Dan Gohman [Mon, 26 Jan 2009 20:27:11 +0000 (20:27 +0000)]
Fix the name of an argument.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63037
91177308-0d34-0410-b5e6-
96231b3b80d8
Anton Korobeynikov [Mon, 26 Jan 2009 19:26:01 +0000 (19:26 +0000)]
During bittest switch lowering emit shift in the test block, which should (theoretically)
allow us to generate more efficient code. We don't do this now though :)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63027
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Mon, 26 Jan 2009 19:18:06 +0000 (19:18 +0000)]
Get rid of a bunch of dead code now that interval reconstruction is enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63026
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Mon, 26 Jan 2009 19:12:06 +0000 (19:12 +0000)]
Fix an issue where LiveIntervals was trying to be smart about removing kill
markers, and ended up foiling the interval reconstruction.
This allows us to turn on reconstruction in the pre alloc splitter, which
fixes a number of miscompilations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63025
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 26 Jan 2009 18:43:34 +0000 (18:43 +0000)]
Enhance logic in X86DAGToDAGISel::PreprocessForRMW which move load inside callseq_start to allow it to be folded into a call. It was not considering the cases where a token factor is between the load and the callseq_start.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63022
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 26 Jan 2009 18:33:51 +0000 (18:33 +0000)]
Silence a bogus compiler warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63021
91177308-0d34-0410-b5e6-
96231b3b80d8
Steve Naroff [Mon, 26 Jan 2009 18:08:55 +0000 (18:08 +0000)]
Update VS project files.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63017
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Levenstein [Mon, 26 Jan 2009 11:07:20 +0000 (11:07 +0000)]
Fix a bug in BitVector.h. All assignment operations (except the usual
assignment operator) were returning a copy of the bit vector, instead of a
reference! This old semantics probably did not meet the expectations.
With this patch, chained assignments happen to the right object.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63012
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 26 Jan 2009 07:53:42 +0000 (07:53 +0000)]
Actually source file has already been uniquified into an id during isel. Eliminate the StringMap.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63009
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 26 Jan 2009 07:41:49 +0000 (07:41 +0000)]
Add data structure to define and track debug location during codegen.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63008
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 26 Jan 2009 07:40:13 +0000 (07:40 +0000)]
Looks like comments were chopped off.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63007
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 26 Jan 2009 07:31:20 +0000 (07:31 +0000)]
LLVM_SUPPORT_DEBUGINFO_H -> LLVM_ANALYSIS_DEBUGINFO_H since DebugInfo.h is under Analysis.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63005
91177308-0d34-0410-b5e6-
96231b3b80d8
Mon P Wang [Mon, 26 Jan 2009 04:39:00 +0000 (04:39 +0000)]
Fixed optimization of combining two shuffles where the first shuffle inputs
has a different number of elements than the output.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62998
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Mon, 26 Jan 2009 04:35:06 +0000 (04:35 +0000)]
Take the next steps in making SDUse more consistent with LLVM Use, and
tidy up SDUse and related code.
- Replace the operator= member functions with a set method, like
LLVM Use has, and variants setInitial and setNode, which take
care up updating use lists, like LLVM Use's does. This simplifies
code that calls these functions.
- getSDValue() is renamed to get(), as in LLVM Use, though most
places can either use the implicit conversion to SDValue or the
convenience functions instead.
- Fix some more node vs. value terminology issues.
Also, eliminate the one remaining use of SDOperandPtr, and
SDOperandPtr itself.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62995
91177308-0d34-0410-b5e6-
96231b3b80d8
Scott Michel [Mon, 26 Jan 2009 03:37:41 +0000 (03:37 +0000)]
Untabify code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62991
91177308-0d34-0410-b5e6-
96231b3b80d8
Scott Michel [Mon, 26 Jan 2009 03:31:40 +0000 (03:31 +0000)]
CellSPU:
- Rename fcmp.ll test to fcmp32.ll, start adding new double tests to fcmp64.ll
- Fix select_bits.ll test
- Capitulate to the DAGCombiner and move i64 constant loads to instruction
selection (SPUISelDAGtoDAG.cpp).
<rant>DAGCombiner will insert all kinds of 64-bit optimizations after
operation legalization occurs and now we have to do most of the work that
instruction selection should be doing twice (once to determine if v2i64
build_vector can be handled by SelectCode(), which then runs all of the
predicates a second time to select the necessary instructions.) But,
CellSPU is a good citizen.</rant>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62990
91177308-0d34-0410-b5e6-
96231b3b80d8
Nate Begeman [Mon, 26 Jan 2009 03:15:54 +0000 (03:15 +0000)]
Fix a typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62989
91177308-0d34-0410-b5e6-
96231b3b80d8
Nate Begeman [Mon, 26 Jan 2009 03:15:31 +0000 (03:15 +0000)]
De-identifying per sabre review
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62988
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Mon, 26 Jan 2009 03:04:57 +0000 (03:04 +0000)]
Build libLTO on any platform so long as PIC is enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62987
91177308-0d34-0410-b5e6-
96231b3b80d8
Nate Begeman [Mon, 26 Jan 2009 02:54:45 +0000 (02:54 +0000)]
Add some documentation for address-space-based access to the segment registers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62986
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 26 Jan 2009 02:18:20 +0000 (02:18 +0000)]
Handle single-entry phi nodes gracefully in condprop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62985
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 26 Jan 2009 02:11:30 +0000 (02:11 +0000)]
Fix PR3408 by making a non-obvious assumption very obvious, and
handling the flaw inherent in that assumption. :)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62984
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 26 Jan 2009 01:57:01 +0000 (01:57 +0000)]
More cleanups and simplifications, no functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62983
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 26 Jan 2009 01:38:24 +0000 (01:38 +0000)]
tidy asserts
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62982
91177308-0d34-0410-b5e6-
96231b3b80d8
Nate Begeman [Mon, 26 Jan 2009 01:24:32 +0000 (01:24 +0000)]
Map address space 256 to gs; similar mappings could be supported for the
other x86 segments. address space 0 is stack/default, 1-255 are reserved for
client use.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62980
91177308-0d34-0410-b5e6-
96231b3b80d8
Nate Begeman [Mon, 26 Jan 2009 00:52:55 +0000 (00:52 +0000)]
Support pattern matching various x86 sse shifts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62979
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 25 Jan 2009 23:08:00 +0000 (23:08 +0000)]
silence a warning when assertions are disabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62976
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sun, 25 Jan 2009 20:29:34 +0000 (20:29 +0000)]
should have removed the + when manually applying a patch!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62973
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sun, 25 Jan 2009 20:21:24 +0000 (20:21 +0000)]
revert this patch for now, because Codegen does still want to generate SSE code,
for example in the case of va-args. XFAIL associated tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62972
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sun, 25 Jan 2009 18:00:06 +0000 (18:00 +0000)]
testcase for llvm-gcc part of PR3402.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62969
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sun, 25 Jan 2009 17:58:56 +0000 (17:58 +0000)]
If user explicitly asks not to use SSE, don't force it. This fixes LLVM part of PR3402.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62967
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sun, 25 Jan 2009 16:29:12 +0000 (16:29 +0000)]
Eliminate the loop that searches through each of the operands
of each use in the SelectionDAG ReplaceAllUses* functions. Thanks
to Chris for spotting this opportunity.
Also, factor out code from all 5 of the ReplaceAllUses* functions
into AddNonLeafNodeToCSEMaps, which is now renamed
AddModifiedNodeToCSEMaps to more accurately reflect its purpose.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62964
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sun, 25 Jan 2009 16:21:38 +0000 (16:21 +0000)]
Whitespace tidiments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62963
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sun, 25 Jan 2009 16:20:37 +0000 (16:20 +0000)]
Move the N->use_empty() assert from DeleteNode to
DeleteNodeNotInCSEMaps, since DeleteNode just calls
DeleteNodeNotInCSEMaps.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62962
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sun, 25 Jan 2009 16:04:50 +0000 (16:04 +0000)]
Reality-check the FAQ entry for "Can I use LLVM to convert C++ to C?"
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62961
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 25 Jan 2009 08:16:27 +0000 (08:16 +0000)]
Start generating arbitrary precision integer SCEVs. This removes the temporary
code that rounded up and capped the size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62958
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 25 Jan 2009 08:05:07 +0000 (08:05 +0000)]
Actually run the test in this directory.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62957
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 25 Jan 2009 07:59:57 +0000 (07:59 +0000)]
The function that does nothing but call malloc is noalias return.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62956
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sun, 25 Jan 2009 06:32:01 +0000 (06:32 +0000)]
Private linkage support for PPC / Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62955
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Sun, 25 Jan 2009 03:53:59 +0000 (03:53 +0000)]
Teach 2addr pass to be do more commuting. If both uses of a two-address instruction are killed, but the first operand has a use before and after the def, commute if the second operand does not suffer from the same issue.
%reg1028<def> = EXTRACT_SUBREG %reg1027<kill>, 1
%reg1029<def> = MOV8rr %reg1028
%reg1029<def> = SHR8ri %reg1029, 7, %EFLAGS<imp-def,dead>
insert => %reg1030<def> = MOV8rr %reg1028
%reg1030<def> = ADD8rr %reg1028<kill>, %reg1029<kill>, %EFLAGS<imp-def,dead>
In this case, it might not be possible to coalesce the second MOV8rr
instruction if the first one is coalesced. So it would be profitable to
commute it:
%reg1028<def> = EXTRACT_SUBREG %reg1027<kill>, 1
%reg1029<def> = MOV8rr %reg1028
%reg1029<def> = SHR8ri %reg1029, 7, %EFLAGS<imp-def,dead>
insert => %reg1030<def> = MOV8rr %reg1029
%reg1030<def> = ADD8rr %reg1029<kill>, %reg1028<kill>, %EFLAGS<imp-def,dead>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62954
91177308-0d34-0410-b5e6-
96231b3b80d8
Nate Begeman [Sat, 24 Jan 2009 22:12:48 +0000 (22:12 +0000)]
Fix an indent and a typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62940
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Sat, 24 Jan 2009 21:49:34 +0000 (21:49 +0000)]
Revert previous change; even this mild and clearly
more accurate change loses more than it gains on
benchmarks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62938
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sat, 24 Jan 2009 19:30:25 +0000 (19:30 +0000)]
add note about possible GEP improvement with fields of size 0.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62925
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sat, 24 Jan 2009 17:32:54 +0000 (17:32 +0000)]
Add a PR comment to this test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62921
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sat, 24 Jan 2009 17:16:04 +0000 (17:16 +0000)]
testcase for PR3381.
Also it was an empty struct, not a void after all.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62920
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sat, 24 Jan 2009 15:58:40 +0000 (15:58 +0000)]
Add more documentation mentioning the limitations due to PR2660.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62919
91177308-0d34-0410-b5e6-
96231b3b80d8
Torok Edwin [Sat, 24 Jan 2009 11:30:49 +0000 (11:30 +0000)]
void* is represented as pointer to empty struct {}.
Thus we need to check whether the struct is empty before trying to index into
it. This fixes PR3381.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62918
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Sat, 24 Jan 2009 10:07:43 +0000 (10:07 +0000)]
Some cleanups. No functional changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62917
91177308-0d34-0410-b5e6-
96231b3b80d8
Dale Johannesen [Sat, 24 Jan 2009 01:27:33 +0000 (01:27 +0000)]
Improve the inlining cost function a bit.
Little practical effect.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62908
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 24 Jan 2009 01:00:13 +0000 (01:00 +0000)]
Make InstCombineStoreToCast handle aggregates more aggressively,
handling the case in Transforms/InstCombine/cast-store-gep.ll, which
is a heavily reduced testcase from Clang on x86-64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62904
91177308-0d34-0410-b5e6-
96231b3b80d8
Mike Stump [Sat, 24 Jan 2009 00:00:41 +0000 (00:00 +0000)]
Perform optional clang building.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62895
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 23 Jan 2009 23:27:33 +0000 (23:27 +0000)]
Refactor code. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62893
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Fri, 23 Jan 2009 22:33:47 +0000 (22:33 +0000)]
Introduce two DWARF attribute extentions DW_AT_APPLE_optimized, DW_AT_APPLE_flags.
DW_AT_APPLE_optimized flag is set when a compile_unit is optimized. The debugger takes advantage of this information some way.
DW_AT_APPLE_flags encodes command line options when certain env. variable is set. This is used by build engineers to track various gcc command lines used by by a project, irrespective of whether the project used makefile, Xcode or something else.
llvm-gcc patch is next.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62888
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 23 Jan 2009 22:33:40 +0000 (22:33 +0000)]
hopefully address PR3379 by making the P modifier work in x86 inline asm.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62887
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Fri, 23 Jan 2009 21:54:18 +0000 (21:54 +0000)]
Fix test case. Use valid file name and directory in global variable's debug info entry.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62883
91177308-0d34-0410-b5e6-
96231b3b80d8
Gabor Greif [Fri, 23 Jan 2009 21:17:04 +0000 (21:17 +0000)]
use CallSite::isCalle instead of slow getOperandNo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62877
91177308-0d34-0410-b5e6-
96231b3b80d8
Gabor Greif [Fri, 23 Jan 2009 19:40:15 +0000 (19:40 +0000)]
Simplify the logic of getting hold of a PHI predecessor block.
There is now a direct way from value-use-iterator to incoming block in PHINode's API.
This way we avoid the iterator->index->iterator trip, and especially the costly
getOperandNo() invocation. Additionally there is now an assertion that the iterator
really refers to one of the PHI's Uses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62869
91177308-0d34-0410-b5e6-
96231b3b80d8
Nate Begeman [Fri, 23 Jan 2009 19:27:28 +0000 (19:27 +0000)]
Add support for deleting a module provider from a JIT in such a way that it does not cause the owned module to be fully materialized.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62864
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Fri, 23 Jan 2009 19:13:31 +0000 (19:13 +0000)]
Empty DIType represents void. In this case no need to construct any type DIE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62861
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 23 Jan 2009 19:10:37 +0000 (19:10 +0000)]
Fold x-0 to x in unsafe-fp-math mode. This comes up in the
testcase from PR3376, and in fact is sufficient to completely
avoid the problem in that testcase.
There's an underlying problem though; TLI.isOperationLegal
considers Custom to be Legal, which might be ok in some
cases, but that's what DAGCombiner is using in many places
to test if something is legal when LegalOperations is true.
When DAGCombiner is running after legalize, this isn't
sufficient. I'll address this in a separate commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62860
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Fri, 23 Jan 2009 08:42:38 +0000 (08:42 +0000)]
Patch to build llvm with srcdir != objdir from
llvm-top. By Michael Schuerig, with some small
tweaks to make the "don't configure if already
configured" logic work.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62842
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 23 Jan 2009 07:12:16 +0000 (07:12 +0000)]
fix two more cases where we could let the NLPDI cache get unsorted.
With this, sqlite3 now passes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62839
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 23 Jan 2009 06:48:41 +0000 (06:48 +0000)]
Unconditionally reset 'cache' to zero, even if we don't need to resort it.
This avoids using a dangling pointer.
Reset NumSortedEntries after restoring Cache to avoid extraneous sorts.
This fixes the reduced sqlite3 testcase, but apparently not the whole app.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62838
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 23 Jan 2009 05:48:59 +0000 (05:48 +0000)]
Only check if coalescing is worthwhile when the result is targeting a more restrictive register class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62837
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 23 Jan 2009 05:45:31 +0000 (05:45 +0000)]
Update test to reflect command line option name change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62836
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Fri, 23 Jan 2009 03:28:53 +0000 (03:28 +0000)]
Stage two of fixing pre-alloc-splitting's code size issues: filter out restores that are just
going to be re-spilled again.
This also helps performance. Pre-alloc-splitting now seems to be an overall win on SPEC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62834
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 23 Jan 2009 02:15:19 +0000 (02:15 +0000)]
Cross register class coalescing. Not yet enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62832
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Fri, 23 Jan 2009 01:25:58 +0000 (01:25 +0000)]
Code did not follow associated comment. not a good idea.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62828
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Fri, 23 Jan 2009 01:19:09 +0000 (01:19 +0000)]
Set appropriate tag for the composite type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62827
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 23 Jan 2009 00:27:03 +0000 (00:27 +0000)]
a minor tweak to my previous patch, handle the invalidation case
when there are multiple iterations of the loop. This fixes PR3375.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62822
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Fri, 23 Jan 2009 00:23:32 +0000 (00:23 +0000)]
Add an initial pass at dead spill/restore removal for pre alloc splitting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62821
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 22 Jan 2009 23:38:45 +0000 (23:38 +0000)]
do not sign extend characters input to isprint. This improves
compatibility with VC++. Patch by Max Burke!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62813
91177308-0d34-0410-b5e6-
96231b3b80d8
Steve Naroff [Thu, 22 Jan 2009 23:05:37 +0000 (23:05 +0000)]
Handful of fixes to the VC project files.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62807
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Thu, 22 Jan 2009 22:05:48 +0000 (22:05 +0000)]
Fix a minor bug in DAGCombiner's folding of SELECT. Folding "select C, 0, 1"
to "C ^ 1" is only valid when C is known to be either 0 or 1. Most of the
similar foldings in this function only handle "i1" types, but this one appears
intentionally written to handle larger integer types. If C has an integer
type larger than "i1", this needs to check if the high bits of a boolean
are known to be zero. I also changed the comment to describe this folding as
"C ^ 1" instead of "~C", since that is what the code does and since the latter
would only be valid for "i1" types. The good news is that most LLVM targets
use TargetLowering::ZeroOrOneBooleanContent so this change will not disable
the optimization; the bad news is that I've been unable to come up with a
testcase to demonstrate the problem.
I have also removed a "FIXME" comment for folding "select C, X, 0" to "C & X",
since the code looks correct to me. It could be made more aggressive by not
limiting the type to "i1", but that would then require checking for
TargetLowering::ZeroOrNegativeOneBooleanContent. Similar changes could be
done for the other SELECT foldings, but it was decided to be not worth the
trouble and complexity (see e.g., r44663).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62790
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Thu, 22 Jan 2009 21:58:43 +0000 (21:58 +0000)]
Don't create ISD::FNEG nodes after legalize if they aren't legal.
Simplify x+0 to x in unsafe-fp-math mode. This avoids a bunch of
redundant work in many cases, because in unsafe-fp-math mode,
ISD::FADD with a constant is considered free to negate, so the
DAGCombiner often negates x+0 to -0-x thinking it's free, when
in reality the end result is -x, which is more expensive than x.
Also, combine x*0 to 0.
This fixes PR3374.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62789
91177308-0d34-0410-b5e6-
96231b3b80d8
Gabor Greif [Thu, 22 Jan 2009 21:35:57 +0000 (21:35 +0000)]
introduce a useful abstraction to find out if a Use is in the call position of an instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62788
91177308-0d34-0410-b5e6-
96231b3b80d8
Tanya Lattner [Thu, 22 Jan 2009 20:11:17 +0000 (20:11 +0000)]
Regenerated configure after backing out 62553 and r62616.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62778
91177308-0d34-0410-b5e6-
96231b3b80d8
Tanya Lattner [Thu, 22 Jan 2009 20:09:20 +0000 (20:09 +0000)]
Revert r62553 and r62616 due to issues with portability.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62777
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 22 Jan 2009 19:53:00 +0000 (19:53 +0000)]
rename methods in System/Host to be more consistent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62776
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Thu, 22 Jan 2009 18:28:11 +0000 (18:28 +0000)]
Do not use buggy llvm-gcc to generate testcases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62770
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Thu, 22 Jan 2009 17:39:32 +0000 (17:39 +0000)]
Add SelectionDAG::getNOT method to construct bitwise NOT operations,
corresponding to the "not" and "vnot" PatFrags. Use the new method
in some places where it seems appropriate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62768
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Thu, 22 Jan 2009 15:37:29 +0000 (15:37 +0000)]
Remove no-longer relevant comment. Pointed out
by Gabor.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62765
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Thu, 22 Jan 2009 15:07:15 +0000 (15:07 +0000)]
This passes on linux.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62764
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjiv Gupta [Thu, 22 Jan 2009 10:14:21 +0000 (10:14 +0000)]
Few targets like the tiny little PIC16 have only 16-bit pointers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62763
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 22 Jan 2009 09:10:11 +0000 (09:10 +0000)]
Eliminate a couple of fields from TargetRegisterClass: SubRegClasses and SuperRegClasses. These are not necessary. Also eliminate getSubRegisterRegClass and getSuperRegisterRegClass. These are slow and their results can change if register file names change. Just use TargetLowering::getRegClassFor() to get the right TargetRegisterClass instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62762
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 22 Jan 2009 07:21:55 +0000 (07:21 +0000)]
fix a typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62761
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 22 Jan 2009 07:16:03 +0000 (07:16 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62760
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 22 Jan 2009 07:08:58 +0000 (07:08 +0000)]
fix a testcase.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62758
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 22 Jan 2009 07:04:01 +0000 (07:04 +0000)]
Fix PR3358, a really nasty bug where recursive phi translated
analyses could be run without the caches properly sorted. This
can fix all sorts of weirdness. Many thanks to Bill for coming
up with the 'issorted' verification idea.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62757
91177308-0d34-0410-b5e6-
96231b3b80d8
Tanya Lattner [Thu, 22 Jan 2009 05:17:59 +0000 (05:17 +0000)]
Bump to 2.6svn.
Regenerate configure (last regen was with the wrong version).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62751
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Thu, 22 Jan 2009 05:10:16 +0000 (05:10 +0000)]
Make tblgen more portable, allowing it to build with ICC.
Patch by Robert Zeh!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62750
91177308-0d34-0410-b5e6-
96231b3b80d8
Mike Stump [Thu, 22 Jan 2009 03:24:22 +0000 (03:24 +0000)]
Don't create .dir files in installation directories. Switch to using
order-only dependancies for installation directories.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62746
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Thu, 22 Jan 2009 03:05:10 +0000 (03:05 +0000)]
The operator<() and operator>() were reversing their tests. Have the test the correct way.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62745
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Thu, 22 Jan 2009 02:10:33 +0000 (02:10 +0000)]
Get rid of warning about implicit 64-to-32 bit conversions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62741
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Thu, 22 Jan 2009 01:39:38 +0000 (01:39 +0000)]
Add a few more notes about LLVM IR features that codegen doesn't
yet support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62739
91177308-0d34-0410-b5e6-
96231b3b80d8
Oscar Fuentes [Thu, 22 Jan 2009 01:20:31 +0000 (01:20 +0000)]
CMake: Tests PERL availability with PERL_EXECUTABLE variable.
On MSVC 64bits, does not put underscore before the symbol name on the
/INCLUDE linker parameter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62738
91177308-0d34-0410-b5e6-
96231b3b80d8
Steve Naroff [Wed, 21 Jan 2009 23:49:23 +0000 (23:49 +0000)]
Add explicit this-> (to make the VS compiler happy).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62731
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Wed, 21 Jan 2009 23:40:54 +0000 (23:40 +0000)]
Recognize inline asm for bswap on x86-64 GLIBC. This allows it
to be supported in the JIT.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62730
91177308-0d34-0410-b5e6-
96231b3b80d8
Douglas Gregor [Wed, 21 Jan 2009 22:38:44 +0000 (22:38 +0000)]
Fix ilist two-phase name lookup
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62724
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Wed, 21 Jan 2009 21:54:44 +0000 (21:54 +0000)]
Fix a missing word.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62720
91177308-0d34-0410-b5e6-
96231b3b80d8