Commit Graph

14190 Commits

Author SHA1 Message Date
Nadav Rotem
f55ef64544 Generate better select code by allowing the target to use scalar select, and not sign-extend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163086 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-02 08:20:07 +00:00
Pete Cooper
d906017c1a Only legalise a VSELECT in to bitwise operations if the vector mask bool is zeros or all ones. A vector bool with just ones isn't suitable for masking with.
No test case unfortunately as i couldn't find a target which fit all
the conditions needed to hit this code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163075 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-01 22:27:48 +00:00
Pete Cooper
0fc44aba18 Revert "Take account of boolean vector contents when promoting a build vector from i1 to some other type. rdar://problem/12210060"
This reverts commit 5dd9e214fb.

Thanks to Duncan for explaining how this should have been done.

Conflicts:

	test/CodeGen/X86/vec_select.ll

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163064 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-01 17:37:55 +00:00
Logan Chien
e2ac552421 Fix typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163059 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-01 12:11:41 +00:00
Owen Anderson
58d5729540 Teach DAG combine a number of tricks to simplify FMA expressions in fast-math mode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163051 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-01 06:04:27 +00:00
Michael Liao
b79bff50bd Fix typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163049 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-01 04:09:16 +00:00
Jakob Stoklund Olesen
94083149fd Add MachineInstr::tieOperands, remove setIsTied().
Manage tied operands entirely internally to MachineInstr. This makes it
possible to change the representation of tied operands, as I will do
shortly.

The constraint that tied uses and defs must be in the same order was too
restrictive.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163021 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-31 20:50:53 +00:00
Craig Topper
89b2ff0b5c Use CloneMachineInstr to make a new MI in commuteInstruction to make the code tolerant of instructions with more than two input operands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163000 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-31 16:30:05 +00:00
Jakob Stoklund Olesen
908c0c01f6 Don't enforce ordered inline asm operands.
I was too optimistic, inline asm can have tied operands that don't
follow the def order.

Fixes PR13742.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162998 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-31 15:34:59 +00:00
Pete Cooper
5dd9e214fb Take account of boolean vector contents when promoting a build vector from i1 to some other type. rdar://problem/12210060
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162960 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-30 23:58:52 +00:00
Owen Anderson
43da6c7f13 Teach the DAG combiner to turn chains of FADDs (x+x+x+x+...) into FMULs by constants. This is only enabled in unsafe FP math mode, since it does not preserve rounding effects for all such constants.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162956 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-30 23:35:16 +00:00
Nadav Rotem
e757f00446 Currently targets that do not support selects with scalar conditions and vector operands - scalarize the code. ARM is such a target
because it does not support CMOV of vectors. To implement this efficientlyi, we broadcast the condition bit and use a sequence of NAND-OR
to select between the two operands. This is the same sequence we use for targets that don't have vector BLENDs (like SSE2).

rdar://12201387



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162926 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-30 19:17:29 +00:00
Jakob Stoklund Olesen
e941df5e35 Don't use MCInstrDesc flags for implicit operands.
When a MachineInstr is constructed, its implicit operands are added
first, then the explicit operands are inserted before the implicits.

MCInstrDesc has oprand flags like early clobber and operand ties that
apply to the explicit operands.

Don't look at those flags when the implicit operands are first added in
the explicit operands's positions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162910 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-30 14:39:06 +00:00
Craig Topper
6b1e1d8b3d Add FMA to switch statement in VectorLegalizer::LegalizeOp so that it can be expanded when it isn't legal.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162894 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-30 07:34:22 +00:00
Craig Topper
3b9dfc9bf7 Add support for FMA to WidenVectorResult.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162893 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-30 07:13:41 +00:00
Jakob Stoklund Olesen
492d76b02e Verify the order of tied operands in inline asm.
When there are multiple tied use-def pairs on an inline asm instruction,
the tied uses must appear in the same order as the defs.

It is possible to write an LLVM IR inline asm instruction that breaks
this constraint, but there is no reason for a front end to emit the
operands out of order.

The gnu inline asm syntax specifies tied operands as a single read/write
constraint "+r", so ouf of order operands are not possible.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162878 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-29 23:52:52 +00:00
Jakob Stoklund Olesen
66390805ad Set the isTied flags when building INLINEASM MachineInstrs.
For normal instructions, isTied() is set automatically by addOperand(),
based on MCInstrDesc, but inline asm has tied operands outside the
descriptor.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162869 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-29 22:02:00 +00:00
Jakob Stoklund Olesen
f036f7a1e7 Rename hasVolatileMemoryRef() to hasOrderedMemoryRef().
Ordered memory operations are more constrained than volatile loads and
stores because they must be ordered with respect to all other memory
operations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162861 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-29 21:19:21 +00:00
Jakob Stoklund Olesen
0d75858395 Don't move normal loads across volatile/atomic loads.
It is technically allowed to move a normal load across a volatile load,
but probably not a good idea.

It is not allowed to move a load across an atomic load with
Ordering > Monotonic, and we model those with MOVolatile as well.

I recently removed the mayStore flag from atomic load instructions, so
they don't need a pseudo-opcode. This patch makes up for the difference.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162857 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-29 20:48:45 +00:00
Jakob Stoklund Olesen
90a4f78d3b Verify the consistency of inline asm operands.
The operands on an INLINEASM machine instruction are divided into groups
headed by immediate flag operands. Verify this structure.

Extract verifyTiedOperands(), and only call it for non-inlineasm
instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162849 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-29 18:11:05 +00:00
Eric Christopher
17dd9a992a Clean this up slightly, doesn't really fall through.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162848 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-29 17:59:32 +00:00
Jakob Stoklund Olesen
ca71c5debe Verify the tied operand flags.
WHen running with -verify-machineinstrs, check that tied operands come
in matching use/def pairs, and that they are consistent with MCInstrDesc
when it applies.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162816 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-29 00:38:03 +00:00
Jakob Stoklund Olesen
699ac04951 Maintain a vaild isTied bit as operands are added and removed.
The isTied bit is set automatically when a tied use is added and
MCInstrDesc indicates a tied operand. The tie is broken when one of the
tied operands is removed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162814 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-29 00:37:58 +00:00
Jakob Stoklund Olesen
4ba6916a98 Add a MachineOperand::isTied() flag.
While in SSA form, a MachineInstr can have pairs of tied defs and uses.
The tied operands are used to represent read-modify-write operands that
must be assigned the same physical register.

Previously, tied operand pairs were computed from fixed MCInstrDesc
fields, or by using black magic on inline assembly instructions.

The isTied flag makes it possible to add tied operands to any
instruction while getting rid of (some of) the inlineasm magic.

Tied operands on normal instructions are needed to represent predicated
individual instructions in SSA form. An extra <tied,imp-use> operand is
required to represent the output value when the instruction predicate is
false.

Adding a predicate to:

  %vreg0<def> = ADD %vreg1, %vreg2

Will look like:

  %vreg0<tied,def> = ADD %vreg1, %vreg2, pred:3, %vreg7<tied,imp-use>

The virtual register %vreg7 is the value given to %vreg0 when the
predicate is false. It will be assigned the same physreg as %vreg0.

This commit adds the isTied flag and sets it based on MCInstrDesc when
building an instruction. The flag is not used for anything yet.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162774 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-28 18:34:41 +00:00
Jakob Stoklund Olesen
190e342cbc Don't allow TargetFlags on MO_Register MachineOperands.
Register operands are manipulated by a lot of target-independent code,
and it is not always possible to preserve target flags. That means it is
not safe to use target flags on register operands.

None of the targets in the tree are using register operand target flags.
External targets should be using immediate operands to annotate
instructions with operand modifiers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162770 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-28 18:05:48 +00:00
Jakob Stoklund Olesen
36d29bc723 Remove extra MayLoad/MayStore flags from atomic_load/store.
These extra flags are not required to properly order the atomic
load/store instructions. SelectionDAGBuilder chains atomics as if they
were volatile, and SelectionDAG::getAtomic() sets the isVolatile bit on
the memory operands of all atomic operations.

The volatile bit is enough to order atomic loads and stores during and
after SelectionDAG.

This means we set mayLoad on atomic_load, mayStore on atomic_store, and
mayLoad+mayStore on the remaining atomic read-modify-write operations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162733 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-28 03:11:32 +00:00
Akira Hatanaka
1d522388bf Fix bug 13532.
In SelectionDAGLegalize::ExpandLegalINT_TO_FP, expand INT_TO_FP nodes without
using any f64 operations if f64 is not a legal type.

Patch by Stefan Kristiansson. 


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162728 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-28 02:12:42 +00:00
Richard Smith
1144af3c9b Fix integer undefined behavior due to signed left shift overflow in LLVM.
Reviewed offline by chandlerc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162623 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-24 23:29:28 +00:00
Jakob Stoklund Olesen
baa74e4b35 Avoid including explicit uses when counting SDNode imp-uses.
It is legal to have a register node as an explicit operand, it shouldn't
be counted as an implicit use.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162591 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-24 20:52:42 +00:00
Manman Ren
1a710fdde1 BranchProb: modify the definition of an edge in BranchProbabilityInfo to handle
the case of multiple edges from one block to another.

A simple example is a switch statement with multiple values to the same
destination. The definition of an edge is modified from a pair of blocks to
a pair of PredBlock and an index into the successors.

Also set the weight correctly when building SelectionDAG from LLVM IR,
especially when converting a Switch.
IntegersSubsetMapping is updated to calculate the weight for each cluster.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162572 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-24 18:14:27 +00:00
Eric Christopher
873cf0a0d7 Use DW_FORM_flag_present to save space in debug information if we're
not in darwin gdb compat mode.

Fixes rdar://10975088

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162526 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-24 01:14:27 +00:00
Eric Christopher
cbbd5b1d86 Remove the DW_AT_MIPS_linkage name attribute when we don't need it
output (we're emitting a specification already and the information
isn't changing) and we're not in old gdb compat mode.

Saves 1% on the debug information for a build of llvm.

Fixes rdar://11043421

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162493 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-23 22:52:55 +00:00
Eric Christopher
20f47ab768 Turn these two options in to trinary state so that they can be
turned on and off separate from the platform if you're on darwin.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162487 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-23 22:36:40 +00:00
Eric Christopher
c1610fa3c4 Add a flag to DwarfDebug to allow it to communicate whether or not
we're using the darwin old gdb compat mode for emitting dwarf.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162486 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-23 22:36:36 +00:00
Eric Christopher
9d9f5a5855 Typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162438 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-23 07:32:06 +00:00
Eric Christopher
9eb1a94c20 Only emit the __debug_inlined section if we're trying to be compatible
with older gdbs on darwin.

rdar://10975874

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162436 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-23 07:32:02 +00:00
Eric Christopher
360f0062bc Emit pubtypes only when going for darwin gdb compatibility.
rdar://10393214

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162434 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-23 07:10:56 +00:00
Eric Christopher
10cb744f19 Add an option for darwin gdb compatibility.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162432 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-23 07:10:46 +00:00
Andrew Trick
ffd2526fa4 Simplify the computeOperandLatency API.
The logic for recomputing latency based on a ScheduleDAG edge was
shady. This bypasses the problem by requiring the client to provide
operand indices. This ensures consistent use of the machine model's
API.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162420 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-23 00:39:43 +00:00
David Blaikie
986d76d7b3 Tidy up a few more uses of MF.getFunction()->getName().
Based on CR feedback from r162301 and Craig Topper's refactoring in r162347
here are a few other places that could use the same API (& in one instance drop
a Function.h dependency).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162367 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-22 17:18:53 +00:00
Benjamin Kramer
05d96f98cb Reduce duplicated hash map lookups.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162362 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-22 15:37:57 +00:00
Stepan Dyatkovskiy
fdeb9fe5e0 Rejected 169195. As Duncan commented, bitcasting to proper type is wrong approach. We need to insert some valid TRANCATE node here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162354 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-22 09:33:55 +00:00
Craig Topper
96601ca332 Add a getName function to MachineFunction. Use it in places that previously did getFunction()->getName(). Remove includes of Function.h that are no longer needed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162347 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-22 06:07:19 +00:00
Richard Smith
cb1f68d7c8 Initialize SelectionDAGBuilder's Context in 'init', not in its constructor. The
SelectionDAG's 'init' has not been called when the SelectionDAGBuilder is
constructed (in SelectionDAGISel's constructor), so this was previously always
initialized with 0.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162333 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-22 00:42:39 +00:00
David Blaikie
79c78de378 Remove unnecessary cast that was also unnecessarily casting away constness.
Even looking at the revision history I couldn't quite piece together why this
cast was ever written in the first place, but I assume it was because of some
change in the inheritance, perhaps this function was reimplemented in a
derived type & this caller was meant to get the base version (& it wasn't
virtual)?

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162301 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-21 18:54:23 +00:00
Chad Rosier
35907e9862 Add support for the --param ssp-buffer-size= driver option.
PR9673

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162284 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-21 16:15:24 +00:00
Jakob Stoklund Olesen
990ca5517f Fix a quadratic algorithm in MachineBranchProbabilityInfo.
The getSumForBlock function was quadratic in the number of successors
because getSuccWeight would perform a linear search for an already known
iterator.

This patch was originally committed as r161460, but reverted again
because of assertion failures. Now that duplicate Machine CFG edges have
been eliminated, this works properly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162233 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-20 22:01:38 +00:00
Jakob Stoklund Olesen
e7fdef420d Don't add CFG edges for redundant conditional branches.
IR that hasn't been through SimplifyCFG can look like this:

  br i1 %b, label %r, label %r

Make sure we don't create duplicate Machine CFG edges in this case.

Fix the machine code verifier to accept conditional branches with a
single CFG edge.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162230 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-20 21:39:52 +00:00
Jakob Stoklund Olesen
228e3f5fac Add a verification pass after ExpandISelPseudos.
This pass often has weird CFG hacks and hand-written MI building code
that can go wrong in many ways.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162224 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-20 20:52:08 +00:00
Jakob Stoklund Olesen
b254c6d77a Add CFG checks to MachineVerifier.
Verify that the predecessor and successor lists are consistent and free
of duplicates.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162223 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-20 20:52:06 +00:00
Stepan Dyatkovskiy
08da177c35 Fixed DAGCombiner bug (found and localized by James Malloy):
The DAGCombiner tries to optimise a BUILD_VECTOR by checking if it
consists purely of get_vector_elts from one or two source vectors. If
so, it either makes a concat_vectors node or a shufflevector node.

However, it doesn't check the element type width of the underlying
vector, so if you have this sequence:

Node0: v4i16 = ...
Node1: i32 = extract_vector_elt Node0
Node2: i32 = extract_vector_elt Node0
Node3: v16i8 = BUILD_VECTOR Node1, Node2, ...

It will attempt to:

Node0:    v4i16 = ...
NewNode1: v16i8 = concat_vectors Node0, ...

Where this is actually invalid because the element width is completely
different. This causes an assertion failure on DAG legalization stage.

Fix:
If output item type of BUILD_VECTOR differs from input item type.
Make concat_vectors based on input element type and then bitcast it to the output vector type. So the case described above will transformed to:
Node0:    v4i16 = ...
NewNode1: v8i16 = concat_vectors Node0, ...
NewNode2: v16i8 = bitcast NewNode1



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162195 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-20 07:57:06 +00:00
Eli Friedman
fd45fa1503 Make atomic load and store of pointers work. Tighten verification of atomic operations
so other unexpected operations don't slip through.  Based on patch by Logan Chien.
PR11786/PR13186.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162146 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-17 23:24:29 +00:00
Bill Wendling
a67eda76c0 Implement stack protectors for structures with character arrays in them.
<rdar://problem/10545247>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162131 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-17 20:59:56 +00:00
Bill Wendling
32811bef95 Change the linker_private_weak_def_auto' linkage to linkonce_odr_auto_hide' to
make it more consistent with its intended semantics.

The `linker_private_weak_def_auto' linkage type was meant to automatically hide
globals which never had their addresses taken. It has nothing to do with the
`linker_private' linkage type, which outputs the symbols with a `l' (ell) prefix
among other things.

The intended semantic is more like the `linkonce_odr' linkage type.

Change the name of the linkage type to `linkonce_odr_auto_hide'. And therefore
changing the semantics so that it produces the correct output for the linker.

Note: The old linkage name `linker_private_weak_def_auto' will still parse but
is not a synonym for `linkonce_odr_auto_hide'. This should be removed in 4.0.
<rdar://problem/11754934>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162114 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-17 18:33:14 +00:00
Benjamin Kramer
b97cebdfcc TargetLowering: Use the large shift amount during legalize types. The legalizer may call us with an overly large type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162101 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-17 15:54:21 +00:00
Jakob Stoklund Olesen
cabc0699ea Use standard pattern for iterate+erase.
Increment the MBB iterator at the top of the loop to properly handle the
current (and previous) instructions getting erased.

This fixes PR13625.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162099 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-17 14:38:59 +00:00
Jakob Stoklund Olesen
f2c64ef519 Add an MCID::Select flag and TII hooks for optimizing selects.
Select instructions pick one of two virtual registers based on a
condition, like x86 cmov. On targets like ARM that support predication,
selects can sometimes be eliminated by predicating the instruction
defining one of the operands.

Teach PeepholeOptimizer to recognize select instructions, and ask the
target to optimize them.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162059 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-16 23:11:47 +00:00
Richard Smith
11a4fa4523 Fix undefined behavior: don't perform array indexing through a potentially null
pointer.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161919 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-15 01:39:31 +00:00
Richard Smith
bb57febe3c Fix undefined behavior: binding null pointer to reference. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161853 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-14 05:31:26 +00:00
Eric Christopher
443c9ed768 Grammar.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161851 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-14 05:13:29 +00:00
Owen Anderson
7c626d3097 Add a roundToIntegral method to APFloat, which can be parameterized over various rounding modes. Use this to implement SelectionDAG constant folding of FFLOOR, FCEIL, and FTRUNC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161807 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-13 23:32:49 +00:00
Jakob Stoklund Olesen
03e593efc6 Transfer weights in transferSuccessorsAndUpdatePHIs().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161805 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-13 23:13:25 +00:00
Jakob Stoklund Olesen
cb6889b23d Print out MachineBasicBlock successor weights when available.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161804 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-13 23:13:23 +00:00
Jakob Stoklund Olesen
43600e95ec Remove the TII::scheduleTwoAddrSource() hook.
It never does anything when running 'make check', and it get's in the
way of updating live intervals in 2-addr.

The hook was originally added to help form IT blocks in Thumb2 code
before register allocation, but the pass ordering has changed since
then, and we run if-conversion after register allocation now.

When the MI scheduler is enabled, there will be no less than two
schedulers between 2-addr and Thumb2ITBlockPass, so this hook is
unlikely to help anything.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161794 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-13 21:52:57 +00:00
Bill Wendling
6d86f3cdfc Whitespace cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161788 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-13 21:20:43 +00:00
Jakob Stoklund Olesen
786556c268 Count triangles and diamonds in early if-conversion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161783 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-13 21:03:27 +00:00
Jakob Stoklund Olesen
5f6d36335b Delete dead typedef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161782 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-13 21:03:25 +00:00
Jakob Stoklund Olesen
bc70ff3cb9 Handle extra Tail predecessors in if-conversion.
It is still possible to if-convert if the tail block has extra
predecessors, but the tail phis must be rewritten instead of being
removed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161781 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-13 20:49:04 +00:00
Benjamin Kramer
5fa2d458af MachineCSE: Hoist isConstantPhysReg out of the loop, it checks for overlaps already.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161729 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-11 20:42:59 +00:00
Benjamin Kramer
cfc0ad6e48 PR13578: Teach MachineCSE that instructions that use a constant register can be CSE'd safely.
This is common e.g. when doing rip-relative addressing on x86_64.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161728 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-11 19:05:13 +00:00
Jakob Stoklund Olesen
eb74c08192 Add a proper if-conversion cost model.
Detect when there is not enough available ILP, so if-conversion can't
speculate instructions for free.

Compute the lengthening of the critical path when inserting a select
instruction that depends on the condition as well as both sides of the
if.

Reject conversions that would stretch the critical path by more than
half a mispredict penalty.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161713 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-10 22:27:31 +00:00
Jakob Stoklund Olesen
a35bf506f4 Give MachineTraceMetrics its own debug tag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161712 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-10 22:27:29 +00:00
Jakob Stoklund Olesen
5413b68b1f Add more trace query functions.
Trace::getResourceLength() computes the number of cycles required to
execute the trace when ignoring data dependencies. The number can be
compared to the critical path to estimate the trace ILP.

Trace::getPHIDepth() computes the data dependency depth of a PHI in a
trace successor that isn't necessarily part of the trace.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161711 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-10 22:27:27 +00:00
Jakob Stoklund Olesen
870da6de2c Add getTPred() and getFPred() functions.
They identify the PHI predecessors in both diamonds and triangles.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161689 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-10 20:19:17 +00:00
Jakob Stoklund Olesen
8828c4ccd4 Include loop-carried dependencies when computing instr heights.
When a trace ends with a back-edge, include PHIs in the loop header in
the height computations. This makes the critical path through a loop
more accurate by including the latencies of the last instructions in the
loop.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161688 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-10 20:11:38 +00:00
Jakob Stoklund Olesen
15121ca0d1 Update edge weights correctly in replaceSuccessor().
When replacing Old with New, it can happen that New is already a
successor. Add the old and new edge weights instead of creating a
duplicate edge.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161653 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-10 03:23:27 +00:00
Jakob Stoklund Olesen
c7908037d8 Reapply r161633-161634 "Partition use lists so defs always come before uses.""
No changes to these patches, MRI needed to be notified when changing
uses into defs and vice versa.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161644 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-10 00:21:30 +00:00
Jakob Stoklund Olesen
d6397eba23 Also update MRI use lists when changing a use to a def and vice versa.
This was the cause of the buildbot failures.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161643 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-10 00:21:26 +00:00
Jakob Stoklund Olesen
1134aae4e7 Revert r161633-161634 "Partition use lists so defs always come before uses."
These commits broke a number of buildbots.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161640 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-09 23:31:36 +00:00
Jakob Stoklund Olesen
81a6995243 Partition use lists so defs always come before uses.
This makes it possible to speed up def_iterator by stopping at the first
use. This makes def_empty() and getUniqueVRegDef() much faster when
there are many uses.

In a +Asserts build, LiveVariables is 100x faster in one case because
getVRegDef() has an assertion that would scan to the end of a
def_iterator chain.

Spill weight calculation is significantly faster (300x in one case)
because isTriviallyReMaterializable() calls MRI->isConstantPhysReg(%RIP)
which calls def_empty(%RIP).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161634 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-09 22:49:46 +00:00
Jakob Stoklund Olesen
46f4c35372 Don't use pointer-pointers for the register use lists.
Use a more conventional doubly linked list where the Prev pointers form
a cycle. This means it is no longer necessary to adjust the Prev
pointers when reallocating the VRegInfo array.

The test changes are required because the register allocation hint is
using the use-list order to break ties.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161633 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-09 22:49:42 +00:00
Jakob Stoklund Olesen
ff2b99afc8 Move use list management into MachineRegisterInfo.
Register MachineOperands are kept in linked lists accessible via MRI's
reg_iterator interfaces. The linked list management was handled partly
by MachineOperand methods, partly by MRI methods.

Move all of the list management into MRI, delete
MO::AddRegOperandToRegInfo() and MO::RemoveRegOperandFromRegInfo().

Be more explicit about handling the cases where an MRI pointer isn't
available.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161632 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-09 22:49:37 +00:00
Jakob Stoklund Olesen
988069e229 Fix a future TwoAddressInstructionPass crash.
No test case, the crash only happens when the default use list order is
changed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161627 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-09 22:08:26 +00:00
Nadav Rotem
0b66bd9b07 Fix the legalization of ExtLoad on ARM. ExpandUnalignedLoad did not properly
handle the cases where the memory value type was illegal. 
PR 13111. 



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161565 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-09 01:56:44 +00:00
Jakob Stoklund Olesen
4e6966266a Don't use getNextOperandForReg() in RAFast.
That particular optimization was probably premature anyway.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161541 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-08 23:44:01 +00:00
Jakob Stoklund Olesen
e723007ee6 Deal with irreducible control flow when building traces.
We filter out MachineLoop back-edges during the trace-building PO
traversals, but it is possible to have CFG cycles that aren't natural
loops, and MachineLoopInfo doesn't include such cycles.

Use a standard visited set to detect such CFG cycles, and completely
ignore them when picking traces.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161532 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-08 22:12:01 +00:00
Jakob Stoklund Olesen
d6cf5f4224 Heed -stress-early-ifcvt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161513 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-08 18:24:23 +00:00
Jakob Stoklund Olesen
0fac6aa076 Get the MispredictPenalty from MCSchedModel.
Thanks, Andy!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161507 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-08 18:19:58 +00:00
Andrew Trick
3c417554ca Minor cleanup of defaultDefLatency API
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161470 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-08 02:44:11 +00:00
Jakob Stoklund Olesen
913ff09a9a Revert "Fix a quadratic algorithm in MachineBranchProbabilityInfo."
It caused an assertion failure when compiling consumer-typeset.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161463 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-08 01:10:31 +00:00
Manman Ren
39ad568c62 X86: enable CSE between CMP and SUB
We perform the following:
1> Use SUB instead of CMP for i8,i16,i32 and i64 in ISel lowering.
2> Modify MachineCSE to correctly handle implicit defs.
3> Convert SUB back to CMP if possible at peephole.

Removed pattern matching of (a>b) ? (a-b):0 and like, since they are handled
by peephole now.

rdar://11873276


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161462 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-08 00:51:41 +00:00
Jakob Stoklund Olesen
5e63d43e48 Fix a quadratic algorithm in MachineBranchProbabilityInfo.
The getSumForBlock function was quadratic in the number of successors
because getSuccWeight would perform a linear search for an already known
iterator.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161460 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-08 00:20:37 +00:00
Jakob Stoklund Olesen
8c5c0733cd Skip tied operand pairs that already have the same register.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161454 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-07 22:47:06 +00:00
Jakob Stoklund Olesen
74500bdba3 Add SelectionDAG::getTargetIndex.
This adds support for TargetIndex operands during isel. The meaning of
these (index, offset, flags) operands is entirely defined by the target.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161453 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-07 22:37:05 +00:00
Bill Wendling
1299422ee1 For non-Darwin platforms, we want to generate stack protectors only for
character arrays. This is in line with what GCC does.
<rdar://problem/10529227>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161446 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-07 20:59:05 +00:00
Jakob Stoklund Olesen
0b40d09ff6 Add a new kind of MachineOperand: MO_TargetIndex.
A target index operand looks a lot like a constant pool reference, but
it is completely target-defined. It contains the 8-bit TargetFlags, a
32-bit index, and a 64-bit offset. It is preserved by all code generator
passes.

TargetIndex operands can be used to carry target-specific information in
cases where immediate operands won't suffice.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161441 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-07 18:56:39 +00:00
Jakob Stoklund Olesen
7a8f311ece Fix a couple of typos.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161437 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-07 18:32:57 +00:00
Jakob Stoklund Olesen
84ef6ba443 Add trace accessor methods, implement primitive if-conversion heuristic.
Compare the critical paths of the two traces through an if-conversion
candidate. If the difference is larger than the branch brediction
penalty, reject the if-conversion. If would never pay.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161433 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-07 18:02:19 +00:00
Chandler Carruth
e6450dc2af Add a much more conservative strategy for aligning branch targets.
Previously, MBP essentially aligned every branch target it could. This
bloats code quite a bit, especially non-looping code which has no real
reason to prefer aligned branch targets so heavily.

As Andy said in review, it's still a bit odd to do this without a real
cost model, but this at least has much more plausible heuristics.

Fixes PR13265.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161409 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-07 09:45:24 +00:00
Manman Ren
ba86b13ad9 MachineCSE: Update the heuristics for isProfitableToCSE.
If the result of a common subexpression is used at all uses of the candidate
expression, CSE should not increase the live range of the common subexpression.

rdar://11393714 and rdar://11819721


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161396 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-07 06:16:46 +00:00
Jakob Stoklund Olesen
cf1823dc25 Delete a dead variable.
TwoAddressInstructionPass doesn't remat any more.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161285 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-04 00:04:03 +00:00
Jakob Stoklund Olesen
ae52faddeb TwoAddressInstructionPass refactoring: Extract another method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161284 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 23:57:58 +00:00
Bob Wilson
53624a2df5 Refactor and check "onlyReadsMemory" before optimizing builtins.
This patch is mostly just refactoring a bunch of copy-and-pasted code, but
it also adds a check that the call instructions are readnone or readonly.
That check was already present for sin, cos, sqrt, log2, and exp2 calls, but
it was missing for the rest of the builtins being handled in this code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161282 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 23:29:17 +00:00
Jakob Stoklund Olesen
6ac8066ae4 TwoAddressInstructionPass refactoring: Extract a method.
No functional change intended, except replacing a DenseMap with a
SmallDenseMap which should behave identically.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161281 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 23:25:45 +00:00
Jakob Stoklund Olesen
5bfdedf6ab Begin adding support for updating LiveIntervals in TwoAddressInstructionPass.
This is far from complete, and only changes behavior when the
-early-live-intervals flag is passed to llc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161273 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 22:58:34 +00:00
Jakob Stoklund Olesen
dcc4436cdd Add an experimental -early-live-intervals option.
This option runs LiveIntervals before TwoAddressInstructionPass which
will eventually learn to exploit and update the analysis.

Eventually, LiveIntervals will run before PHIElimination, and we can get
rid of LiveVariables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161270 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 22:12:54 +00:00
Jakob Stoklund Olesen
e744ac49f4 Delete merged physreg copies in joinReservedPhysReg().
Previously, the identity copy would survive through register allocation
before it was removed by the rewriter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161269 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 22:12:51 +00:00
Bob Wilson
982dc84762 Try to reduce the compile time impact of r161232.
The previous change caused fast isel to not attempt handling any calls to
builtin functions.  That included things like "printf" and caused some
noticable regressions in compile time.  I wanted to avoid having fast isel
keep a separate list of functions that had to be kept in sync with what the
code in SelectionDAGBuilder.cpp was handling.  I've resolved that here by
moving the list into TargetLibraryInfo.  This is somewhat redundant in
SelectionDAGBuilder but it will ensure that we keep things consistent.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161263 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 21:26:24 +00:00
Bob Wilson
772af92cb1 Fix memcmp code-gen to honor -fno-builtin.
I noticed that SelectionDAGBuilder::visitCall was missing a check for memcmp
in TargetLibraryInfo, so that it would use custom code for memcmp calls even
with -fno-builtin.  I also had to add a new -disable-simplify-libcalls option
to llc so that I could write a test for this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161262 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 21:26:18 +00:00
Jakob Stoklund Olesen
b2beac2b96 Completely eliminate VNInfo flags.
The 'unused' state of a value number can be represented as an invalid
def SlotIndex. This also exposed code that shouldn't have been looking
at unused value VNInfos.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161258 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 20:59:32 +00:00
Jakob Stoklund Olesen
1599a64277 Fix a couple of loops that were processing unused value numbers.
Unused VNInfos should be left alone. Their def SlotIndex doesn't point
to anything.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161257 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 20:59:29 +00:00
Matt Beaumont-Gay
a5b51f5694 Silence unused variable warning in -asserts build
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161256 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 20:54:11 +00:00
Jakob Stoklund Olesen
bf60aa9db5 Eliminate the VNInfo::hasPHIKill() flag.
The only real user of the flag was removeCopyByCommutingDef(), and it
has been switched to LiveIntervals::hasPHIKill().

All the code changed by this patch was only concerned with computing and
propagating the flag.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161255 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 20:19:44 +00:00
Jakob Stoklund Olesen
0ab7103e06 Make the hasPHIKills flag a computed property.
The VNInfo::HAS_PHI_KILL is only half supported. We precompute it in
LiveIntervalAnalysis, but it isn't properly updated by live range
splitting and functions like shrinkToUses().

It is only used in one place: RegisterCoalescer::removeCopyByCommutingDef().

This patch changes that function to use a new LiveIntervals::hasPHIKill()
function that computes the flag for a given value number.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161254 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 20:10:24 +00:00
Jakob Stoklund Olesen
3cd9f572ed Delete dead function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161242 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 15:21:21 +00:00
Jakob Stoklund Olesen
bc8d827b2b Don't delete dead code in TwoAddressInstructionPass.
This functionality was added before we started running
DeadMachineInstructionElim on all targets. It serves no purpose now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161241 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 15:11:57 +00:00
Bob Wilson
d49edb7ab0 Fall back to selection DAG isel for calls to builtin functions.
Fast isel doesn't currently have support for translating builtin function
calls to target instructions.  For embedded environments where the library
functions are not available, this is a matter of correctness and not
just optimization.  Most of this patch is just arranging to make the
TargetLibraryInfo available in fast isel.  <rdar://problem/12008746>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161232 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 04:06:28 +00:00
Manman Ren
127eea87d6 X86 Peephole: fold loads to the source register operand if possible.
Add more comments and use early returns to reduce nesting in isLoadFoldable.
Also disable folding for V_SET0 to avoid introducing a const pool entry and
a const pool load.

rdar://10554090 and rdar://11873276


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161207 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-02 19:37:32 +00:00
Jakob Stoklund Olesen
79a20ce6f0 Compute the critical path length through a trace.
Whenever both instruction depths and instruction heights are known in a
block, it is possible to compute the length of the critical path as
max(depth+height) over the instructions in the block.

The stored live-in lists make it possible to accurately compute the
length of a critical path that bypasses the current (small) block.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161197 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-02 18:45:54 +00:00
Jakob Stoklund Olesen
804468916a Verify regunit intervals along with virtreg intervals.
Don't cause regunit intervals to be computed just to verify them. Only
check the already cached intervals.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161183 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-02 16:36:50 +00:00
Jakob Stoklund Olesen
e2b83ab056 Avoid creating dangling physreg live ranges during DCE.
LiveRangeEdit::eliminateDeadDefs() can delete a dead instruction that
reads unreserved physregs. This would leave the corresponding regunit
live interval dangling because we don't have shrinkToUses() for physical
registers.

Fix this problem by turning the instruction into a KILL instead of
deleting it. This happens in a landing pad in
test/CodeGen/X86/2012-05-19-CoalescerCrash.ll:

  %vreg27<def,dead> = COPY %EDX<kill>; GR32:%vreg27

becomes:

  KILL %EDX<kill>

An upcoming fix to the machine verifier will catch problems like this by
verifying regunit live intervals.

This fixes PR13498. I am not including the test case from the PR since
we already have one exposing the problem once the verifier is fixed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161182 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-02 16:36:47 +00:00
Jakob Stoklund Olesen
79240f9539 Add report() functions that take a LiveInterval argument.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161178 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-02 14:31:49 +00:00
Manman Ren
d7d003c2b7 X86 Peephole: fold loads to the source register operand if possible.
Machine CSE and other optimizations can remove instructions so folding
is possible at peephole while not possible at ISel.

This patch is a rework of r160919 and was tested on clang self-host on my local
machine.

rdar://10554090 and rdar://11873276


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161152 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-02 00:56:42 +00:00
Jakob Stoklund Olesen
e5c79a5c25 Extract some methods from verifyLiveIntervals.
No functional change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161149 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-02 00:20:20 +00:00
Jakob Stoklund Olesen
a62e1e8bc6 Also verify RegUnit intervals at uses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161147 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-01 23:52:40 +00:00
Jakob Stoklund Olesen
c7f44b8b8f Compute instruction heights through a trace.
The height on an instruction is the minimum number of cycles from the
instruction is issued to the end of the trace. Heights are computed for
all instructions in and below the trace center block.

The method for computing heights is different from the depth
computation. As we visit instructions in the trace bottom-up, heights of
used instructions are pushed upwards. This way, we avoid scanning long
use lists, looking for uses in the current trace.

At each basic block boundary, a list of live-in registers and their
minimum heights is saved in the trace block info. These live-in lists
are used when restarting depth computations on a trace that
converges with an already computed trace. They will also be used to
accurately compute the critical path length.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161138 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-01 22:36:00 +00:00
Eric Christopher
6635cad548 Temporarily revert c23b933d5f. It's causing
failures in the debug testsuite and possibly PR13486.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161121 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-01 18:19:01 +00:00
Jakob Stoklund Olesen
9dae457e2b Add DataDep constructors. Explicitly check SSA form.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161115 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-01 16:02:59 +00:00
Elena Demikhovsky
1503aba4a0 Added FMA functionality to X86 target.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161110 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-01 12:06:00 +00:00
Manman Ren
f99efdf329 MachineSink: Sort the successors before trying to find SuccToSinkTo.
Use stable_sort instead of sort. Follow-up to r161062.

rdar://11980766


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161075 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-31 20:45:38 +00:00
Jakob Stoklund Olesen
5f8e8bd656 Compute instruction depths through the current trace.
Assuming infinite issue width, compute the earliest each instruction in
the trace can issue, when considering the latency of data dependencies.
The issue cycle is record as a 'depth' from the beginning of the trace.

This is half the computation required to find the length of the critical
path through the trace. Heights are next.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161074 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-31 20:44:38 +00:00
Jakob Stoklund Olesen
64e2973bf7 Rename CT -> MTM. MachineTraceMetrics is abbreviated MTM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161072 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-31 20:25:13 +00:00
Manman Ren
53b59d1d97 MachineSink: Sort the successors before trying to find SuccToSinkTo.
One motivating example is to sink an instruction from a basic block which has
two successors: one outside the loop, the other inside the loop. We should try
to sink the instruction outside the loop.

rdar://11980766


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161062 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-31 18:10:39 +00:00
Micah Villmow
8c574be2fe Conform to LLVM coding style.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161061 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-31 18:07:43 +00:00
Micah Villmow
683826765c Don't generate ordered or unordered comparison operations if it is not legal to do so.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161053 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-31 16:48:03 +00:00
Jakob Stoklund Olesen
34af6f597b Clear kill flags in removeCopyByCommutingDef().
We are extending live ranges, so kill flags are not accurate. They
aren't needed until they are recomputed after RA anyway.

<rdar://problem/11950722>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161023 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-31 02:47:24 +00:00
Manman Ren
1123614317 Reverse order of the two branches at end of a basic block if it is profitable.
We branch to the successor with higher edge weight first.
Convert from
     je    LBB4_8  --> to outer loop
     jmp   LBB4_14 --> to inner loop
to
     jne   LBB4_14
     jmp   LBB4_8

PR12750
rdar: 11393714


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161018 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-31 01:11:07 +00:00
Andrew Trick
8b5704f958 Use the latest MachineRegisterInfo APIs. No functionality.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161010 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 23:48:17 +00:00
Andrew Trick
269120cd9b Inline MachineRegisterInfo::hasOneUse
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161007 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 23:48:12 +00:00
Jakob Stoklund Olesen
2f6b62b09f Avoid looking at stale data in verifyAnalysis().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161004 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 23:15:12 +00:00
Jakob Stoklund Olesen
1c899cf47c Allow traces to enter nested loops.
This lets traces include the final iteration of a nested loop above the
center block, and the first iteration of a nested loop below the center
block.

We still don't allow traces to contain backedges, and traces are
truncated where they would leave a loop, as seen from the center block.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161003 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 23:15:10 +00:00
Jakob Stoklund Olesen
20f13c50d8 Clarify invalidation strategy in comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160997 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 21:16:22 +00:00
Jakob Stoklund Olesen
8e54ab5c76 Assert that all trace candidate blocks have been visited by the PO.
When computing a trace, all the candidates for pred/succ must have been
visited. Filter out back-edges first, though. The PO traversal ignores
them.

Thanks to Andy for spotting this in review.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160995 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 21:10:27 +00:00
Jakob Stoklund Olesen
ef6c76c984 Hook into PassManager's analysis verification.
By overriding Pass::verifyAnalysis(), the pass contents will be verified
by the pass manager.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160994 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 20:57:50 +00:00
Pete Cooper
32ecfb4158 Consider address spaces for hashing and CSEing DAG nodes. Otherwise two loads from different x86 segments but the same address would get CSEd
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160987 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 20:23:19 +00:00
Jakob Stoklund Olesen
3f63a58978 Add MachineInstr::isTransient().
This is a cleaned up version of the isFree() function in
MachineTraceMetrics.cpp.

Transient instructions are very unlikely to produce any code in the
final output. Either because they get eliminated by RegisterCoalescing,
or because they are pseudo-instructions like labels and debug values.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160977 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 18:34:14 +00:00
Jakob Stoklund Olesen
a1b2bf7979 Add MachineTraceMetrics::verify().
This function verifies the consistency of cached data in the
MachineTraceMetrics analysis.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160976 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 18:34:11 +00:00
Jakob Stoklund Olesen
ee31ae12e8 Verify that the CFG hasn't changed during invalidate().
The MachineTraceMetrics analysis must be invalidated before modifying
the CFG. This will catch some of the violations of that rule.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160969 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 17:36:49 +00:00
Jakob Stoklund Olesen
f192b507a3 Add MachineBasicBlock::isPredecessor().
A->isPredecessor(B) is the same as B->isSuccessor(A), but it can
tolerate a B that is null or dangling. This shouldn't happen normally,
but it it useful for verification code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160968 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-30 17:36:47 +00:00
Manman Ren
e8b4a4a9d1 Revert r160920 and r160919 due to dragonegg and clang selfhost failure
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160927 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-29 02:44:09 +00:00
Manman Ren
0eb3edea9c X86 Peephole: fold loads to the source register operand if possible.
Machine CSE and other optimizations can remove instructions so folding
is possible at peephole while not possible at ISel.

rdar://10554090 and rdar://11873276


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160919 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-28 16:48:01 +00:00
Andrew Trick
4b72ada1f4 Reenable a basic SSA DAG builder optimization.
Jakob fixed ProcessImplicifDefs in r159149.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160910 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-28 01:48:15 +00:00
Jakob Stoklund Olesen
08f6ef6a78 Add more debug output to MachineTraceMetrics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160905 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-27 23:58:38 +00:00
Jakob Stoklund Olesen
0271a5fa29 Keep track of the head and tail of the trace through each block.
This makes it possible to quickly detect blocks that are outside the
trace.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160904 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-27 23:58:36 +00:00
Eric Christopher
c23b933d5f Add a DW_AT_high_pc for CUs that are a single address range. Update
all tests accordingly.

Fixes PR13351.

Patch by shinichiro hamaji!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160899 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-27 22:00:05 +00:00
Jakob Stoklund Olesen
c16bf79303 Also compute register mask lists under -new-live-intervals.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160898 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-27 21:56:39 +00:00
Jakob Stoklund Olesen
b18d779b35 Eliminate the IS_PHI_DEF flag and VNInfo::setIsPHIDef().
A value number is a PHI def if and only if it begins at a block
boundary. This can be derived from the def slot, a separate flag is not
necessary.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160893 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-27 21:11:14 +00:00
Jakob Stoklund Olesen
3dfa38a5bf Add a -new-live-intervals experimental option.
This option replaces the existing live interval computation with one
based on LiveRangeCalc.cpp. The new algorithm does not depend on
LiveVariables, and it can be run at any time, before or after leaving
SSA form.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160892 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-27 20:58:46 +00:00
Jakob Stoklund Olesen
72e7dbf88b Add <imp-def> of super-register when lowering SUBREG_TO_REG.
Patch by Tyler Nowicki!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160888 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-27 20:19:49 +00:00
Jakob Stoklund Olesen
0fc44869ff Use an otherwise unused variable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160798 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-26 19:42:56 +00:00
Jakob Stoklund Olesen
9f63e10427 Start scaffolding for a MachineTraceMetrics analysis pass.
This is still a work in progress.

Out-of-order CPUs usually execute instructions from multiple basic
blocks simultaneously, so it is necessary to look at longer traces when
estimating the performance effects of code transformations.

The MachineTraceMetrics analysis will pick a typical trace through a
given basic block and provide performance metrics for the trace. Metrics
will include:

- Instruction count through the trace.
- Issue count per functional unit.
- Critical path length, and per-instruction 'slack'.

These metrics can be used to determine the performance limiting factor
when executing the trace, and how it will be affected by a code
transformation.

Initially, this will be used by the early if-conversion pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160796 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-26 18:38:11 +00:00
Dan Gohman
27db99fcee Add a floor intrinsic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160791 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-26 17:43:27 +00:00
Manman Ren
d68e8cda24 Disable rematerialization in TwoAddressInstructionPass.
It is redundant; RegisterCoalescer will do the remat if it can't eliminate
the copy. Collected instruction counts before and after this. A few extra
instructions are generated due to spilling but it is normal to see these kinds
of changes with almost any small codegen change, according to Jakob.

This also fixed rdar://11830760 where xor is expected instead of movi0.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160749 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-25 18:28:13 +00:00
Jakob Stoklund Olesen
84315f03cb Preserve 2-addr constraints in ConnectedVNInfoEqClasses.
When a live range splits into multiple connected components, we would
arbitrarily assign <undef> uses to component 0. This is wrong when the
use is tied to a def that gets assigned to a different component:

  %vreg69<def> = ADD8ri %vreg68<undef>, 1

The use and def must get the same virtual register.

Fix this by assigning <undef> uses to the same component as the value
defined by the instruction, if any:

  %vreg69<def> = ADD8ri %vreg69<undef>, 1

This fixes PR13402. The PR has a test case which I am not including
because it is unlikely to keep exposing this behavior in the future.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160739 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-25 17:15:15 +00:00
Jakob Stoklund Olesen
eba2bbb58c Verify two-address constraints more carefully.
Include <undef> operands and virtual registers after leaving SSA form.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160734 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-25 16:49:11 +00:00
Craig Topper
ea0ca84647 Change llvm_unreachable in SplitVectorOperand to report_fatal_error. Keeps release builds from crashing if code uses an intrinsic with an illegal type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160661 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-24 04:11:21 +00:00
Sylvestre Ledru
c8e41c5917 Fix a typo (the the => the)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160621 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-23 08:51:15 +00:00
Nadav Rotem
ed1a335ece Fixed DAGCombine optimizations which generate select_cc for targets
that do not support it (X86 does not lower select_cc).

PR: 13428

Together with Michael Kuperstein <michael.m.kuperstein@intel.com>



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160619 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-23 07:59:50 +00:00
Craig Topper
f093393b9a Tidy up. Fix indentation and remove trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160617 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-23 05:38:07 +00:00
Craig Topper
200e04c3fd Change llvm_unreachable in SplitVectorResult to report_fatal_error. Keeps release builds from crashing if code uses an intrinsic with an illegal type. For instance 256-bit AVX intrinsics without having AVX enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160616 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-23 04:34:49 +00:00
Benjamin Kramer
e288cd100f Remove unused private member variables uncovered by the recent changes to clang's -Wunused-private-field.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160583 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-20 22:05:57 +00:00
Jakob Stoklund Olesen
2ec0cda5d5 Avoid folding loads that are unsafe to move.
LiveRangeEdit::foldAsLoad() can eliminate a register by folding a load
into its only use. Only do that when the load is safe to move, and it
won't extend any live ranges.

This fixes PR13414.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160575 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-20 21:29:31 +00:00
Jakob Stoklund Olesen
c321a20b2e Split loop exiting edges more aggressively.
PHIElimination splits critical edges when it predicts it can resolve
interference and eliminate copies. It doesn't split the edge if the
interference wouldn't be resolved anyway because the phi-use register is
live in the critical edge anyway.

Teach PHIElimination to split loop exiting edges with interference, even
if it wouldn't resolve the interference. This removes the necessary
copies from the loop, which is still an improvement from injecting the
copies into the loop.

The test case demonstrates the improvement. Before:

LBB0_1:
  cmpb  $0, (%rdx)
  leaq  1(%rdx), %rdx
  movl  %esi, %eax
  je  LBB0_1

After:

LBB0_1:
  cmpb  $0, (%rdx)
  leaq  1(%rdx), %rdx
  je  LBB0_1

  movl  %esi, %eax

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160571 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-20 20:49:53 +00:00
Pete Cooper
b97c57a8b7 Fix crash in machine verifier when trying to print the def of a register which has no def
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160531 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-19 23:40:38 +00:00
Benjamin Kramer
b26e2916c9 Replace some explicit compare loops with std::equal.
No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160501 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-19 10:46:05 +00:00
Galina Kistanova
72ea0c9ffa Fixed few warnings.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160493 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-19 04:50:12 +00:00
Bill Wendling
96cb112852 Remove tabs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160475 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-19 00:04:14 +00:00
Chandler Carruth
32d75bec4b Fix a somewhat nasty crasher in PR13378. This crashes inside of
LiveIntervals due to the two-addr pass generating bogus MI code.

The crux of the issue was a loop nesting problem. The intent of the code
which attempts to transform instructions before converting them to
two-addr form is to defer and reprocess any transformed instructions as
the second processing is likely to have more opportunities to coalesce
copies, etc. Unfortunately, there was one section of processing that was
not deferred -- the INSERT_SUBREG rewriting. Due to quirks of how this
rewriting proceeded, not only did it occur early, it removed the bits of
information needed for the deferred processing to correctly generate the
necessary two address form (specifically inserting a copy), but didn't
trigger any immediate assertions and produced what appeared to be
already valid two-address from code. Thus, the assertion only fired much
later in the pipeline.

The fix is to hoist the transformation logic up layer to where it can
more firmly defer all further processing, and to teach the normal
processing to handle an edge case previously handled as part of the
transformation logic. This edge case (already matched tied register
operands) needs to *not* defer any steps.

As has been brought up repeatedly in the process: wow does this code
need refactoring. I *may* squeeze in some time to at least bring sanity
to this loop... but wow... =]

Thanks to Jakob for helpful hints on the way here, and the review.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160443 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-18 18:58:22 +00:00
Nuno Lopes
4532bf6ecf ignore 'invoke @llvm.donothing', but still keep the edge to the continuation BB
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160411 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-18 00:07:17 +00:00
Evan Cheng
a9e13ba3c8 Back out r160101 and instead implement a dag combine to recover from instcombine transformation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160387 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-17 18:54:11 +00:00
Jakob Stoklund Olesen
a532bcec2f Add some trace output to TwoAddressInstructionPass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160380 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-17 17:57:23 +00:00
Benjamin Kramer
de39671ebb Remove unused variable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160372 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-17 17:00:11 +00:00
Nadav Rotem
5589a69f0a Fix a crash in the legalization of large vectors.
When truncating a result of a vector that is split we need
to use the result of the split vector, and not re-split the dead node.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160357 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-17 09:07:37 +00:00
Evan Cheng
f5c0539092 Implement r160312 as target indepedenet dag combine.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160354 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-17 08:31:11 +00:00
Evan Cheng
b4d4959fdd Make sure constant bitwidth is <= 64 bit before calling getSExtValue().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160350 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-17 07:47:50 +00:00
Evan Cheng
70e10d3fe4 This is another case where instcombine demanded bits optimization created
large immediates. Add dag combine logic to recover in case the large
immediates doesn't fit in cmp immediate operand field.

int foo(unsigned long l) {
  return (l>> 47) == 1;
}

we produce

  %shr.mask = and i64 %l, -140737488355328
  %cmp = icmp eq i64 %shr.mask, 140737488355328
  %conv = zext i1 %cmp to i32
  ret i32 %conv

which codegens to

movq    $0xffff800000000000,%rax
andq    %rdi,%rax
movq    $0x0000800000000000,%rcx
cmpq    %rcx,%rax
sete    %al
movzbl    %al,%eax
ret

TargetLowering::SimplifySetCC would transform
(X & -256) == 256 -> (X >> 8) == 1
if the immediate fails the isLegalICmpImmediate() test. For x86,
that's immediates which are not a signed 32-bit immediate.

Based on a patch by Eli Friedman.

PR10328
rdar://9758774


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160346 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-17 06:53:39 +00:00
Nadav Rotem
c76fa8937d Minor cleanup and docs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160311 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-16 18:56:39 +00:00
Nadav Rotem
7ee0e5ae60 Make ComputeDemandedBits return a deterministic result when computing an AssertZext value.
In the added testcase the constant 55 was behind an AssertZext of type i1, and ComputeDemandedBits
reported that some of the bits were both known to be one and known to be zero.

Together with Michael Kuperstein <michael.m.kuperstein@intel.com>



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160305 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-16 18:34:53 +00:00
Nadav Rotem
46646572f7 Fix a bug in the scalarization of BUILD_VECTOR. BUILD_VECTOR elements may be wider than the output element type. Make sure to trunc them if needed.
Together with Michael Kuperstein <michael.m.kuperstein@intel.com>



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160235 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-15 20:39:08 +00:00
Nadav Rotem
b87bdac6a3 Refactor the code that checks that all operands of a node are UNDEFs.
Add a micro-optimization to getNode of CONCAT_VECTORS when both operands are undefs.
Can't find a testcase for this because VECTOR_SHUFFLE already handles undef operands, but Duncan suggested that we add this.

Together with Michael Kuperstein <michael.m.kuperstein@intel.com>



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160229 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-15 08:38:23 +00:00
Chandler Carruth
7d532c8d07 Reapply r160194, switching to use LV information for finding local kills.
The notable fix is to look at any dependencies attached to the kill
instruction (or other instructions between MI nad the kill) where the
dependencies are specific to the register in question.

The old code implicitly handled this by rejecting the transform if *any*
other uses were found within the block, but after the start point. The
new code directly finds the kill, and has to re-use the existing
dependency scan to check for non-kill uses.

This was caught by self-host, but I found the bug via inspection and use
of absurd assert scaffolding to compute the kills in two ways and
compare them. So I have no useful testcase for this other than
"bootstrap". I'd work harder to reduce a test case if this particular
code were likely to live for a long time.

Thanks to Benjamin Kramer for reviewing the fix itself.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160228 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-15 03:29:46 +00:00
Nadav Rotem
b7e230d999 Add a dagcombine optimization to convert concat_vectors of undefs into a single undef.
The unoptimized concat_vectors isd prevented the canonicalization of the vector_shuffle node.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160221 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-14 21:30:27 +00:00
Jakob Stoklund Olesen
27982e1108 Account for early-clobber reload instructions.
No test case, there are no in-tree targets that require this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160219 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-14 18:45:35 +00:00
Jakob Stoklund Olesen
c8981f2e3d Be more verbose when detecting dominance problems.
Catch uses of undefined physregs that haven't been added to basic block
live-in lists. Run the verifier to pinpoint the problem.

Also run the verifier when a virtual register use is not jointly
dominated by defs.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160207 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-13 23:39:05 +00:00
Chandler Carruth
cb41e5f6f2 Revert r160194, which switched to use LV information for finding local
kills.

This is causing miscompiles that I'm working on tracking down.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-13 22:23:32 +00:00
Chandler Carruth
5583415909 Use the LiveVariables information to efficiently get local kills. This
removes the largest scaling problem in the test cases from PR13225 when
ASan is switched to insert basic blocks in the natural CFG order.

It may also solve some scaling problems for more normal code with large
numbers of basic blocks and variables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160194 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-13 21:18:38 +00:00
Jim Grosbach
f51f1a7dd6 Provide function name in 'Cannot select' fatal error.
When dumping the DAG for a fatal 'Cannot select' back-end error, also
provide the name of the function the construct is in. Useful when dealing
with large testcases, as the next step is to llvm-extract the function
in question to get a small(er) testcase.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160152 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-13 00:29:09 +00:00
Eric Christopher
5cf55e1c6a The end of the prologue should be marked with is_stmt.
Fixes PR13303.

Patch by Paul Robinson!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160148 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-12 23:30:25 +00:00
Duncan Sands
4e8982a34d The result type of EXTRACT_VECTOR_ELT doesn't have to match the element type of
the input vector, it can be bigger (this is helpful for powerpc where <2 x i16>
is a legal vector type but i16 isn't a legal type, IIRC).  However this wasn't
being taken into account by ExpandRes_EXTRACT_VECTOR_ELT, causing PR13220.
Lightly tweaked version of a patch by Michael Liao.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160116 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-12 09:01:35 +00:00
Evan Cheng
875913439c InstrEmitter::EmitSubregNode() optimize extract_subreg in this case:
r1025 = s/zext r1024, 4
r1026 = extract_subreg r1025, 4

to a copy:
r1026 = copy r1024

This is correct. However it uses TII->isCoalescableExtInstr() which can return
true for instructions which essentially does a sext_in_reg so this can end up
with an illegal copy where the source and destination register classes do not
match. Add a check to avoid it. Sorry, no test case possible at this time.

rdar://11849816


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@160059 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-11 18:55:07 +00:00