Commit Graph

7783 Commits

Author SHA1 Message Date
Dan Gohman
e298d55930 Special-case static allocas in IndVarSimplify's loop invariant
sinking code, since they are special. If the loop preheader happens
to be the entry block of a function, don't sink static allocas
out of it. This fixes PR4775.

llvm-svn: 80010
2009-08-25 17:42:10 +00:00
David Goodwin
047f69da86 Fixup register kills after scheduling.
llvm-svn: 80002
2009-08-25 17:03:05 +00:00
Anton Korobeynikov
b388728ba7 Provide dynamic_stackalloc lowering for MSP430.
This fixes PR4769

llvm-svn: 80001
2009-08-25 17:00:23 +00:00
Torok Edwin
b3aea7de1a Remove target triple from this test, it fails on x86-64 with
--enable-targets=host-only otherwise.

llvm-svn: 80000
2009-08-25 16:53:15 +00:00
Dan Gohman
b75bb8c2fb Fix a few typos from the removal of -f.
llvm-svn: 79994
2009-08-25 15:45:44 +00:00
Dan Gohman
bf08e82d8e Remove obsolete -f flags.
llvm-svn: 79992
2009-08-25 15:38:29 +00:00
Devang Patel
7d42bfab6c Update DebugInfo interface to use metadata, instead of special named llvm.dbg.... global variables, to encode debugging information in llvm IR. This is mostly a mechanical change that tests metadata support very well.
This change speeds up llvm-gcc by more then 6% at "-O0 -g" (measured by compiling InstructionCombining.cpp!)

llvm-svn: 79977
2009-08-25 05:24:07 +00:00
Dale Johannesen
b8888477db Fix PR 4751, another difficulty with %a modifier on x86.
llvm-svn: 79961
2009-08-25 00:16:14 +00:00
Scott Michel
ad0e6ad80f - Remove SelectSEXTi128 from SPUISelDAGToDAG.cpp, evidently, this is redundant
code, according to Anton (I'm not totally convinced, but we can always
  resurrect patches if we need to do so.)
- Start moving CellSPU's tests to prefer FileCheck.

llvm-svn: 79958
2009-08-24 23:57:35 +00:00
Scott Michel
3444be8f21 Prefer 'FileCheck' over 'grep'.
llvm-svn: 79953
2009-08-24 22:49:22 +00:00
Scott Michel
ee51c50e21 128-bit sign extension and vector shift cleanups, contributed by Ken Werner
(IBM).

llvm-svn: 79949
2009-08-24 22:28:53 +00:00
Bob Wilson
af26d40dc0 Fix a typo. Somehow I thought this had passed before, but I guess not.
llvm-svn: 79937
2009-08-24 21:17:17 +00:00
Bob Wilson
fadb065745 Convert slow test to use FileCheck.
llvm-svn: 79935
2009-08-24 20:33:47 +00:00
Daniel Dunbar
60ce8bad43 Convert two gratuitous abuses of poor helpless CPU cycles to FileCheck.
llvm-svn: 79933
2009-08-24 20:08:27 +00:00
Dale Johannesen
add8a314dd Split test into 3.
llvm-svn: 79926
2009-08-24 17:51:19 +00:00
Daniel Dunbar
937b745b68 llvm-mc/Mach-O: Support symbol attributes.
- This is mostly complete, the main thing missing is .indirect_symbol support
   (which would be straight-forward, except that the way it is implemented in
   'as' makes getting an exact .o match interesting).

llvm-svn: 79899
2009-08-24 08:40:12 +00:00
Chris Lattner
c0e1f7be5f Remove this test now that the interpreter doesn't support
interpreting the wrong endianness anymore.  This fixes a 
failure on clang-powerpc-darwin8

llvm-svn: 79880
2009-08-24 02:05:59 +00:00
Dale Johannesen
3a5e5c7edd Make linkerprivate work for ARM and PPC. Testcase covers
all Darwin targets; could be split into separate tests for
the chip subdirectories, but from Chris' last mail on testing
I assume he'd rather have only one test.  Generic seems to be
the best available, maybe there should be a Darwin subdirectory?

llvm-svn: 79877
2009-08-24 01:03:42 +00:00
Chris Lattner
d8999e8673 remove this test: it is testing for a "feature" I just removed.
llvm-svn: 79874
2009-08-23 23:01:02 +00:00
Daniel Dunbar
cb45f3cb4a Rerevert (r75663 and r76805), seems there is more non-determinism.
llvm-svn: 79856
2009-08-23 17:26:24 +00:00
Jakob Stoklund Olesen
291647e0d9 Fix PR4753.
When undoing a reuse in ReuseInfo::GetRegForReload, check if it was only a
sub-register being used. The MachineOperand::getSubReg() method is only valid
for virtual registers, so we have to recover the sub-register index manually.

llvm-svn: 79855
2009-08-23 13:01:45 +00:00
Daniel Dunbar
f2e39b8c6d Speculatively revert r76823 (i.e., reapply r75663 and r76805) to see if the real
problem is fixed by the TableGen determinism fix.

llvm-svn: 79851
2009-08-23 10:44:51 +00:00
Daniel Dunbar
4c3cf93d1d llvm-mc/Mach-O: Improve symbol table support:
- Honor .globl.

 - Set symbol type and section correctly ('nm' now works), and order symbols
   appropriately.

 - Take care to the string table so that the .o matches 'as' exactly (for ease
   of testing).

llvm-svn: 79740
2009-08-22 11:41:10 +00:00
Daniel Dunbar
eef8b8ed05 Fix typo.
llvm-svn: 79738
2009-08-22 10:09:17 +00:00
Daniel Dunbar
260235bc1a Force triple for these tests.
llvm-svn: 79737
2009-08-22 09:45:43 +00:00
Daniel Dunbar
b960969b44 llvm-mc: Hopefully fix Mach-O tests on big-endian platforms, force values to be
converted to Python ints if possible.

llvm-svn: 79736
2009-08-22 09:28:33 +00:00
Daniel Dunbar
2bb8df3b5d llvm-mc: Clean up some handling of symbol/section association to be more correct
(external was really undefined and there wasn't an explicit representation for
absolute symbols).
 - This still needs some cleanup to how the absolute "pseudo" section is dealt
   with, but I haven't figured out the nicest approach yet.

llvm-svn: 79733
2009-08-22 07:22:36 +00:00
Eli Friedman
79615641f1 Make x86 test actually test x86 code generation. Fix the
construct on ARM, which was breaking by coincidence, and add a similar 
testcase for ARM.

llvm-svn: 79719
2009-08-22 03:13:10 +00:00
Chris Lattner
cf54f19140 rename test, make more specific.
llvm-svn: 79712
2009-08-22 00:44:24 +00:00
Anton Korobeynikov
0fe74b95a5 Add missing RUN line
llvm-svn: 79707
2009-08-22 00:28:50 +00:00
Anton Korobeynikov
833c9c6163 Reduce the test
llvm-svn: 79703
2009-08-22 00:18:11 +00:00
Bob Wilson
79c0af15d0 Use CHECK-NEXT to make sure we're only getting one copy of each shuffle
instruction.

llvm-svn: 79702
2009-08-22 00:13:23 +00:00
Daniel Dunbar
15dbbf3644 llvm-mc: Improve handling of implicit alignment for magic section directives
(e.g., .objc_message_refs).
 - Just emit a .align when we see the directive; this isn't exactly what 'as'
   does but in practice it should be ok, at least for now. See FIXME.

llvm-svn: 79697
2009-08-21 23:30:15 +00:00
Daniel Dunbar
aa60fad397 llvm-mc/Mach-O: Support .o emission for .org and .align.
llvm-svn: 79684
2009-08-21 23:07:38 +00:00
Bob Wilson
6d4400e852 Match VTRN, VZIP, and VUZP shuffles. Restore the tests for these operations,
now using shuffles instead of intrinsics.

llvm-svn: 79673
2009-08-21 20:54:19 +00:00
Anton Korobeynikov
9458a8c84d Add fcopysign instructions
llvm-svn: 79664
2009-08-21 20:02:37 +00:00
Daniel Dunbar
ca9ad524dc llvm-mc/Mach-O: Support byte and fill value emission.
llvm-svn: 79652
2009-08-21 18:29:01 +00:00
Anton Korobeynikov
a706ea5720 Handle 'r' inline asm constraint
llvm-svn: 79648
2009-08-21 18:15:41 +00:00
Dan Gohman
d240c19451 Change getelementptr folding to use APInt instead of uint64_t for
offset computations. This fixes a truncation bug on targets that
don't have 64-bit pointers.

llvm-svn: 79639
2009-08-21 16:52:54 +00:00
Bob Wilson
0da4ec0046 Add some tests for vext.16 and vext.32.
llvm-svn: 79638
2009-08-21 16:35:24 +00:00
Daniel Dunbar
fe93385495 llvm-mc: Accept .fill size of 8.
llvm-svn: 79635
2009-08-21 15:43:35 +00:00
Andreas Neustifter
fb55ceecf3 Removed profiling test, lli not available on all platforms.
llvm-svn: 79633
2009-08-21 15:27:35 +00:00
Andreas Neustifter
b11ab6b931 Added tests for Profiling Infrastructure.
llvm-svn: 79615
2009-08-21 09:36:28 +00:00
Bill Wendling
6f187b7ee9 Make for 32-bit only.
llvm-svn: 79614
2009-08-21 09:29:26 +00:00
Daniel Dunbar
7e81ff0f87 llvm-mc: Start MCAssembler and MCMachOStreamer.
- Together these form the (Mach-O) back end of the assembler.

 - MCAssembler is the actual assembler backend, which is designed to have a
   reasonable API. This will eventually grow to support multiple object file
   implementations, but for now its Mach-O/i386 only.

 - MCMachOStreamer adapts the MCStreamer "actions" API to the MCAssembler API,
   e.g. converting the various directives into fragments, managing state like
   the current section, and so on.

 - llvm-mc will use the new backend via '-filetype=obj', which may eventually
   be, but is not yet, since I hear that people like assemblers which actually
   assemble.

 - The only thing that works at the moment is changing sections. For the time
   being I have a Python Mach-O dumping tool in test/scripts so this stuff can
   be easily tested, eventually I expect to replace this with a real LLVM tool.

 - More doxyments to come.

I assume that since this stuff doesn't touch any of the things which are part of
2.6 that it is ok to put this in not so long before the freeze, but if someone
objects let me know, I can pull it.

llvm-svn: 79612
2009-08-21 09:11:24 +00:00
Bill Wendling
1df766abc3 Probably should emit bitcode instead of text.
llvm-svn: 79607
2009-08-21 08:02:19 +00:00
Bill Wendling
a6999a9051 Testcase for r79597.
llvm-svn: 79598
2009-08-21 06:12:30 +00:00
Bob Wilson
c046b62f1a Remove Neon intrinsics for VZIP, VUZP, and VTRN. We will represent these as
vector shuffles.  Temporarily remove the tests for these operations until the
new implementation is working.

llvm-svn: 79579
2009-08-21 00:01:42 +00:00
Dale Johannesen
f6b2759c5e Use FileCheck even though this means testing for something
that has nothing to do with the point of the test, per Chris.

llvm-svn: 79569
2009-08-20 22:12:08 +00:00
Dan Gohman
f2f57ebc84 Fix an x86 code size regression: prefer RIP-relative addressing
over absolute addressing even in non-PIC mode (unless the address
has an index or something else incompatible), because it has a
smaller encoding.

llvm-svn: 79553
2009-08-20 18:23:44 +00:00