidation bug I introduced in the last commit.

llvm-svn: 60258
2008-11-29 23:30:39 +00:00
Nick Lewycky af67df5881 Add protected visibility to libLTO.
llvm-svn: 60257
2008-11-29 22:49:59 +00:00
Chris Lattner e7d7e13bf7 implement some fixme's: when deleting an instruction with
an entry in the nonlocal deps map, don't reset entries
referencing that instruction to [dirty, null], instead, set
them to [dirty,next] where next is the instruction after the
deleted one.  Use this information in the non-local deps
code to avoid rescanning entire blocks.

This speeds up GVN slightly by avoiding pointless work.  On
403.gcc this makes GVN 1.5% faster. 

llvm-svn: 60256
2008-11-29 22:02:15 +00:00
Chris Lattner 1c6b62eb4d Change MemDep::getNonLocalDependency to return its results as
a smallvector instead of a DenseMap.  This speeds up GVN by 5%
on 403.gcc.

llvm-svn: 60255
2008-11-29 21:33:22 +00:00
Chris Lattner b8ec75bc35 move MemoryDependenceAnalysis::verifyRemoved to the end of the file,
no functionality/code change.

llvm-svn: 60254
2008-11-29 21:25:10 +00:00
Chris Lattner f280b0c729 reimplement getNonLocalDependency with a simpler worklist
formulation that is faster and doesn't require nonLazyHelper.
Much less code.

llvm-svn: 60253
2008-11-29 21:22:42 +00:00
Chris Lattner c40039c736 don't require GVN to work on dead values, just make the
test return the loaded value.

llvm-svn: 60252
2008-11-29 21:21:48 +00:00
Chris Lattner 8c5ff516c6 Fix a thinko that manifested as a crash on clamav last night.
llvm-svn: 60251
2008-11-29 20:29:04 +00:00
Nick Lewycky 35847809b7 Fix spelling mistake.
llvm-svn: 60250
2008-11-29 20:13:25 +00:00
Chris Lattner 5661fead0b tidy up some variable names.
llvm-svn: 60243
2008-11-29 09:22:14 +00:00
Chris Lattner 9f1988ab6c rename some maps.
llvm-svn: 60242
2008-11-29 09:20:15 +00:00
Chris Lattner 5cd1cfad11 rename some variables.
llvm-svn: 60241
2008-11-29 09:15:21 +00:00
Chris Lattner 80c081828f eliminate a bunch of code in favor of using AliasAnalysis::getModRefInfo.
Put a some code back to handle buggy behavior that GVN expects: it wants
loads to depend on each other, and accesses to depend on their allocations.

llvm-svn: 60240
2008-11-29 09:09:48 +00:00
Torok Edwin 96ce5a0bdf protect against negative values that would exceed allowed bit width
llvm-svn: 60239
2008-11-29 08:52:45 +00:00
Chris Lattner 81f19e9aa4 simplify some code and rename some variables. Reduce nesting.
Use getTypeStoreSize instead of ABITypeSize for in-memory size
in a couple places.

llvm-svn: 60238
2008-11-29 08:51:16 +00:00
Chris Lattner e1822ef660 apparently GCC doesn't believe that I understand C
precedence rules.  Pacify it.

llvm-svn: 60237
2008-11-29 08:36:39 +00:00
Duncan Sands 4f96a92145 Typo fix.
llvm-svn: 60236
2008-11-29 08:03:35 +00:00
Chris Lattner 51ba8d0630 Split getDependency into getDependency and getDependencyFrom, the
former does caching, the later doesn't.  This dramatically simplifies
the logic in getDependency and getDependencyFrom.

llvm-svn: 60234
2008-11-29 03:47:00 +00:00
Bill Wendling 469e3aa696 Temporarily revert r60195. It's causing an optimized bootstrap of llvm-gcc to fail.
llvm-svn: 60233
2008-11-29 03:43:04 +00:00
Chris Lattner e4d32791ef Now that DepType is private, we can start cleaning up some of its uses:
Document the Dirty value more precisely, use it for the uninitialized
DepResultTy value.  Change reverse mappings to be from an instruction*
instead of DepResultTy, and stop tracking other forms.  This makes it more
clear that we only care about the instruction cases.

Eliminate a DepResultTy,bool pair by using Dirty in the local case as well,
shrinking the map and simplifying the code.

This speeds up GVN by ~3% on 403.gcc.

llvm-svn: 60232
2008-11-29 03:22:12 +00:00
Chris Lattner 7f9c8a0f05 Introduce and use a new MemDepResult class to hold the results of a memdep
query.  This makes it crystal clear what cases can escape from MemDep that
the clients have to handle.  This also gives the clients a nice simplified
interface to it that is easy to poke at.

This patch also makes DepResultTy and MemoryDependenceAnalysis::DepType
private, yay.

llvm-svn: 60231
2008-11-29 02:29:27 +00:00
Chris Lattner de04e1173a Reimplement the internal abstraction used by MemDep in terms
of a pointer/int pair instead of a manually bitmangled pointer.
This forces clients to think a little more about checking the 
appropriate pieces and will be useful for internal 
implementation improvements later.

I'm not particularly happy with this.  After going through this
I don't think that the clients of memdep should be exposed to
the internal type at all.  I'll fix this in a subsequent commit.

This has no functionality change.

llvm-svn: 60230
2008-11-29 01:43:36 +00:00
Chris Lattner 482713e14a Fix sentinels to use correctly 'aligned' pointers.
llvm-svn: 60229
2008-11-29 01:36:16 +00:00
Chris Lattner 2774a41b96 Fix spello, add DenseMapInfo specialization for PointerIntPair.
llvm-svn: 60228
2008-11-29 01:18:05 +00:00
Chris Lattner 06db5255c3 fix comment typo
llvm-svn: 60227
2008-11-28 23:57:26 +00:00
Chris Lattner 845d731670 fix a bug.
llvm-svn: 60225
2008-11-28 23:36:15 +00:00
Chris Lattner 7602e4bdbf add a generic "bitmangled pointer" class, which allows a parameterized
pointer and integer type to be used.

llvm-svn: 60224
2008-11-28 23:31:44 +00:00
Chris Lattner d3d9111ede Fix PR3141 by ensuring that MemoryDependenceAnalysis::removeInstruction
properly updates the reverse dependency map when it installs updated 
dependencies for instructions that depend on the removed instruction.

llvm-svn: 60222
2008-11-28 22:51:08 +00:00
Chris Lattner f3f6a801cc don't revisit instructions off the beginning of the block.
llvm-svn: 60221
2008-11-28 22:50:08 +00:00
Chris Lattner 08f3c00562 comment cleanups.
llvm-svn: 60220
2008-11-28 22:41:36 +00:00
Chris Lattner 73c254593e more cleanups for MemoryDependenceAnalysis::removeInstruction,
no functionality change.

llvm-svn: 60219
2008-11-28 22:28:27 +00:00
Chris Lattner a25d3952c6 random cleanups, no functionality change.
llvm-svn: 60218
2008-11-28 22:04:47 +00:00
Chris Lattner 2916a4b589 forward declare CallSite instead of #includ'ing it.
llvm-svn: 60217
2008-11-28 21:47:19 +00:00
Chris Lattner 554d1221aa Run verifyRemoved from removeInstruction when -debug is specified.
This shows the root problem behind PR3141.

llvm-svn: 60216
2008-11-28 21:45:17 +00:00
Chris Lattner e5fd5c29de rename "ping" to "verifyRemoved". I don't know why 'ping' what chosen,
but it doesn't make any sense at all.

Also make the method const, private, and fit in 80 cols while we're at it.

llvm-svn: 60215
2008-11-28 21:42:09 +00:00
Chris Lattner cfa414fe9e comment and indentation improvements.
llvm-svn: 60214
2008-11-28 21:36:43 +00:00
Chris Lattner f2a8ba4cf0 simplify some code, remove escaped newline.
llvm-svn: 60213
2008-11-28 21:29:52 +00:00
Chris Lattner dca2cd3562 remove mysterious escaped newlines.
llvm-svn: 60211
2008-11-28 21:16:44 +00:00
Chris Lattner 8a172daa55 don't call MergeBasicBlockIntoOnlyPred on a block whose only
predecessor is itself.  This doesn't make sense, and this is
a dead infinite loop anyway.

llvm-svn: 60210
2008-11-28 19:54:49 +00:00
Duncan Sands 71ecd67b5d Add include files needed when building with
gcc 4.4 (due to use of sprintf).

llvm-svn: 60209
2008-11-28 10:20:03 +00:00
Duncan Sands 595a4423dc Fix build with gcc-4.4: it doesn't like PICStyle
being both a namespace and a variable name.

llvm-svn: 60208
2008-11-28 09:29:37 +00:00
Chris Lattner e9f6c355bf rewrite RecursivelyDeleteTriviallyDeadInstructions to use a more efficient
formulation that doesn't require set lookups or scanning a set.

llvm-svn: 60203
2008-11-28 01:20:46 +00:00
Chris Lattner d4b5ba615e remove some weirdness that came from the LSR code that has
nothing to do with dead instruction elimination.  No tests in
dejagnu depend on this, so I don't know what it was needed for.

llvm-svn: 60202
2008-11-28 00:58:15 +00:00
Chris Lattner 1adb6759ef rewrite a big chunk of how DSE does recursive dead operand
elimination to use more modern infrastructure.  Also do a bunch
of small cleanups.

llvm-svn: 60201
2008-11-28 00:27:14 +00:00
Mikhail Glushenkov 7283d44569 Scrap some boilerplate.
llvm-svn: 60200
2008-11-28 00:14:11 +00:00
Mikhail Glushenkov cc2d0b2c4c Support multiple compilation graph definitions. Not terribly useful, but makes the code more generic.
llvm-svn: 60199
2008-11-28 00:13:47 +00:00
Mikhail Glushenkov 3bb3da6f4c Add 'hidden' and 'really_hidden' option properties.
llvm-svn: 60198
2008-11-28 00:13:25 +00:00
Mikhail Glushenkov 4ad34cbdc0 Documentation: clarify what is meant by 'multiple edges'.
llvm-svn: 60197
2008-11-28 00:12:09 +00:00
Chris Lattner 8e84c129ce delete ErasePossiblyDeadInstructionTree, replacing uses of it with
RecursivelyDeleteTriviallyDeadInstructions.

llvm-svn: 60196
2008-11-27 23:25:44 +00:00
Chris Lattner c077a2a535 Simplify LoopStrengthReduce::DeleteTriviallyDeadInstructions by
making it use RecursivelyDeleteTriviallyDeadInstructions to do
the heavy lifting.

llvm-svn: 60195
2008-11-27 23:23:35 +00:00
Chris Lattner a1bbdff933 enhance RecursivelyDeleteTriviallyDeadInstructions to make
PHIs dead if they are single-value.

llvm-svn: 60194
2008-11-27 23:18:11 +00:00
Chris Lattner 1cb4f72706 Enhance RecursivelyDeleteTriviallyDeadInstructions to optionally
return a list of deleted instructions.

llvm-svn: 60193
2008-11-27 23:14:34 +00:00
Chris Lattner 96e2dbe008 use continue to reduce indentation
llvm-svn: 60192
2008-11-27 23:00:20 +00:00
Chris Lattner c6c481cdfc remove doConstantPropagation and dceInstruction, they are just
wrappers around the interesting code and use an obscure iterator
abstraction that dates back many many years.

Move EraseDeadInstructions to Transforms/Utils and name it
RecursivelyDeleteTriviallyDeadInstructions.

llvm-svn: 60191
2008-11-27 22:57:53 +00:00
Chris Lattner 5ef9ebf787 simplify code.
llvm-svn: 60190
2008-11-27 22:56:14 +00:00
Chris Lattner c92fa42ddd simplify this logic.
llvm-svn: 60189
2008-11-27 22:46:09 +00:00
Nick Lewycky edd5d3e4e9 Also update the README.
llvm-svn: 60188
2008-11-27 22:41:45 +00:00
Nick Lewycky 4ab50b93c8 Chris prefers icmp/select over udiv!
llvm-svn: 60187
2008-11-27 22:41:10 +00:00
Nick Lewycky b3dc4ad5b4 Add a synthetic missed optimization.
llvm-svn: 60186
2008-11-27 22:12:22 +00:00
Nick Lewycky 69941fd0a0 Add a couple of missed optimizations on integer vectors. Multiply and divide
by 1, as well as multiply by -1.

llvm-svn: 60182
2008-11-27 20:21:08 +00:00
Chris Lattner 4059f43b74 defensive patch: if CGP is merging a block with the entry block, make sure
it ends up being the entry block.

llvm-svn: 60180
2008-11-27 19:29:14 +00:00
Chris Lattner 5dfbfcd80d Fix PR3138: if we merge the entry block into another block, make sure to
move the other block back up into the entry position!

llvm-svn: 60179
2008-11-27 19:25:19 +00:00
Nick Lewycky 2c96bdd8d6 Silence a warning.
Despite changing the order of evaluation, this doesn't actually change the
meaning of the statement.

llvm-svn: 60177
2008-11-27 17:29:52 +00:00
Nuno Lopes 50343cd2fe fix build on some machines. thanks buildbot
llvm-svn: 60175
2008-11-27 16:42:44 +00:00
Nuno Lopes d5c2a144e1 fix my previous commit r60064: compare strings instead of pointers
llvm-svn: 60174
2008-11-27 16:37:02 +00:00
Chris Lattner e0d019def6 switch InstCombine::visitLoadInst to use
FindAvailableLoadedValue

llvm-svn: 60169
2008-11-27 08:56:30 +00:00
Chris Lattner 378b041f03 improve const correctness.
llvm-svn: 60168
2008-11-27 08:39:18 +00:00
Chris Lattner c6ae56d23f enhance FindAvailableLoadedValue to make use of AliasAnalysis
if it has it.

llvm-svn: 60167
2008-11-27 08:18:12 +00:00
Chris Lattner 72f16e70f0 move FindAvailableLoadedValue from JumpThreading to Transforms/Utils.
llvm-svn: 60166
2008-11-27 08:10:05 +00:00
Bill Wendling c6075401c2 Get rid of bogus "control may reach end of non-void function ‘...’ being
inlined" message.

llvm-svn: 60165
2008-11-27 08:00:12 +00:00
Chris Lattner d6204bed3d simplify this code a bit.
llvm-svn: 60164
2008-11-27 07:54:38 +00:00
Chris Lattner 206250284d Use the new MergeBasicBlockIntoOnlyPred function.
llvm-svn: 60163
2008-11-27 07:54:12 +00:00
Chris Lattner 99d6809ac1 move MergeBasicBlockIntoOnlyPred to Transforms/Utils.
llvm-svn: 60162
2008-11-27 07:43:12 +00:00
Bill Wendling 077eb6fcc2 XFAil test due to reverting of patch.
llvm-svn: 60161
2008-11-27 07:34:10 +00:00
Chris Lattner 240051aace rename ThreadBlock to ProcessBlock, since it does other things than
just simple threading.

llvm-svn: 60157
2008-11-27 07:20:04 +00:00
Bill Wendling 128f032cc8 Comment out code that isn't entirely correct.
llvm-svn: 60156
2008-11-27 07:18:35 +00:00
Misha Brukman c9813bda47 Fixed HTML closing tag, cleaned up some spacing.
llvm-svn: 60153
2008-11-27 06:41:20 +00:00
Sanjiv Gupta 7ae1a84465 Removing redundant semicolons. No functionality change.
llvm-svn: 60149
2008-11-27 05:58:04 +00:00
Chris Lattner 98d89d1b1b Make jump threading substantially more powerful, in the following ways:
1. Make it fold blocks separated by an unconditional branch.  This enables
   jump threading to see a broader scope.
2. Make jump threading able to eliminate locally redundant loads when they
   feed the branch condition of a block.  This frequently occurs due to
   reg2mem running.
3. Make jump threading able to eliminate *partially redundant* loads when
   they feed the branch condition of a block.  This is common in code with
   lots of loads and stores like C++ code and 255.vortex.

This implements thread-loads.ll and rdar://6402033.

Per the fixme's, several pieces of this should be moved into Transforms/Utils.

llvm-svn: 60148
2008-11-27 05:07:53 +00:00
Evan Cheng b133907e61 Eliminate a compile time warning.
llvm-svn: 60145
2008-11-27 02:29:25 +00:00
Evan Cheng 3761143755 Avoid inserting noop's in the middle of a loop.
llvm-svn: 60141
2008-11-27 01:16:00 +00:00
Evan Cheng 83bdb38965 On x86 favors folding short immediate into some arithmetic operations (e.g. add, and, xor, etc.) because materializing an immediate in a register is expensive in turns of code size.
e.g.
movl 4(%esp), %eax
addl $4, %eax

is 2 bytes shorter than

movl $4, %eax
addl 4(%esp), %eax

llvm-svn: 60139
2008-11-27 00:49:46 +00:00
Dale Johannesen 73bc0ba4c9 Add a missing case in visitADD.
llvm-svn: 60137
2008-11-27 00:43:21 +00:00
Evan Cheng d1dda5339d Add -march=x86.
llvm-svn: 60135
2008-11-27 00:37:06 +00:00
Ted Kremenek 143b058a8e Add typedef to StringMapEntry.
llvm-svn: 60134
2008-11-27 00:17:25 +00:00
Mikhail Glushenkov 6f4cb52c71 Disallow multiple edges.
llvm-svn: 60127
2008-11-26 22:59:45 +00:00
Bill Wendling a69ced6b68 Add x86-specific test for add-with-overflow intrinsics.
llvm-svn: 60125
2008-11-26 22:42:19 +00:00
Bill Wendling 751a694ad3 Generate something sensible for an [SU]ADDO op when the overflow/carry flag is
the conditional for the BRCOND statement. For instance, it will generate:

    addl %eax, %ecx
    jo LOF

instead of

    addl %eax, %ecx
    ; About 10 instructions to compare the signs of LHS, RHS, and sum.
    jl LOF

llvm-svn: 60123
2008-11-26 22:37:40 +00:00
Chris Lattner 397a11ccd8 Turn on my codegen prepare heuristic by default. It doesn't affect
performance in most cases on the Grawp tester, but does speed some 
things up (like shootout/hash by 15%).  This also doesn't impact 
compile time in a noticable way on the Grawp tester.

It also, of course, gets the testcase it was designed for right :)

llvm-svn: 60120
2008-11-26 22:16:44 +00:00
Bill Wendling 6e41adddab Small formatting change.
llvm-svn: 60113
2008-11-26 19:19:05 +00:00
Bill Wendling 0f5541e4cf Update to explain how ssp and sspreq attributes override each other.
llvm-svn: 60112
2008-11-26 19:07:40 +00:00
Devang Patel 1e916900ba Fix typo.
llvm-svn: 60111
2008-11-26 18:13:11 +00:00
Evan Cheng fc371c6c1d Cosmetic.
llvm-svn: 60110
2008-11-26 18:00:00 +00:00
Duncan Sands d1ba7908cf Check that running the DAG combiner between type
and operation legalization does something useful.

llvm-svn: 60108
2008-11-26 16:44:30 +00:00
Mikhail Glushenkov b21abb9d48 Describe some more options in the man page.
llvm-svn: 60105
2008-11-26 13:40:08 +00:00
Sanjiv Gupta 80810f8c6b Allow custom lowering of ADDE/ADDC/SUBE/SUBC operations.
llvm-svn: 60102
2008-11-26 11:19:00 +00:00
Mikhail Glushenkov 5b4e3b88eb Fix the -I option (llvmc -I dir1 -I dir2 didn't work).
llvm-svn: 60101
2008-11-26 10:57:31 +00:00
Mikhail Glushenkov f9e0513415 Refactor Tools.td to remove repetition.
llvm-svn: 60100
2008-11-26 10:56:56 +00:00
Mikhail Glushenkov 3ac10202c0 Small fix: the error message was incorrect in some cases.
llvm-svn: 60099
2008-11-26 10:55:45 +00:00
Sanjiv Gupta 83c70fa3dc Emit declaration for globals and externs.
Custom lower AND, OR, XOR bitwise operations.

llvm-svn: 60098
2008-11-26 10:53:50 +00:00
Dan Gohman 002a2cb207 Fish kill flag annotations in PUSH instructions.
llvm-svn: 60095
2008-11-26 06:39:12 +00:00
Dan Gohman 3336b1f06b LiveRanges are represented as half-open ranges. Fix the findLiveInMBBs code
and the LiveInterval.h top-level comment and accordingly. This fixes blocks
having spurious live-in registers in boundary cases.

llvm-svn: 60092
2008-11-26 05:50:31 +00:00
Chris Lattner fef04acc50 teach the new heuristic how to handle inline asm.
llvm-svn: 60088
2008-11-26 04:59:11 +00:00
Devang Patel 3bc9e25df2 Disable -loop-index-split for now.
llvm-svn: 60087
2008-11-26 04:58:14 +00:00
Ted Kremenek e076257b8c Add 'tell' method to raw_fd_ostream that clients can use to query the current location in the file the stream is writing to.
llvm-svn: 60085
2008-11-26 03:33:13 +00:00
Chris Lattner 6d71b7fb95 Improve ValueAlreadyLiveAtInst with a cheap and dirty, but effective
heuristic: the value is already live at the new memory operation if
it is used by some other instruction in the memop's block.  This is
cheap and simple to compute (moreso than full liveness).

This improves the new heuristic even more.  For example, it cuts two
out of three new instructions out of 255.vortex:DbmFileInGrpHdr, 
which is one of the functions that the heuristic regressed.  This
overall eliminates another 40 instructions from 403.gcc and visibly
reduces register pressure in 255.vortex (though this only actually
ends up saving the 2 instructions from the whole program).

llvm-svn: 60084
2008-11-26 03:20:37 +00:00
Nick Lewycky ea0bd51cae __fastcall and __stdcall are mingw extensions to gcc for windows. Use the
__attribute__ notation which is supported on more platforms.

llvm-svn: 60083
2008-11-26 03:17:27 +00:00
Chris Lattner e34fe2c52d Start rewroking a subpiece of the profitability heuristic to be
phrased in terms of liveness instead of as a horrible hack.  :)

In pratice, this doesn't change the generated code for either 
255.vortex or 403.gcc, but it could cause minor code changes in 
theory.  This is framework for coming changes.

llvm-svn: 60082
2008-11-26 03:02:41 +00:00
Zhongxing Xu 50e6f82ce4 Adjust indent.
llvm-svn: 60081
2008-11-26 02:57:24 +00:00
Chris Lattner 8b291e66eb add a long-overdue AllocaInst::isStaticAlloca method.
llvm-svn: 60080
2008-11-26 02:54:17 +00:00
Bill Wendling 3d14916b3e Add test for rdar://6394879.
llvm-svn: 60079
2008-11-26 02:21:12 +00:00
Chris Lattner 383a797f42 add a comment, make save/restore logic more obvious.
llvm-svn: 60076
2008-11-26 02:11:11 +00:00
Chris Lattner eb3e4fb6fb This adds in some code (currently disabled unless you pass
-enable-smarter-addr-folding to llc) that gives CGP a better
cost model for when to sink computations into addressing modes.
The basic observation is that sinking increases register 
pressure when part of the addr computation has to be available
for other reasons, such as having a use that is a non-memory
operation.  In cases where it works, it can substantially reduce
register pressure.

This code is currently an overall win on 403.gcc and 255.vortex
(the two things I've been looking at), but there are several 
things I want to do before enabling it by default:

1. This isn't doing any caching of results, so it is much slower 
   than it could be.  It currently slows down release-asserts llc 
   by 1.7% on 176.gcc: 27.12s -> 27.60s.
2. This doesn't think about inline asm memory operands yet.
3. The cost model botches the case when the needed value is live
   across the computation for other reasons.

I'll continue poking at this, and eventually turn it on as llcbeta.

llvm-svn: 60074
2008-11-26 02:00:14 +00:00
Evan Cheng 496b042e20 Revert r60042. IndVarSimplify should check if APFloat is PPCDoubleDouble first before trying to convert it to an integer.
llvm-svn: 60072
2008-11-26 01:11:57 +00:00
Chris Lattner a9ab165b08 Teach CodeGenPrepare to look through Bitcast instructions when attempting to
optimize addressing modes.  This allows us to optimize things like isel-sink2.ll
into:

	movl	4(%esp), %eax
	cmpb	$0, 4(%eax)
	jne	LBB1_2	## F
LBB1_1:	## TB
	movl	$4, %eax
	ret
LBB1_2:	## F
	movzbl	7(%eax), %eax
	ret

instead of:

_test:
	movl	4(%esp), %eax
	cmpb	$0, 4(%eax)
	leal	4(%eax), %eax
	jne	LBB1_2	## F
LBB1_1:	## TB
	movl	$4, %eax
	ret
LBB1_2:	## F
	movzbl	3(%eax), %eax
	ret

This shrinks (e.g.) 403.gcc from 1133510 to 1128345 lines of .s.

Note that the 2008-10-16-SpillerBug.ll testcase is dubious at best, I doubt
it is really testing what it thinks it is.

llvm-svn: 60068
2008-11-26 00:26:16 +00:00
Chris Lattner f0e01def8c fix an over-reduced test.
llvm-svn: 60067
2008-11-26 00:12:08 +00:00
Chris Lattner 0f98f74c74 this doesn't need EH
llvm-svn: 60066
2008-11-26 00:03:26 +00:00
Nuno Lopes b472c9fac7 change AnnotationManager to use 'const char*' instead of std::string. this fixes the leakage of those strings and avoids the creation of such strings in static cosntructors (should result in a little improvement of startup time)
llvm-svn: 60064
2008-11-26 00:00:44 +00:00
Oscar Fuentes c4430484bc CMake: llvmc2 is now known as llvmc.
llvm-svn: 60052
2008-11-25 22:18:49 +00:00
Mikhail Glushenkov e9eeb0d562 Add a man page for llvmc. Really basic for now, will be updated later.
llvm-svn: 60049
2008-11-25 21:38:38 +00:00
Mikhail Glushenkov 98d5ed5cb7 Since the old llvmc was removed, rename llvmc2 to llvmc.
llvm-svn: 60048
2008-11-25 21:38:12 +00:00
Mikhail Glushenkov 67630080b9 Make -fsyntax-only, -include and -emit-llvm work for C++ and Objective-C/C++.
llvm-svn: 60047
2008-11-25 21:35:20 +00:00
Mikhail Glushenkov 86d5fa8f28 docs: Add author info + fix incorrect code example.
llvm-svn: 60046
2008-11-25 21:34:53 +00:00
Mikhail Glushenkov eafa1dd9d9 Small documentation update.
llvm-svn: 60045
2008-11-25 21:34:29 +00:00
Mikhail Glushenkov cb0ffa0182 Document the plugin priority feature.
llvm-svn: 60044
2008-11-25 21:34:01 +00:00
Bill Wendling b4ff5322c1 A simplification for checking whether the signs of the operands and sum differ. Thanks, Duncan.
llvm-svn: 60043
2008-11-25 19:40:17 +00:00
Evan Cheng 2e5aeff676 convertToSignExtendedInteger should return opInvalidOp instead of asserting if sematics of float does not allow arithmetics.
llvm-svn: 60042
2008-11-25 19:00:29 +00:00
Dan Gohman bb1298e6d4 Suppress warnings.
llvm-svn: 60041
2008-11-25 18:53:54 +00:00
Chris Lattner c09f2c2bb0 This method got renamed, thanks to Mattias Holm for pointing this out.
llvm-svn: 60039
2008-11-25 18:34:50 +00:00
Scott Michel 910046d174 CellSPU:
(a) Remove conditionally removed code in SelectXAddr. Basically, hope for the
    best that the A-form and D-form address predicates catch everything before
    the code decides to emit a X-form address.
(b) Expand vector store test cases to include the usual suspects.

llvm-svn: 60034
2008-11-25 17:29:43 +00:00
Nuno Lopes ab6d607ff7 add info about how to run the tests with valgrind
llvm-svn: 60030
2008-11-25 15:57:52 +00:00