Bill Wendling [Mon, 20 Jun 2011 22:12:24 +0000 (22:12 +0000)]
Remove the subclassing. This will be moved to the ASM printer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133473
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 20 Jun 2011 20:56:05 +0000 (20:56 +0000)]
Skip fields that don't exist in the Register class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133470
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Holewinski [Mon, 20 Jun 2011 18:42:48 +0000 (18:42 +0000)]
PTX: Fix conversion between predicates and value types
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133454
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Mon, 20 Jun 2011 18:33:26 +0000 (18:33 +0000)]
Emit movq for 64-bit register to XMM register moves, but continue to accept
movd when assembling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133452
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Mon, 20 Jun 2011 17:46:19 +0000 (17:46 +0000)]
Fix a check for PHINodes with two incoming values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133449
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Holewinski [Mon, 20 Jun 2011 17:08:56 +0000 (17:08 +0000)]
PTX: Fix if-then-else formatting and add missing asserts
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133447
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 20 Jun 2011 17:04:08 +0000 (17:04 +0000)]
Disable again.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133446
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Holewinski [Mon, 20 Jun 2011 15:56:20 +0000 (15:56 +0000)]
PTX: Add basic register spilling code
The current implementation generates stack loads/stores, which are
really just mov instructions from/to "special" registers. This may
not be the most efficient implementation, compared to an approach where
the stack registers are directly folded into instructions, but this is
easier to implement and I have yet to see a case where ptxas is unable
to see through this kind of register usage and know what is really
going on.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133443
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Mon, 20 Jun 2011 15:28:39 +0000 (15:28 +0000)]
Don't apply on PPC64 the 32bit ADDIC optimizations as there's no overflow
with 32bit values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133439
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Mon, 20 Jun 2011 14:46:47 +0000 (14:46 +0000)]
This is an automatically reduced test case that crashed in GVN, at some
point during the development of the phi operand changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133436
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Mon, 20 Jun 2011 14:38:01 +0000 (14:38 +0000)]
Change how PHINodes store their operands.
Change PHINodes to store simple pointers to their incoming basic blocks,
instead of full-blown Uses.
Note that this loses an optimization in SplitCriticalEdge(), because we
can no longer walk the use list of a BasicBlock to find phi nodes. See
the comment I removed starting "However, the foreach loop is slow for
blocks with lots of predecessors".
Extend replaceAllUsesWith() on a BasicBlock to also update any phi
nodes in the block's successors. This mimics what would have happened
when PHINodes were proper Users of their incoming blocks. (Note that
this only works if OldBB->replaceAllUsesWith(NewBB) is called when
OldBB still has a terminator instruction, so it still has some
successors.)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133435
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Mon, 20 Jun 2011 14:18:48 +0000 (14:18 +0000)]
Make better use of the PHINode API.
Change various bits of code to make better use of the existing PHINode
API, to insulate them from forthcoming changes in how PHINodes store
their operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133434
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Mon, 20 Jun 2011 14:12:33 +0000 (14:12 +0000)]
Remove the AugmentedUse struct.
I don't think the AugmentedUse struct buys us much, either in
correctness or in ease of use. Ditch it, and simplify Use::getUser() and
User::allocHungoffUses().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133433
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 20 Jun 2011 14:11:42 +0000 (14:11 +0000)]
Re enable 133415 with two fixes
* Don't introduce a duplicated bb in the CFG
* When making a branch unconditional, clear the PredCond array so that it
is really unconditional.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133432
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Mon, 20 Jun 2011 09:26:23 +0000 (09:26 +0000)]
Disable the logic added by rafael in commit 133415 to see if it brings the
dragonegg buildbots back to life. Original commit message:
Teach early dup how to duplicate basic blocks with one successor and only phi instructions
into more complex blocks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133430
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 20 Jun 2011 07:15:58 +0000 (07:15 +0000)]
Fix PromoteIntRes_TRUNCATE: Add support for cases where the
source vector type is to be split while the target vector is to be promoted.
(eg: <4 x i64> -> <4 x i8> )
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133424
91177308-0d34-0410-b5e6-
96231b3b80d8
Francois Pichet [Mon, 20 Jun 2011 05:19:37 +0000 (05:19 +0000)]
Fix MSVC build. next() function already exists in the MSVC headers. This create a overload conflict. Make sure we pick up the llvm one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133416
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Mon, 20 Jun 2011 04:16:35 +0000 (04:16 +0000)]
Teach early dup how to duplicate basic blocks with one successor and only phi instructions
into more complex blocks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133415
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 20 Jun 2011 04:01:31 +0000 (04:01 +0000)]
Revamp the "ConstantStruct::get" methods. Previously, these were scattered
all over the place in different styles and variants. Standardize on two
preferred entrypoints: one that takes a StructType and ArrayRef, and one that
takes StructType and varargs.
In cases where there isn't a struct type convenient, we now add a
ConstantStruct::getAnon method (whose name will make more sense after a few
more patches land).
It would be "really really nice" if the ConstantStruct::get and
ConstantVector::get methods didn't make temporary std::vectors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133412
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 20 Jun 2011 03:51:04 +0000 (03:51 +0000)]
introduce an isLayoutIdentical() method, which is currently just a pointer
equality check.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133409
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 20 Jun 2011 02:50:54 +0000 (02:50 +0000)]
Add a RegisterTuples class to Target.td and TableGen.
A RegisterTuples instance is used to synthesize super-registers by
zipping together lists of sub-registers. This is useful for generating
pseudo-registers representing register sequence constraints like 'two
consecutive GPRs', or 'an even-odd pair of floating point registers'.
The RegisterTuples def can be used in register set operations when
building register classes. That is the only way of accessing the
synthesized super-registers.
For example, the ARM QQ register class of pseudo-registers could have
been formed like this:
// Form pairs Q0_Q1, Q2_Q3, ...
def QQPairs : RegisterTuples<[qsub_0, qsub_1],
[(decimate QPR, 2),
(decimate (shl QPR, 1), 2)]>;
def QQ : RegisterClass<..., (add QQPairs)>;
Similarly, pseudo-registers representing '3 consecutive D-regs with
wraparound' look like:
// Form D0_D1_D2, D1_D2_D3, ..., D30_D31_D0, D31_D0_D1.
def DSeqTriples : RegisterTuples<[dsub_0, dsub_1, dsub_2],
[(rotl DPR, 0),
(rotl DPR, 1),
(rotl DPR, 2)]>;
TableGen automatically computes aliasing information for the synthesized
registers.
Register tuples are still somewhat experimental. We still need to see
how they interact with MC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133407
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Sun, 19 Jun 2011 18:37:11 +0000 (18:37 +0000)]
Fix a FIXME by making GlobalVariable::getInitializer() return a
const Constant *.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133400
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 19 Jun 2011 12:14:34 +0000 (12:14 +0000)]
Update test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133390
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 19 Jun 2011 10:49:57 +0000 (10:49 +0000)]
Code cleanups: Remove duplicated logic in PromotInteRes_BITCAST, reserve vector space, reuse types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133389
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 19 Jun 2011 10:22:39 +0000 (10:22 +0000)]
Calls to AssertZext and getZeroExtendInReg must be made using scalar types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133388
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 19 Jun 2011 08:49:38 +0000 (08:49 +0000)]
When promoting the vector elements in CopyToParts, use vector trunc
instead of scalarizing, and doing an element-by-element truncat.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133382
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 19 Jun 2011 08:12:43 +0000 (08:12 +0000)]
Reduce the runtime of the test. Keep only the interesting cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133381
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 19 Jun 2011 03:30:32 +0000 (03:30 +0000)]
Revert r133373. I was going to use this to teach the Verifier to verify constant
expressions, but Chris wants to instead reduce the set of possible constant
expression types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133374
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sun, 19 Jun 2011 02:26:33 +0000 (02:26 +0000)]
Add the remaining instructions/constant expressions as Operators so that code
can manipulate instructions and constantexpr's uniformly. No users yet though.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133373
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 19 Jun 2011 00:03:46 +0000 (00:03 +0000)]
Remove support for parsing the "type i32" syntax for defining a numbered
top level type without a specified number. This syntax isn't documented
and blocks forward progress.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133371
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 23:51:31 +0000 (23:51 +0000)]
revert r133368, apparently I missed the tests to be updated.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133369
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 23:38:57 +0000 (23:38 +0000)]
Remove support for parsing the "type i32" syntax for defining a numbered
top level type without a specified number. This asmprinter has never
generated this, as you can tell by no tests being updated. It also isn't
documented.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133368
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 22:48:56 +0000 (22:48 +0000)]
fix the varargs version of StructType::get to not require an LLVMContext, making usage
much cleaner.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133364
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 22:15:47 +0000 (22:15 +0000)]
eliminate some pointless virtual methods.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133363
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 21:46:23 +0000 (21:46 +0000)]
simplify some code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133362
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 21:23:04 +0000 (21:23 +0000)]
now that Type::getDescription() is dead, the TypePrinting class can move from Assembly/Writer.h to being
a private class in AsmWriter.cpp.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133361
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 21:18:23 +0000 (21:18 +0000)]
eliminate the Type::getDescription() method, using "<<" instead. This
removes some gunk from LLVMContext.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133360
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 21:02:49 +0000 (21:02 +0000)]
improve some comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133359
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 19:12:59 +0000 (19:12 +0000)]
remove an unreduced testcase.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133356
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 18:56:39 +0000 (18:56 +0000)]
rework the remaining autoupgrade logic to use a StringRef instead of creating a
temporary std::string for every function being checked.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133355
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 18 Jun 2011 14:42:47 +0000 (14:42 +0000)]
Directly print to a raw_ostream instead of printing to a buffer first.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133352
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 18 Jun 2011 14:42:42 +0000 (14:42 +0000)]
Simplify code. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133351
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 18 Jun 2011 13:53:47 +0000 (13:53 +0000)]
Simplify code. No change in functionality.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133350
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Sat, 18 Jun 2011 13:51:54 +0000 (13:51 +0000)]
MC: Allow .common as alias for .comm assembler directive. PR10116.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133349
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 18 Jun 2011 13:13:44 +0000 (13:13 +0000)]
Don't allocate empty read-only SmallVectors during SelectionDAG deallocation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133348
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 18 Jun 2011 11:09:41 +0000 (11:09 +0000)]
Remove unused but set variables.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133347
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Sat, 18 Jun 2011 10:28:47 +0000 (10:28 +0000)]
Fix PR10103: Less code for enum type translation.
In cases such as the attached test, where the case value for a switch
destination is used in a phi node that follows the destination, it
might be better to replace that value with the condition value of the
switch, so that more blocks can be folded away with
TryToSimplifyUncondBranchFromEmptyBlock because there are less
conflicts in the phi node.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133344
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Sat, 18 Jun 2011 07:23:25 +0000 (07:23 +0000)]
Add test for r133251.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133339
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sat, 18 Jun 2011 06:17:51 +0000 (06:17 +0000)]
When scalar replacement returns a vector type, only accept it if the vector
type's bitwidth matches the (allocated) size of the alloca. This severely
pessimizes vector scalar replacement when the only vector type being used is
something like <3 x float> on x86 or ARM whose allocated size matches a
<4 x float>.
I hope to fix some of the flawed assumptions about allocated size throughout
scalar replacement and reenable this in most cases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133338
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 18 Jun 2011 06:05:24 +0000 (06:05 +0000)]
rip out a ton of intrinsic modernization logic from AutoUpgrade.cpp, which is
for pre-2.9 bitcode files. We keep x86 unaligned loads, movnt, crc32, and the
target indep prefetch change.
As usual, updating the testsuite is a PITA.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133337
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sat, 18 Jun 2011 05:47:49 +0000 (05:47 +0000)]
Fix an invalid bitcast crash that occurs when doing a partial memset of a vector
alloca. Fixes part of <rdar://problem/
9580800>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133336
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron Zwarich [Sat, 18 Jun 2011 05:47:45 +0000 (05:47 +0000)]
Remove a pointless assignment. Nothing checks the value of VectorTy anymore now
unless ScalarKind is Vector.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133335
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 18 Jun 2011 05:44:55 +0000 (05:44 +0000)]
Use the correct comparator to avoid depending on pointer values.
This should fix the Linux buildbots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133334
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 18 Jun 2011 04:26:06 +0000 (04:26 +0000)]
Store CodeGenRegisters as pointers so they won't be reallocated.
Reuse the CodeGenRegBank DenseMap in a few places that would build their
own or use linear search.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133333
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 18 Jun 2011 03:08:20 +0000 (03:08 +0000)]
Remove MethodProtos/MethodBodies and allocation_order_begin/end.
Targets that need to change the default allocation order should use the
AltOrders mechanism instead. See the X86 and ARM targets for examples.
The allocation_order_begin() and allocation_order_end() methods have been
replaced with getRawAllocationOrder(), and there is further support
functions in RegisterClassInfo.
It is no longer possible to insert arbitrary code into generated
register classes. This is a feature.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133332
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 18 Jun 2011 02:30:02 +0000 (02:30 +0000)]
Delete unneeded allocation order override.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133331
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 18 Jun 2011 01:14:46 +0000 (01:14 +0000)]
Switch ARM to using AltOrders instead of MethodBodies.
This slightly changes the GPR allocation order on Darwin where R9 is not
a callee-saved register:
Before: %R0 %R1 %R2 %R3 %R12 %R9 %LR %R4 %R5 %R6 %R8 %R10 %R11
After: %R0 %R1 %R2 %R3 %R9 %R12 %LR %R4 %R5 %R6 %R8 %R10 %R11
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133326
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 18 Jun 2011 01:14:43 +0000 (01:14 +0000)]
Switch x86 to using AltOrders instead of MethodBodies.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133325
91177308-0d34-0410-b5e6-
96231b3b80d8
Galina Kistanova [Sat, 18 Jun 2011 00:59:37 +0000 (00:59 +0000)]
Moved to the right place.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133324
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 18 Jun 2011 00:53:27 +0000 (00:53 +0000)]
Reserve D16-D13 on subtargets that don't support them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133321
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 18 Jun 2011 00:50:49 +0000 (00:50 +0000)]
Provide AltOrders for specifying alternative allocation orders.
A register class can define AltOrders and AltOrderSelect instead of
defining method protos and bodies. The AltOrders lists can be defined
with set operations, and TableGen can verify that the alternative
allocation orders only contain valid registers.
This is currently an opt-in feature, and it is still possible to
override allocation_order_begin/end. That will not be true for long.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133320
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 18 Jun 2011 00:19:35 +0000 (00:19 +0000)]
* Override the "EmitBytes" function, since it can sneak values in that way.
* Make this used only if CFI is used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133319
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Sat, 18 Jun 2011 00:09:57 +0000 (00:09 +0000)]
Fix UMULO support for 2x register width to allow the full
range without a libcall to a new mulo<mode> libcall
that we'd have to create.
Finishes the rest of rdar://
9090077 and rdar://
9210061
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133318
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 17 Jun 2011 23:42:01 +0000 (23:42 +0000)]
Remove false assertion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133314
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 17 Jun 2011 23:26:52 +0000 (23:26 +0000)]
Only call TRI::getRawAllocationOrder to resolve a target-dependent hint.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133313
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 17 Jun 2011 23:17:13 +0000 (23:17 +0000)]
Zap the last reference to allocation_order_begin().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133310
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 17 Jun 2011 23:15:00 +0000 (23:15 +0000)]
SI, DI, BP, and SP don't have 8-bit sub-registers in x86 mode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133308
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 17 Jun 2011 22:35:59 +0000 (22:35 +0000)]
Fix comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133307
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Beaumont-Gay [Fri, 17 Jun 2011 22:21:12 +0000 (22:21 +0000)]
Fix -Asserts build
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133305
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 17 Jun 2011 22:08:25 +0000 (22:08 +0000)]
Revert r133285. Causing odd failures on Dragonegg.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133301
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 17 Jun 2011 21:44:15 +0000 (21:44 +0000)]
Disable for another investigation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133299
91177308-0d34-0410-b5e6-
96231b3b80d8
Devang Patel [Fri, 17 Jun 2011 21:36:44 +0000 (21:36 +0000)]
Set debug loc for new preheader's terminator.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133298
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 17 Jun 2011 21:31:43 +0000 (21:31 +0000)]
Support only DwarfCFI or SjLj exception handling in LSDA decoder.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133297
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 17 Jun 2011 21:29:06 +0000 (21:29 +0000)]
SjLj exception handling LSDA decoding support wasn't represented correctly. Use
the correct values, etc. In particular, the exception handling type is SjLj, not
ARM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133296
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 17 Jun 2011 21:09:50 +0000 (21:09 +0000)]
Disable to investigate ARM failure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133293
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 17 Jun 2011 20:55:01 +0000 (20:55 +0000)]
Use the verbose asm flag instead of a new flag for decoding the LSDA.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133292
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Fri, 17 Jun 2011 20:54:12 +0000 (20:54 +0000)]
Fix a bug in the type-lowering of integer-promoted elements. Add a check that
the newly created simple type is valid before checking its legality.
Re-commit the test file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133291
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 17 Jun 2011 20:47:21 +0000 (20:47 +0000)]
Add an alternative rev16 pattern. We should figure out a better way to handle these complex rev patterns. rdar://
9609108
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133289
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 17 Jun 2011 20:41:29 +0000 (20:41 +0000)]
Lower multiply with overflow checking to __mulo<mode>
calls if we haven't been able to lower them any
other way.
Fixes rdar://
9090077 and rdar://
9210061
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133288
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Fri, 17 Jun 2011 20:35:21 +0000 (20:35 +0000)]
Add an option that allows one to "decode" the LSDA.
The LSDA is a bit difficult for the non-initiated to read. Even with comments,
it's not always clear what's going on. This wraps the ASM streamer in a class
that retains the LSDA and then emits a human-readable description of what's
going on in it.
So instead of having to make sense of:
Lexception1:
.byte 255
.byte 155
.byte 168
.space 1
.byte 3
.byte 26
Lset0 = Ltmp7-Leh_func_begin1
.long Lset0
Lset1 = Ltmp812-Ltmp7
.long Lset1
Lset2 = Ltmp913-Leh_func_begin1
.long Lset2
.byte 3
Lset3 = Ltmp812-Leh_func_begin1
.long Lset3
Lset4 = Leh_func_end1-Ltmp812
.long Lset4
.long 0
.byte 0
.byte 1
.byte 0
.byte 2
.byte 125
.long __ZTIi@GOTPCREL+4
.long __ZTIPKc@GOTPCREL+4
you can read this instead:
## Exception Handling Table: Lexception1
## @LPStart Encoding: omit
## @TType Encoding: indirect pcrel sdata4
## @TType Base: 40 bytes
## @CallSite Encoding: udata4
## @Action Table Size: 26 bytes
## Action 1:
## A throw between Ltmp7 and Ltmp812 jumps to Ltmp913 on an exception.
## For type(s): __ZTIi@GOTPCREL+4 __ZTIPKc@GOTPCREL+4
## Action 2:
## A throw between Ltmp812 and Leh_func_end1 does not have a landing pad.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133286
91177308-0d34-0410-b5e6-
96231b3b80d8
Stuart Hastings [Fri, 17 Jun 2011 20:21:52 +0000 (20:21 +0000)]
Relocate NUW test to cover all binary ops in a dynamic alloca expr.
Followup to 132926. rdar://problem/
9265821
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133285
91177308-0d34-0410-b5e6-
96231b3b80d8
Galina Kistanova [Fri, 17 Jun 2011 18:26:23 +0000 (18:26 +0000)]
est 2008-06-04-indirectmem.ll is X86-specific. Move to X86 folder.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133275
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 17 Jun 2011 18:17:37 +0000 (18:17 +0000)]
Drop the "2" suffix on some enums.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133274
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 17 Jun 2011 18:09:11 +0000 (18:09 +0000)]
remove support for a bunch of obsolete instruction encodings
and other backward compatibility hacks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133273
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 17 Jun 2011 18:05:30 +0000 (18:05 +0000)]
Remove a useless copy of MCELFStreamer. Patch by Logan Chien!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133272
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Fri, 17 Jun 2011 18:00:21 +0000 (18:00 +0000)]
getSuccWeight returns now default 0 if Weights vector is empty.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133271
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 17 Jun 2011 17:56:00 +0000 (17:56 +0000)]
missed a file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133270
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 17 Jun 2011 17:50:30 +0000 (17:50 +0000)]
Remove some "2" suffixes from the metadata enums now that "1" is gone.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133269
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 17 Jun 2011 17:48:53 +0000 (17:48 +0000)]
remove bitcode reader support for LLVM 2.7 metadata encoding.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133268
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 17 Jun 2011 17:40:18 +0000 (17:40 +0000)]
remove another old and dead hunk of code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133267
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Fri, 17 Jun 2011 17:37:13 +0000 (17:37 +0000)]
Stop accepting and ignoring attributes in function types. Attributes are applied
to functions and call/invokes, not to types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133266
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakub Staszak [Fri, 17 Jun 2011 17:30:10 +0000 (17:30 +0000)]
Allow empty Weights vector.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133265
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Fri, 17 Jun 2011 15:21:10 +0000 (15:21 +0000)]
Fix a few places where 32bit instructions/registerset were used on PPC64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133260
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 17 Jun 2011 14:16:17 +0000 (14:16 +0000)]
Test for previous commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133256
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 17 Jun 2011 13:59:43 +0000 (13:59 +0000)]
Two fixes relating to debug value:
* We should change the generated code because of a debug use.
* Avoid creating debug uses of undef, as they become a kill.
Test to follow.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133255
91177308-0d34-0410-b5e6-
96231b3b80d8
Jay Foad [Fri, 17 Jun 2011 13:36:06 +0000 (13:36 +0000)]
Fix typo in comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133254
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Holewinski [Fri, 17 Jun 2011 12:12:42 +0000 (12:12 +0000)]
PTX: Adjust rounding modes
* rounding modes for fp add, mul, sub now use .rn
* float -> int rounding correctly uses .rzi not .rni
* 32bit fdiv for sm13 uses div.rn (instead of div.approx)
* 32bit fdiv for sm10 now uses div (instead of div.approx)
Approx is not IEEE 754 compatible (and should be optionally set by a flag to the backend instead). The .rn rounding modifier is the PTX default anyway, but it's better to be explicit.
All these modifiers should be available by using __fmul_rz functions for example, but support will need to be added for this in the backend.
Patch by Dan Bailey
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133253
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Fri, 17 Jun 2011 11:08:09 +0000 (11:08 +0000)]
Don't force remove config.cache on reconfiguration.
config.cache will be used by the person who specifies '-C' to configure.
config.cache's inconsistency should be responsible to him.
Re-configuration would spend so much on cygming without '-C', esp. cygwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133252
91177308-0d34-0410-b5e6-
96231b3b80d8
Nick Lewycky [Fri, 17 Jun 2011 10:09:00 +0000 (10:09 +0000)]
When promoting an alloca to registers discard any lifetime intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133251
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Fri, 17 Jun 2011 07:09:01 +0000 (07:09 +0000)]
Add a hook for PBQP clients to run a custom pre-alloc pass to run prior to PBQP allocation. Patch by Arnaud Allard de Grandmaison.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@133249
91177308-0d34-0410-b5e6-
96231b3b80d8