Michel Danzer [Mon, 11 Feb 2013 15:58:21 +0000 (15:58 +0000)]
R600/SI: Use V_ADD_F32 instead of V_MOV_B32 for clamp/neg/abs modifiers.
The modifiers don't seem to have any effect with V_MOV_B32, supposedly it's
meant to just move bits untouched.
Fixes 46 piglit tests with radeonsi, though unfortunately 11 of those had
just regressed because they started using the clamp modifier.
NOTE: This is a candidate for the Mesa stable branch.
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174890
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Mon, 11 Feb 2013 14:36:01 +0000 (14:36 +0000)]
[asan] added a flag -mllvm asan-short-64bit-mapping-offset=1 (0 by default)
This flag makes asan use a small (<2G) offset for 64-bit asan shadow mapping.
On x86_64 this saves us a register, thus achieving ~2/3 of the
zero-base-offset's benefits in both performance and code size.
Thanks Jakub Jelinek for the idea.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174886
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 11 Feb 2013 14:25:52 +0000 (14:25 +0000)]
AArch64: fix build on some MSVC versions
This does two things:
It removes a call to abs() which may have "long long" parameter on Windows,
which is not necessarily available in C++03.
It also corrects the signedness of Amount, which was relying on
implementation-defined conversions previously.
Code was already tested (albeit in an implemnetation defined way) so no extra
tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174885
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Mon, 11 Feb 2013 14:14:32 +0000 (14:14 +0000)]
Restore the resurrected doc link previously deleted
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174884
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Mon, 11 Feb 2013 13:16:30 +0000 (13:16 +0000)]
Remove a dead PowerPC doc link
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174881
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 11 Feb 2013 12:32:18 +0000 (12:32 +0000)]
AArch64: Simplify logic in deciding whether bfi is valid
Previous code had a confusing comment which was mostly an implementation
detail. This condition corresponds to "lsb up to register width" and "width not
ridiculous".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174877
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 11 Feb 2013 11:16:02 +0000 (11:16 +0000)]
AArch64: Add basic relocation processing for llvm-dwarfdump.
This allows llvm-dwarfdump to handle the relocations needed, at least
for LLVM-produced code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174874
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 11 Feb 2013 10:51:41 +0000 (10:51 +0000)]
AArch64: Undo change to how test was run
This broke on Windows, presumably due to interleaving of output streams.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174873
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 11 Feb 2013 09:29:37 +0000 (09:29 +0000)]
Make use of DiagnosticType to provide better AArch64 diagnostics.
This gives a DiagnosticType to all AsmOperands in sight. This replaces all
"invalid operand" diagnostics with something more specific. The messages given
should still be sufficiently vague that they're not usually actively misleading
when LLVM guesses your instruction incorrectly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174871
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Mon, 11 Feb 2013 09:24:47 +0000 (09:24 +0000)]
Add support for updating LiveIntervals to MachineBasicBlock::SplitCriticalEdge().
This is currently a bit hairier than it needs to be, since depending on where the
split block resides the end ListEntry of the split block may be the end ListEntry
of the original block or a new entry. Some changes to the SlotIndexes updating
should make it possible to eliminate the two cases here.
This also isn't as optimized as it could be. In the future Liveinterval should
probably get a flag that indicates whether the LiveInterval is within a single
basic block. We could ignore all such intervals when splitting an edge.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174870
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Mon, 11 Feb 2013 09:24:45 +0000 (09:24 +0000)]
Update SlotIndexes after updateTerminator() possibly removes instructions. I am
really trying to avoid piping SlotIndexes through to RemoveBranch() and friends.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174869
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Mon, 11 Feb 2013 09:24:42 +0000 (09:24 +0000)]
Fix some problems with the updating of SlotIndexes after adding a new MBB. In
particular, holes were being left between two blocks after splitting an edge.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174868
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 11 Feb 2013 08:43:33 +0000 (08:43 +0000)]
Add support for printing out the attribute groups.
This emits the attribute groups that are used by the functions. (It currently
doesn't print out return type or parameter attributes within attribute groups.)
Note: The functions still retrieve their attributes from the "old" bitcode
format (using the deprecated 'Raw()' method). This means that string attributes
within an attribute group will not show up during a disassembly. This will be
addressed in a future commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174867
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 11 Feb 2013 08:34:57 +0000 (08:34 +0000)]
FileCheck-ize the tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174865
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Mon, 11 Feb 2013 08:13:54 +0000 (08:13 +0000)]
[tsan/msan] adding thread_safety and uninitialized_checks attributes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174864
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Mon, 11 Feb 2013 05:37:07 +0000 (05:37 +0000)]
Revert "Rename LLVMContext diagnostic handler types and functions."
This reverts my commit 171047. Now that I've removed my misguided attempt to
support backend warnings, these diagnostics are only about inline assembly.
It would take quite a bit more work to generalize them properly, so I'm
just reverting this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174860
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Feb 2013 05:29:51 +0000 (05:29 +0000)]
BBVectorize: Remove the linear searches from pair connection searching
This removes the last of the linear searches over ranges of std::multimap
iterators, giving a 7% speedup on the doduc.bc input from PR15222.
No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174859
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Feb 2013 05:29:49 +0000 (05:29 +0000)]
BBVectorize: Avoid linear searches within the load-move set
This is another cleanup aimed at eliminating linear searches
in ranges of std::multimap.
No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174858
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Feb 2013 05:29:48 +0000 (05:29 +0000)]
BBVectorize: isa/cast cleanup in getInstructionTypes
Profiling suggests that getInstructionTypes is performance-sensitive,
this cleans up some double-casting in that function in favor of
using dyn_cast.
No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174857
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Feb 2013 05:29:41 +0000 (05:29 +0000)]
BBVectorize: Make the bookkeeping to support full cycle checking less expensive
By itself, this does not have much of an effect, but only because in the default
configuration the full cycle checks are used only for small problem sizes.
This is part of a general cleanup of uses of iteration over std::multimap
ranges only for the purpose of checking membership.
No functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174856
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 11 Feb 2013 01:27:15 +0000 (01:27 +0000)]
Currently, codegen may spent some time in SDISel passes even if an entire
function is successfully handled by fast-isel. That's because function
arguments are *always* handled by SDISel. Introduce FastLowerArguments to
allow each target to provide hook to handle formal argument lowering.
As a proof-of-concept, add ARMFastIsel::FastLowerArguments to handle
functions with 4 or fewer scalar integer (i8, i16, or i32) arguments. It
completely eliminates the need for SDISel for trivial functions.
rdar://
13163905
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174855
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 11 Feb 2013 01:18:26 +0000 (01:18 +0000)]
Remove unnecessary code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174854
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 11 Feb 2013 01:16:51 +0000 (01:16 +0000)]
Fix unnecessary removal of const through cast machinery
I have some uncommitted changes to the cast code that catch this sort of thing
at compile-time but I still need to do some other cleanup before I can enable
it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174853
91177308-0d34-0410-b5e6-
96231b3b80d8
Joel Jones [Sun, 10 Feb 2013 23:56:30 +0000 (23:56 +0000)]
Spelling correction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174852
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sun, 10 Feb 2013 23:29:54 +0000 (23:29 +0000)]
Fix the unused but nearly correct method SlotIndexes::insertMBBInMaps() and add
support for updating SlotIndexes to MachineBasicBlock::SplitCriticalEdge(). This
calls renumberIndexes() every time; it should be improved to only renumber
locally.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174851
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sun, 10 Feb 2013 23:29:49 +0000 (23:29 +0000)]
Abstract the liveness checking in PHIElimination::SplitPHIEdges() to support
both LiveVariables and LiveIntervals.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174850
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 23:24:25 +0000 (23:24 +0000)]
Add support in the bitcode reader to read the attribute groups.
This reads the attribute groups. It currently doesn't do anything with them.
NOTE: In the commit to the bitcode writer, the format *may* change in the near
future. Which means that this code would also change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174849
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 23:18:05 +0000 (23:18 +0000)]
The 'Raw' method cannot handle 'string' attributes. Don't even try.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174848
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 23:17:10 +0000 (23:17 +0000)]
Update with attribute group IDs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174847
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 23:15:51 +0000 (23:15 +0000)]
Eat the alignment keyword if we're in an attribute group.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174846
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 23:09:32 +0000 (23:09 +0000)]
Add code for emitting the attribute groups.
This is some initial code for emitting the attribute groups into the bitcode.
NOTE: This format *may* change! Do not rely upon the attribute groups' bitcode
not changing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174845
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 23:06:02 +0000 (23:06 +0000)]
Add support for attribute groups in the value enumerator.
Attribute groups are essentially all AttributeSets which are used by the
program. Enumerate them here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174844
91177308-0d34-0410-b5e6-
96231b3b80d8
Vincent Lejeune [Sun, 10 Feb 2013 17:57:33 +0000 (17:57 +0000)]
Test Commit - Remove some trailing whitespace in R600Instructions.td
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174839
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 10:13:23 +0000 (10:13 +0000)]
Handle string attributes in the AttrBuilder.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174834
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 10:12:50 +0000 (10:12 +0000)]
Use a 'continue' here to stop from double lexing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174833
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 10:12:06 +0000 (10:12 +0000)]
Add 'empty' query methods to the builder and use them in the verifier.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174832
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sun, 10 Feb 2013 06:42:36 +0000 (06:42 +0000)]
Add support for updating LiveIntervals to PHIElimination. If LiveIntervals are
present, it currently verifies them with the MachineVerifier, and this passed
all of the test cases in 'make check' (when accounting for existing verifier
errors). There were some assertion failures in the two-address pass, but they
also happened on code without phis and look like they are caused by different
kill flags from LiveIntervals.
The only part that doesn't work is the critical edge splitting heuristic,
because there isn't currently an efficient way to update LiveIntervals after
splitting an edge. I'll probably start by implementing the slow fallback and
test that it works before tackling the fast path for single-block ranges. The
existing code that updates LiveVariables is fairly slow as it is.
There isn't a command-line option for enabling this; instead, just edit
PHIElimination.cpp to require LiveIntervals.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174831
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sun, 10 Feb 2013 06:42:34 +0000 (06:42 +0000)]
Fix a typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174830
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sun, 10 Feb 2013 06:42:32 +0000 (06:42 +0000)]
Remove ancient references to 'atomic' phis in PHIElimination that don't really
make sense anymore.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174829
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sun, 10 Feb 2013 06:42:30 +0000 (06:42 +0000)]
Make LiveVariables an instance variable of PHIElimination.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174828
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 10 Feb 2013 06:36:29 +0000 (06:36 +0000)]
ok, ok, stop fighting type punning warnings by just using a union.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174827
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 10 Feb 2013 06:07:16 +0000 (06:07 +0000)]
hopefully "really" fix a type punning warning by defining the buffer as
type char, which can't have TBAA tags.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174826
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 10 Feb 2013 05:45:34 +0000 (05:45 +0000)]
attempt to defeat a gcc warning that is breaking a -Werror buildbot.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174825
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sun, 10 Feb 2013 05:00:40 +0000 (05:00 +0000)]
Add accessor for the LLVMContext.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174824
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Sat, 9 Feb 2013 20:54:05 +0000 (20:54 +0000)]
Remove unneeded "TargetMachine.h" #includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174817
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 9 Feb 2013 15:48:49 +0000 (15:48 +0000)]
TEMPORARY SYNTAX CHANGE!
The original syntax for the attribute groups was ambiguous. For example:
declare void @foo() #1
#0 = attributes { noinline }
The '#0' would be parsed as an attribute reference for '@foo' and not as a
top-level entity. In order to continue forward while waiting for a decision on
what the correct syntax is, I'm changing it to this instead:
declare void @foo() #1
attributes #0 = { noinline }
Repeat: This is TEMPORARY until we decide what the correct syntax should be.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174813
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 9 Feb 2013 15:42:51 +0000 (15:42 +0000)]
Add a DenseMapInfo class for the AttributeSet.
We are going to place the AttributeSet into a DenseMap during assembly writing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174812
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Holewinski [Sat, 9 Feb 2013 13:34:15 +0000 (13:34 +0000)]
[NVPTX] Make address space errors more explicit (llvm_unreachable -> report_fatal_error)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174808
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Sat, 9 Feb 2013 13:29:31 +0000 (13:29 +0000)]
Simplify code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174807
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Sat, 9 Feb 2013 13:29:10 +0000 (13:29 +0000)]
Remove unneeded #includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174806
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 9 Feb 2013 07:37:59 +0000 (07:37 +0000)]
This is the correct version of r174802.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174804
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 9 Feb 2013 07:37:26 +0000 (07:37 +0000)]
Fix a nasty off-by one error that only manifests with 64-bit word size (which is
not enabled yet).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174803
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 9 Feb 2013 07:07:29 +0000 (07:07 +0000)]
Fix the underlying problem that was causing read(0) to be called: sometimes the
bitcode writer would generate abbrev records saying that the abbrev should be
filled with fixed zero-bit bitfields (this happens in the .bc writer when
the number of types used in a module is exactly one, since log2(1) == 0).
In this case, just handle it as a literal zero. We can't "just fix" the writer
without breaking compatibility with existing bc files, so have the abbrev reader
do the substitution.
Strengthen the assert in read to reject reads of zero bits so we catch such
crimes in the future, and remove the special case designed to handle this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174801
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 9 Feb 2013 06:52:14 +0000 (06:52 +0000)]
recommit r173072 (preparing bitstream reader to read a machine word at a time,
instead of always 32-bits at a time) with two changes:
1. Make Read(0) always return zero without affecting the state of our cursor.
2. Hack word_t to always be 32 bits, as staging.
These two caveats will change shortly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174800
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Sat, 9 Feb 2013 01:19:12 +0000 (01:19 +0000)]
Remove trailing spaces.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174791
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Sat, 9 Feb 2013 01:15:18 +0000 (01:15 +0000)]
Remove unneeded #includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174790
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Sat, 9 Feb 2013 01:11:01 +0000 (01:11 +0000)]
LSR IVChain improvement.
Handle chains in which the same offset is used for both loads and
stores to the same array.
Fixes rdar://
11410078.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174789
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Sat, 9 Feb 2013 01:04:28 +0000 (01:04 +0000)]
Remove #includes from the commonly used LoopInfo.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174786
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Sat, 9 Feb 2013 00:41:44 +0000 (00:41 +0000)]
Dwarf: do not use line_table_start in at_stmt_list since we do not always emit
line table entries in assembly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174785
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 9 Feb 2013 00:04:07 +0000 (00:04 +0000)]
Remove the old liveness algorithm.
This is part of the plan to delete LiveVariables.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174783
91177308-0d34-0410-b5e6-
96231b3b80d8
Sergei Larin [Fri, 8 Feb 2013 23:37:41 +0000 (23:37 +0000)]
Enable *BasicBlockPass::createPrinterPass()
Enables raw_ostream I/O for BasicBlockPass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174776
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 8 Feb 2013 23:18:37 +0000 (23:18 +0000)]
Turn on -new-live-intervals by default.
This uses a liveness algorithm that does not depend on data from the
LiveVariables analysis, it is the first step towards removing
LiveVariables completely.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174774
91177308-0d34-0410-b5e6-
96231b3b80d8
Jordan Rose [Fri, 8 Feb 2013 23:17:31 +0000 (23:17 +0000)]
CMake: Use check_symbol_exists instead of check_cxx_symbol_exists.
check_cxx_symbol_exists requires CMake 2.8.6, so even though I
recommended it to Owen it's probably better to stay away for now.
This check is not technically correct because we're checking <math.h>
but then using <cmath> in the actual code, but if we run into problems we
can do the same sort of dance as isinf() and isnan() where we check /both/
headers and then write a wrapper header around them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174773
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Fri, 8 Feb 2013 22:55:10 +0000 (22:55 +0000)]
configure: when performing a compiler feature test for a -Wno-foo flag, attempt
to use -Wfoo instead of -Wno-foo. This works around a bug in some versions of
gcc, where it will silently accept an unknown -Wno-foo option, but will
generate an error for a compile which uses -Wno-foo if that compile also
triggers any warnings.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174770
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 8 Feb 2013 22:24:41 +0000 (22:24 +0000)]
ReleaseNotes: Add section for R600 backend
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174764
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 8 Feb 2013 22:24:40 +0000 (22:24 +0000)]
R600: Dump the function name when TargetLowering::LowerCall() fails
Also output a more useful error message.
NOTE: This is a candidate for the Mesa stable branch
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174763
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 8 Feb 2013 22:24:38 +0000 (22:24 +0000)]
R600: rework flow creation in the structurizer v2
This fixes a couple of bugs and incorrect assumptions,
in total four more piglit tests now pass.
v2: fix small bug in the dominator updating
Patch by: Christian König
Signed-off-by: Christian König <christian.koenig@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174762
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 8 Feb 2013 22:24:37 +0000 (22:24 +0000)]
R600: fix loop analyses in the structurizer
Patch by: Christian König
Intersecting loop handling was wrong.
Signed-off-by: Christian König <christian.koenig@amd.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174761
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 8 Feb 2013 22:24:35 +0000 (22:24 +0000)]
R600: fix PHI value adding in the structurizer
Otherwise we sometimes produce invalid code.
Patch by: Christian König
Signed-off-by: Christian König <christian.koenig@amd.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174760
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 8 Feb 2013 22:01:47 +0000 (22:01 +0000)]
Minor cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174756
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Fri, 8 Feb 2013 21:51:26 +0000 (21:51 +0000)]
[docs] Tweaks for clarity, readability, and correctness.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174749
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Fri, 8 Feb 2013 21:48:29 +0000 (21:48 +0000)]
Revert "Add LLVMContext::emitWarning methods and use them. <rdar://problem/
12867368>"
This reverts r171041. This was a nice idea that didn't work out well.
Clang warnings need to be associated with warning groups so that they can
be selectively disabled, promoted to errors, etc. This simplistic patch didn't
allow for that. Enhancing it to provide some way for the backend to specify
a front-end warning type seems like overkill for the few uses of this, at
least for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174748
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Fri, 8 Feb 2013 21:42:56 +0000 (21:42 +0000)]
Add the 16 bit version of addiu. To the assembler, the 16 and 32 bit are the
same so we put in the comment field an indicator when we think we are
emitting the 16 bit version. For the direct object emitter, the difference is
important as well as for other passes which need an accurate count of
program size. There will be other similar putbacks to this for various
instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174747
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Feb 2013 21:35:47 +0000 (21:35 +0000)]
DAGCombiner: Constant folding around pre-increment loads/stores
Previously, even when a pre-increment load or store was generated,
we often needed to keep a copy of the original base register for use
with other offsets. If all of these offsets are constants (including
the offset which was combined into the addressing mode), then this is
clearly unnecessary. This change adjusts these other offsets to use the
new incremented address.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174746
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Feb 2013 21:13:39 +0000 (21:13 +0000)]
BBVectorize: Use TTI->getAddressComputationCost
This is a follow-up to the cost-model change in r174713 which splits
the cost of a memory operation between the address computation and the
actual memory access. In r174713, this cost is always added to the
memory operation cost, and so BBVectorize will do the same.
Currently, this new cost function is used only by ARM, and I don't
have any ARM test cases for BBVectorize. Assistance in generating some
good ARM test cases for BBVectorize would be greatly appreciated!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174743
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Fri, 8 Feb 2013 21:08:22 +0000 (21:08 +0000)]
More modifications to PowerPC doc links
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174742
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Fri, 8 Feb 2013 20:35:15 +0000 (20:35 +0000)]
Revert 172027 and 174336. Remove diagnostics about over-aligned stack objects.
Aside from the question of whether we report a warning or an error when we
can't satisfy a requested stack object alignment, the current implementation
of this is not good. We're not providing any source location in the diagnostics
and the current warning is not connected to any warning group so you can't
control it. We could improve the source location somewhat, but we can do a
much better job if this check is implemented in the front-end, so let's do that
instead. <rdar://problem/
13127907>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174741
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Feb 2013 20:24:46 +0000 (20:24 +0000)]
Update PowerPC links in CompilerWriterInfo.rst
This updates the current references to links that work for me.
In the future, we should update the list of references itself to provide
information on newer architecture variants.
Thanks to Sean Silva for pointing out that the current links were broken!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174739
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Fri, 8 Feb 2013 18:19:17 +0000 (18:19 +0000)]
Refine fix to bug 15041.
Thanks to help from Nadav and Hal, I have a more reasonable (and even
correct!) approach. This specifically penalizes the insertelement
and extractelement operations for the performance hit that will occur
on PowerPC processors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174725
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 8 Feb 2013 18:00:14 +0000 (18:00 +0000)]
[SimplifyLibCalls] Library call simplification doen't work if the call site
isn't using the default calling convention. However, if the transformation is
from a call to inline IR, then the calling convention doesn't matter.
rdar://
13157990
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174724
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 8 Feb 2013 17:43:32 +0000 (17:43 +0000)]
Typos.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174723
91177308-0d34-0410-b5e6-
96231b3b80d8
David Tweed [Fri, 8 Feb 2013 16:35:10 +0000 (16:35 +0000)]
The patch to fix some issues in r174543 fixed the lines failing the test, but missed a couple
of lines which weren't being explicitly looked at and were printing incorrect results. These
values clearly must lie within 32 bits, so the casts are definitely safe.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174717
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Fri, 8 Feb 2013 14:50:48 +0000 (14:50 +0000)]
ARM cost model: Address computation in vector mem ops not free
Adds a function to target transform info to query for the cost of address
computation. The cost model analysis pass now also queries this interface.
The code in LoopVectorize adds the cost of address computation as part of the
memory instruction cost calculation. Only there, we know whether the instruction
will be scalarized or not.
Increase the penality for inserting in to D registers on swift. This becomes
necessary because we now always assume that address computation has a cost and
three is a closer value to the architecture.
radar://
13097204
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174713
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Fri, 8 Feb 2013 14:34:33 +0000 (14:34 +0000)]
Update tests for DWARF parser: store sources next to pre-built object files and provide build instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174711
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Kuperstein [Fri, 8 Feb 2013 12:58:29 +0000 (12:58 +0000)]
Test Commit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174709
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 8 Feb 2013 06:32:06 +0000 (06:32 +0000)]
Parse the attribute group reference on a function.
Attribute references are of this form:
define void @foo() #0 #1 #2 { ... }
Parse them for function attributes. If there's more than one reference, then
they are merged together.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174697
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Fri, 8 Feb 2013 03:57:41 +0000 (03:57 +0000)]
When Mips16 frames grow large, the immediate field may exceed the maximum
allowed size for the instruction. This code uses RegScavenger to fix this.
We sometimes need 2 registers for Mips16 so we must handle things
differently than how register scavenger is normally used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174696
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 8 Feb 2013 01:55:39 +0000 (01:55 +0000)]
Revert "Have InstCombine call SipmlifyCall when handling calls. Test case included."
This reverts commit
3854a5d90fee52af1065edbed34521fff6cdc18d.
This causes a clang unit test to hang: vtable-available-externally.cpp.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174692
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Fri, 8 Feb 2013 01:05:48 +0000 (01:05 +0000)]
[readobj] Fix memory leak.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174687
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 8 Feb 2013 00:52:31 +0000 (00:52 +0000)]
Use ParseFnAttributeValuePairs instead of ParseOptionalFuncAttrs
The functionality of ParseOptionalFuncAttrs was there in
ParseFnAttributeValuePairs. So just use that instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174686
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 7 Feb 2013 23:37:58 +0000 (23:37 +0000)]
[Object][ELF] Add a way to get the dynamic symbol table section.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174682
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 7 Feb 2013 23:37:46 +0000 (23:37 +0000)]
[Object][ELF] Add {begin,end}_elf_dynamic_symbols to non-virtually iterate over dynamic symbols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174681
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Ilseman [Thu, 7 Feb 2013 23:01:35 +0000 (23:01 +0000)]
Have InstCombine call SipmlifyCall when handling calls. Test case included.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174675
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Thu, 7 Feb 2013 22:34:07 +0000 (22:34 +0000)]
fix 80-col violation and fix the docs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174671
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Thu, 7 Feb 2013 21:34:34 +0000 (21:34 +0000)]
[lit] Add a display function for lit.Test.TestResult.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174667
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Thu, 7 Feb 2013 21:32:32 +0000 (21:32 +0000)]
[mips] Make Filler a class and reduce indentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174666
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 7 Feb 2013 21:19:56 +0000 (21:19 +0000)]
Formatting.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174664
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 7 Feb 2013 21:19:50 +0000 (21:19 +0000)]
"Clean up" line section symbol emission by emitting the section
syms before constructing the compile units so we're not emitting
section references to sections not there already.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174663
91177308-0d34-0410-b5e6-
96231b3b80d8
Will Schmidt [Thu, 7 Feb 2013 20:43:33 +0000 (20:43 +0000)]
[patch] bug 15055 Add Unistd.h to OProfileWrapper.cpp
Add #include <unistd.h> to OProfileWrapper.cpp. This provides the declarations for 'read' and 'close' that are otherwise missing, and result in 'error: <foo> was not declared in this scope'.
This matches the issue as reported in bug 15055 "Can no longer compile LLVM with --with-oprofile"
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174661
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Schmidt [Thu, 7 Feb 2013 20:33:57 +0000 (20:33 +0000)]
Constrain PowerPC autovectorization to fix bug 15041.
Certain vector operations don't vectorize well with the current
PowerPC implementation. Element insert/extract performs poorly
without VSX support because Altivec requires going through memory.
SREM, UREM, and VSELECT all produce bad scalar code.
There's a lot of work to do for the cost model before
autovectorization will be tuned well, and this is not an attempt to
address the larger problem.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174660
91177308-0d34-0410-b5e6-
96231b3b80d8