Evan Cheng [Fri, 21 Jan 2011 18:55:51 +0000 (18:55 +0000)]
Last round of fixes for movw + movt global address codegen.
1. Fixed ARM pc adjustment.
2. Fixed dynamic-no-pic codegen
3. CSE of pc-relative load of global addresses.
It's now enabled by default for Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123991
91177308-0d34-0410-b5e6-
96231b3b80d8
Renato Golin [Fri, 21 Jan 2011 18:25:47 +0000 (18:25 +0000)]
Clang was not parsing target triples involving EABI and was generating wrong IR (wrong PCS) and passing the wrong information down llc via the target-triple printed in IR. I've fixed this by adding the parsing of EABI into LLVM's Triple class and using it to choose the correct PCS in Clang's Tools. A Clang patch is on its way to use this infrastructure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123990
91177308-0d34-0410-b5e6-
96231b3b80d8
Oscar Fuentes [Fri, 21 Jan 2011 15:42:54 +0000 (15:42 +0000)]
Handles libffi on the CMake build.
Patch by arrowdodger!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123976
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Fri, 21 Jan 2011 14:07:40 +0000 (14:07 +0000)]
Fix the encoding of QADD/SUB, QDADD/SUB. While qadd16, qadd8 use "rd, rn, rm",
qadd and qdadd uses "rd, rm, rn", the same applies to the 'sub' variants. This
is described in ARM manuals and matches the encoding used by the gnu assembler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123975
91177308-0d34-0410-b5e6-
96231b3b80d8
Venkatraman Govindaraju [Fri, 21 Jan 2011 14:00:01 +0000 (14:00 +0000)]
Implement support for byval arguments in Sparc backend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123974
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Fri, 21 Jan 2011 08:38:09 +0000 (08:38 +0000)]
SCCP doesn't actually preserve the CFG. It will delete and insert terminator
instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123973
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Fri, 21 Jan 2011 06:27:04 +0000 (06:27 +0000)]
Revert "Object: Renable the tests now that none of the build bots complain about aliasing."
This reverts commit
281f3901b7b0869929caf8946c1ad1228bc38922.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123972
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 21 Jan 2011 06:19:05 +0000 (06:19 +0000)]
Enable support for precise scheduling of the instruction selection
DAG. Disable using "-disable-sched-cycles".
For ARM, this enables a framework for modeling the cpu pipeline and
counting stalls. It also activates several heuristics to drive
scheduling based on the model. Scheduling is inherently imprecise at
this stage, and until spilling is improved it may defeat attempts to
schedule. However, this framework provides greater control over
tuning codegen.
Although the flag is not target-specific, it should have very little
affect on the default scheduler used by x86. The only two changes that
affect x86 are:
- scheduling a high-latency operation bumps the current cycle so independent
operations can have their latency covered. i.e. two independent 4
cycle operations can produce results in 4 cycles, not 8 cycles.
- Two operations with equal register pressure impact and no
latency-based stalls on their uses will be prioritized by depth before height
(height is irrelevant if no stalls occur in the schedule below this point).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123971
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 21 Jan 2011 05:51:33 +0000 (05:51 +0000)]
Convert -enable-sched-cycles and -enable-sched-hazard to -disable
flags. They are still not enable in this revision.
Added TargetInstrInfo::isZeroCost() to fix a fundamental problem with
the scheduler's model of operand latency in the selection DAG.
Generalized unit tests to work with sched-cycles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123969
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 21 Jan 2011 05:29:50 +0000 (05:29 +0000)]
fix PR9013, an infinite loop in instcombine.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123968
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 21 Jan 2011 05:08:26 +0000 (05:08 +0000)]
update obsolete comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123965
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Fri, 21 Jan 2011 05:07:13 +0000 (05:07 +0000)]
Object: Renable the tests now that none of the build bots complain about aliasing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123964
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Fri, 21 Jan 2011 02:30:43 +0000 (02:30 +0000)]
Don't try to pull vector bitcasts that change the number of elements through
a select. A vector select is pairwise on each element so we'd need a new
condition with the right number of elements to select on. Fixes PR8994.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123963
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Fri, 21 Jan 2011 02:27:02 +0000 (02:27 +0000)]
Object: Fix type punned pointer issues by making DataRefImpl a union and using intptr_t.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123962
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Fri, 21 Jan 2011 02:08:26 +0000 (02:08 +0000)]
tblgen: Add support for non-inheritable attributes
This patch makes the necessary changes to TableGen to support
non-inheritable attributes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123958
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Fri, 21 Jan 2011 01:12:09 +0000 (01:12 +0000)]
Add a constant folding of casts from zero to zero. Fixes PR9011!
While here, I'd like to complain about how vector is not an aggregate type
according to llvm::Type::isAggregateType(), but they're listed under aggregate
types in the LangRef and zero vectors are stored as ConstantAggregateZero.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123956
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 20 Jan 2011 23:55:07 +0000 (23:55 +0000)]
Don't be overly aggressive with CSE of "ldr constantpool". If it's a pc-relative
value, the "add pc" must be CSE'ed at the same time. We could follow the same
approach as T2 by adding pseudo instructions that combine the ldr + "add pc".
But the better approach is to use movw + movt (which I will enable soon), so
I'll leave this as a TODO.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123949
91177308-0d34-0410-b5e6-
96231b3b80d8
Tobias Grosser [Thu, 20 Jan 2011 21:03:22 +0000 (21:03 +0000)]
Implement requiredTransitive
The PassManager did not implement the transitivity of requiredTransitive. This
was unnoticed since 2006.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123942
91177308-0d34-0410-b5e6-
96231b3b80d8
Tobias Grosser [Thu, 20 Jan 2011 21:03:20 +0000 (21:03 +0000)]
RegionPassPrinter should contain the name of the pass printed
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123941
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Thu, 20 Jan 2011 19:27:16 +0000 (19:27 +0000)]
Add testcases for clz encoding
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123937
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Thu, 20 Jan 2011 19:18:32 +0000 (19:18 +0000)]
Fix the encoding and parsing of clrex instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123936
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Thu, 20 Jan 2011 18:38:10 +0000 (18:38 +0000)]
Move InstAlias check of argument types to a separate loop.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123934
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Thu, 20 Jan 2011 18:38:07 +0000 (18:38 +0000)]
Tidy comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123933
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Thu, 20 Jan 2011 18:38:05 +0000 (18:38 +0000)]
Fix broken check for InstAlias argument used with different types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123932
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Thu, 20 Jan 2011 18:38:02 +0000 (18:38 +0000)]
Precompute InstAlias operand mapping to result instruction operand indices.
There should be no functional change from this, but I think it's simpler this
way.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123931
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Thu, 20 Jan 2011 18:36:07 +0000 (18:36 +0000)]
Change instruction names for consistency
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123930
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Thu, 20 Jan 2011 18:32:09 +0000 (18:32 +0000)]
Add cdp/cdp2 instructions for thumb/thumb2
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123929
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Thu, 20 Jan 2011 18:08:44 +0000 (18:08 +0000)]
Disable objdump-trivial-object.test. It is broken on powerpc-darwin9.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123928
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Thu, 20 Jan 2011 18:06:58 +0000 (18:06 +0000)]
- Use a more appropriate name for Owen's ARM Parser isMCR hack since the same operands can be present
in cdp/cdp2 instructions. Also increase the hack with cdp/cdp2 instructions.
- Fix the encoding of cdp/cdp2 instructions for ARM (no thumb and thumb2 yet) and add testcases for t
hem.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123927
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Jan 2011 17:45:23 +0000 (17:45 +0000)]
SplitKit requires that all defs are in place before calling useIntv().
The value mapping gets confused about which original values have multiple new
definitions so they may need phi insertions.
This could probably be simplified by letting enterIntvBefore() take a live range
to be added following the instruction. As long as the range stays inside the
same basic block, value mapping shouldn't be a problem.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123926
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Jan 2011 17:45:20 +0000 (17:45 +0000)]
Add LiveIntervalMap::dumpCache() to print out the cache used by the ssa update algorithm.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123925
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Thu, 20 Jan 2011 16:58:48 +0000 (16:58 +0000)]
Add mcr*2 and mr*c2 support to thumb2 targets
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123919
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Thu, 20 Jan 2011 16:35:57 +0000 (16:35 +0000)]
Add mcr* and mr*c support to thumb targets
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123917
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 16:24:07 +0000 (16:24 +0000)]
Disable this test until I can figure out why it's broken. Not xfailed because it
usese 100% CPU and times out, so it's annoying to run it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123915
91177308-0d34-0410-b5e6-
96231b3b80d8
Kalle Raiskila [Thu, 20 Jan 2011 15:49:06 +0000 (15:49 +0000)]
Allow sign-extending of i8 and i16 to i128 on SPU.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123912
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Thu, 20 Jan 2011 13:21:55 +0000 (13:21 +0000)]
At -O123 the early-cse pass is run before instcombine has run. According to my
auto-simplier the transform most missed by early-cse is (zext X) != 0 -> X != 0.
This patch adds this transform and some related logic to InstructionSimplify
and removes some of the logic from instcombine (unfortunately not all because
there are several situations in which instcombine can improve things by making
new instructions, whereas instsimplify is not allowed to do this). At -O2 this
often results in more than 15% more simplifications by early-cse, and results in
hundreds of lines of bitcode being eliminated from the testsuite. I did see some
small negative effects in the testsuite, for example a few additional instructions
in three programs. One program, 483.xalancbmk, got an additional 35 instructions,
which seems to be due to a function getting an additional instruction and then
being inlined all over the place.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123911
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Thu, 20 Jan 2011 13:17:59 +0000 (13:17 +0000)]
Refactor mcr* and mr*c instructions into classes with the same encoding. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123910
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 20 Jan 2011 08:56:34 +0000 (08:56 +0000)]
My editor's indent went crazy. Fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123909
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 20 Jan 2011 08:54:28 +0000 (08:54 +0000)]
Expand invalid return values for umulo and smulo. Handle these similarly
to add/sub by doing the normal operation and then checking for overflow
afterwards. This generally relies on the DAG handling the later invalid
operations as well.
Fixes the 64-bit part of rdar://
8622122 and rdar://
8774702.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123908
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 20 Jan 2011 08:43:03 +0000 (08:43 +0000)]
Correct itinerary entry for t2MOV_pic_ga_add_pc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123907
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 20 Jan 2011 08:38:21 +0000 (08:38 +0000)]
Add test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123906
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 20 Jan 2011 08:34:58 +0000 (08:34 +0000)]
Sorry, several patches in one.
TargetInstrInfo:
Change produceSameValue() to take MachineRegisterInfo as an optional argument.
When in SSA form, targets can use it to make more aggressive equality analysis.
Machine LICM:
1. Eliminate isLoadFromConstantMemory, use MI.isInvariantLoad instead.
2. Fix a bug which prevent CSE of instructions which are not re-materializable.
3. Use improved form of produceSameValue.
ARM:
1. Teach ARM produceSameValue to look pass some PIC labels.
2. Look for operands from different loads of different constant pool entries
which have same values.
3. Re-implement PIC GA materialization using movw + movt. Combine the pair with
a "add pc" or "ldr [pc]" to form pseudo instructions. This makes it possible
to re-materialize the instruction, allow machine LICM to hoist the set of
instructions out of the loop and make it possible to CSE them. It's a bit
hacky, but it significantly improve code quality.
4. Some minor bug fixes as well.
With the fixes, using movw + movt to materialize GAs significantly outperform the
load from constantpool method. 186.crafty and 255.vortex improved > 20%, 254.gap
and 176.gcc ~10%.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123905
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 07:22:13 +0000 (07:22 +0000)]
llvm-objdump: Remove redundant includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123902
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 07:22:04 +0000 (07:22 +0000)]
llvm-nm: Fix warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123901
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 06:39:15 +0000 (06:39 +0000)]
Object: Add some tests!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123899
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 06:39:06 +0000 (06:39 +0000)]
Add llvm-objdump
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123898
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 06:38:57 +0000 (06:38 +0000)]
llvm-nm: Update to use the new LLVMObject library.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123897
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 06:38:47 +0000 (06:38 +0000)]
Object: Add ELF support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123896
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 06:38:34 +0000 (06:38 +0000)]
Object: Add COFF Support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123895
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 20 Jan 2011 06:21:59 +0000 (06:21 +0000)]
Selection DAG scheduler register pressure heuristic fixes.
Added a check for already live regs before claiming HighRegPressure.
Fixed a few cases of checking the wrong number of successors.
Added some tracing until these heuristics are better understood.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123892
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Jan 2011 06:20:02 +0000 (06:20 +0000)]
Check that a live range exists before shortening it. This fixes PR8989.
The live range may have been deleted earlier because of rematerialization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123891
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Jan 2011 06:20:00 +0000 (06:20 +0000)]
Add hidden -verify-coalescing to run the machine code verifier before and after
register coalescing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123890
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 20 Jan 2011 05:43:16 +0000 (05:43 +0000)]
gold: MinGW fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123886
91177308-0d34-0410-b5e6-
96231b3b80d8
Venkatraman Govindaraju [Thu, 20 Jan 2011 05:08:26 +0000 (05:08 +0000)]
Sparc backend: Implements a delay slot filler that attempt to fill delay slots
with useful instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123884
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Thu, 20 Jan 2011 03:58:43 +0000 (03:58 +0000)]
Update a comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123879
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Thu, 20 Jan 2011 03:56:35 +0000 (03:56 +0000)]
Remove an unnecessary #include.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123877
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Jan 2011 02:43:19 +0000 (02:43 +0000)]
Fix bug found by new clang warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123872
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 20 Jan 2011 01:29:23 +0000 (01:29 +0000)]
Use only one API at a time.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123866
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 20 Jan 2011 00:29:24 +0000 (00:29 +0000)]
If we can, lower the multiply part of a umulo/smulo call to a libcall
with an invalid type then split the result and perform the overflow check
normally.
Fixes the 32-bit parts of rdar://
8622122 and rdar://
8774702.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123864
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Thu, 20 Jan 2011 00:02:16 +0000 (00:02 +0000)]
Fix debug info for merged global.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123862
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 19 Jan 2011 23:14:59 +0000 (23:14 +0000)]
Divert Hopfield network debug output. It is very noisy.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123859
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 19 Jan 2011 23:14:56 +0000 (23:14 +0000)]
Don't accidentally leave small gaps in the live ranges when leaving the active
interval after an instruction. The leaveIntvAfter() method only adds liveness
from the instruction's boundary index to the inserted copy.
Ideally, SplitKit should be smarter about this, perhaps by combining useIntv()
and leaveIntvAfter() into one method that guarantees continuity.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123858
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Wed, 19 Jan 2011 23:06:07 +0000 (23:06 +0000)]
Make sure to propogate the error code when we fail to parse a modifier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123857
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Wed, 19 Jan 2011 23:04:47 +0000 (23:04 +0000)]
Fix register address expression. Patch by Ken Dyck.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123856
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 19 Jan 2011 22:11:48 +0000 (22:11 +0000)]
Implement RAGreedy::splitAroundRegion and remove loop splitting.
Region splitting includes loop splitting as a subset, and it is more generic.
The splitting heuristics for variables that are live in more than one block are
now:
1. Try to create a region that covers multiple basic blocks.
2. Try to create a new live range for each block with multiple uses.
3. Spill.
Steps 2 and 3 are similar to what the standard spiller is doing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123853
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Wed, 19 Jan 2011 18:56:00 +0000 (18:56 +0000)]
Similarly, analyze truncate through multiply.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123842
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Wed, 19 Jan 2011 16:59:46 +0000 (16:59 +0000)]
Add a missed SCEV fold that is required to continue analyzing the IR produced
by indvars through the scev expander.
trunc(add x, y) --> add(trunc x, y). Currently SCEV largely folds the other way
which is probably wrong, but preserved to minimize churn. Instcombine doesn't
do this fold either, demonstrating a missed optz'n opportunity on code doing
add+trunc+add.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123838
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Wed, 19 Jan 2011 16:56:52 +0000 (16:56 +0000)]
Fix the encoding of mrrc and mcrr family of instructions. Also add testcases for mcr and mrc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123837
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 19 Jan 2011 16:32:21 +0000 (16:32 +0000)]
Add unnamed_addr when we can show that address of a global is not used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123834
91177308-0d34-0410-b5e6-
96231b3b80d8
Douglas Gregor [Wed, 19 Jan 2011 15:57:47 +0000 (15:57 +0000)]
Fix comment for gen-clang-decl-nodes tblgen backend, from Michael Han
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123833
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Wed, 19 Jan 2011 15:56:12 +0000 (15:56 +0000)]
Add a missing SCEV simplification sext(zext x) --> zext x.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123832
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Wed, 19 Jan 2011 15:12:16 +0000 (15:12 +0000)]
ARM/ISel: Factor out isScaledConstantInRange() helper.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123823
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 19 Jan 2011 02:35:27 +0000 (02:35 +0000)]
For ARM subtargets with useNEONForSinglePrecisionFP, double count uses
of the floating point types less than 64-bits. It's somewhat of a temporary
hack but forces more accurate modeling of register pressure and results
in fewer spills.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123811
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 19 Jan 2011 02:26:13 +0000 (02:26 +0000)]
whitespace
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123810
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 19 Jan 2011 02:16:49 +0000 (02:16 +0000)]
Don't forget to emit the load from indirect symbol when using movw + movt to materialize GA indirect symbols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123809
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Tue, 18 Jan 2011 23:01:21 +0000 (23:01 +0000)]
When matching asm operands, always try to match the most restricted type first.
Unfortunately, while this is the "right" thing to do, it breaks some ARM
asm parsing tests because MemMode5 and ThumbMemModeReg are ambiguous. This
is tricky to resolve since neither is a subset of the other.
XFAIL the test for now. The old way was broken in other ways, just ways
we didn't happen to be testing, and our ARM asm parsing is going to require
significant revisiting at a later point anyways.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123786
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 21:58:20 +0000 (21:58 +0000)]
Create two new generic classes to represent the following VMRS/VMSR variations:
vmrs reg, fpexc
vmrs reg, fpsid
vmsr fpexc, reg
vmsr fpsid, reg
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123783
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 21:31:35 +0000 (21:31 +0000)]
Fix MRS encoding for arm and thumb.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123778
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 21:17:09 +0000 (21:17 +0000)]
Fix the encoding of t2ISB by using the right class and also parse it correctly
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123776
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Tue, 18 Jan 2011 21:16:06 +0000 (21:16 +0000)]
Teach BasicAA to return PartialAlias in cases where both pointers
are pointing to the same object, one pointer is accessing the entire
object, and the other is access has a non-zero size. This prevents
TBAA from kicking in and saying NoAlias in such cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123775
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 18 Jan 2011 21:13:27 +0000 (21:13 +0000)]
Add RAGreedy methods for splitting live ranges around regions.
Analyze the live range's behavior entering and leaving basic blocks. Compute an
interference pattern for each allocation candidate, and use SpillPlacement to
find an optimal region where that register can be live.
This code is still not enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123774
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 20:55:11 +0000 (20:55 +0000)]
Follow the current hack set and enable the correct parsing of bkpt while in thumb mode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123772
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 18 Jan 2011 20:53:04 +0000 (20:53 +0000)]
fix rdar://
8878965, a regression I introduced with the recent
llvm.objectsize changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123771
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 20:45:56 +0000 (20:45 +0000)]
Add support for parsing and encoding ARM's official syntax for the BFI instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123770
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Tue, 18 Jan 2011 19:59:19 +0000 (19:59 +0000)]
Add a FIXME.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123769
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 19:50:18 +0000 (19:50 +0000)]
Ensure Mips::GP is properly reloaded after a function call. Patch by Sasa Stankovic
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123768
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 19:41:41 +0000 (19:41 +0000)]
Negative zero is not legal on mips. Patch by Sasa Stankovic
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123766
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 19:38:25 +0000 (19:38 +0000)]
Handle (i32,i32) => f64 in a cleaner way. Patch by Sasa Stankovic
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123763
91177308-0d34-0410-b5e6-
96231b3b80d8
Bruno Cardoso Lopes [Tue, 18 Jan 2011 19:29:17 +0000 (19:29 +0000)]
Add support for mips32 madd and msub instructions. Patch by Akira Hatanaka
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123760
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Tue, 18 Jan 2011 11:50:19 +0000 (11:50 +0000)]
For completeness, generalize the (X + Y) - Y -> X transform and add X - (X + 1) -> -1.
These were not recommended by my auto-simplifier since they don't fire often enough.
However they do fire from time to time, for example they remove one subtraction from
the final bitcode for 483.xalancbmk.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123755
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Tue, 18 Jan 2011 09:24:58 +0000 (09:24 +0000)]
Simplify (X<<1)-X into X. According to my auto-simplier this is the most common missed
simplification in fully optimized code. It occurs sporadically in the testsuite, and
many times in 403.gcc: the final bitcode has 131 fewer subtractions after this change.
The reason that the multiplies are not eliminated is the same reason that instcombine
did not catch this: they are used by other instructions (instcombine catches this with
a more general transform which in general is only profitable if the operands have only
one use).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123754
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 18 Jan 2011 07:47:48 +0000 (07:47 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123752
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 18 Jan 2011 06:12:10 +0000 (06:12 +0000)]
finish a sentence.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123750
91177308-0d34-0410-b5e6-
96231b3b80d8
Venkatraman Govindaraju [Tue, 18 Jan 2011 06:09:55 +0000 (06:09 +0000)]
SPARC backend: Modified LowerCall and LowerFormalArguments so that they use CallingConv assignments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123749
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Tue, 18 Jan 2011 06:07:18 +0000 (06:07 +0000)]
Remove an unnecessary #include.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123748
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Tue, 18 Jan 2011 06:06:27 +0000 (06:06 +0000)]
Move DominanceFrontier from VMCore to Analysis.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123747
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 18 Jan 2011 05:55:27 +0000 (05:55 +0000)]
McARM: Use accessors where appropriate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123746
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 18 Jan 2011 05:55:21 +0000 (05:55 +0000)]
McARM: Fill in ASMOperand::dump() for memory operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123745
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Tue, 18 Jan 2011 05:55:15 +0000 (05:55 +0000)]
McARM: Make ARMOperand use a union where appropriate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123744
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Tue, 18 Jan 2011 05:44:04 +0000 (05:44 +0000)]
There is no point in verifying an analysis that is never updated.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123743
91177308-0d34-0410-b5e6-
96231b3b80d8