llvm-svn: 60850

On x86 favors folding short immediate into some arithmetic operations (e.g. add, and, xor, etc.) because materializing an immediate in a register is expensive in turns of code size.
e.g.
movl 4(%esp), %eax
addl $4, %eax

is 2 bytes shorter than

movl $4, %eax
addl 4(%esp), %eax

llvm-svn: 60139

special-purpose hook to a new pass. Also, add check to see if any
x87 virtual registers are used, to avoid doing any work in the
common case that no x87 code is needed.

llvm-svn: 59190

to be sign-extended when it is promoted to 64 bits for intermediate
offset calculations. The offset calculations are done as uint64_t so that
overflow conditions are well defined.

This fixes a problem which is currently hidden by the x86 AsmPrinter but
which was exposed by r58917 (which is temporarily reverted).  See PR3027
for details.

llvm-svn: 59044

priority function. Instead, just iterate over the AllNodes list, which is
already in topological order. This eliminates a fair amount of bookkeeping,
and speeds up the isel phase by about 15% on many testcases.

The impact on most targets is that AddToISelQueue calls can be simply removed.

In the x86 target, there are two additional notable changes.

The rule-bending AND+SHIFT optimization in MatchAddress that creates new
pre-isel nodes during isel is now a little more verbose, but more robust.
Instead of either creating an invalid DAG or creating an invalid topological
sort, as it has historically done, it can now just insert the new nodes into
the node list at a position where they will be consistent with the topological
ordering.

Also, the address-matching code has logic that checked to see if a node was
"already selected". However, when a node is selected, it has all its uses
taken away via ReplaceAllUsesWith or equivalent, so it won't recieve any
further visits from MatchAddress. This code is now removed.

llvm-svn: 58748

have its node id set. The new and and shift nodes are the nodes that need
the IDs. This fixes PR2982.

llvm-svn: 58655

Have TableGen emit setSubgraphColor calls under control of a -gen-debug
flag.  Then in a debugger developers can set breakpoints at these calls
to see waht is about to be selected and what the resulting subgraph
looks like.  This really helps when debugging instruction selection.

llvm-svn: 58278

and add a TargetLowering hook for it to use to determine when this
is legal (i.e. not in PIC mode, etc.)

This allows instruction selection to emit folded constant offsets
in more cases, such as the included testcase, eliminating the need
for explicit arithmetic instructions.

This eliminates the need for the C++ code in X86ISelDAGToDAG.cpp
that attempted to achieve the same effect, but wasn't as effective.

Also, fix handling of offsets in GlobalAddressSDNodes in several
places, including changing GlobalAddressSDNode's offset from
int to int64_t.

The Mips, Alpha, Sparc, and CellSPU targets appear to be
unaware of GlobalAddress offsets currently, so set the hook to
false on those targets.

llvm-svn: 57748

llvm-svn: 57649

llvm-svn: 57508

create a new DAG node to represent the new shift to keep the
DAG consistent, even though it'll almost always be folded into
the address.

If a user of the resulting address has multiple uses, the
nodes may get revisited by a later MatchAddress call, in which
case DAG inconsistencies do matter.

This fixes PR2849.

llvm-svn: 57465

optimized for size. Set OptForSize for each function separately.

llvm-svn: 57182

Make it all work in non-pic mode.

llvm-svn: 57034

incidentally making the case where the memop is a
pointer deref work.  Fix cmp-and-swap regression.

llvm-svn: 57027

Instead, just create one, and make sure everything that needs
it can access it. Previously most of the SelectionDAGISel
subclasses all had their own TargetLowering object, which was
redundant with the TargetLowering object in the TargetMachine
subclasses, except on Sparc, where SparcTargetMachine
didn't have a TargetLowering object. Change Sparc to work
more like the other targets here.

llvm-svn: 57016

llvm-svn: 57014

isReg, etc., from isRegister, etc.

llvm-svn: 57006

llvm-svn: 56963

llvm-svn: 56937

Completely eliminate the TopOrder std::vector. Instead, sort
the AllNodes list in place. This also eliminates the need to
call AllNodes.size(), a linear-time operation, before
performing the sort.

Also, eliminate the Sources temporary std::vector, since it
essentially duplicates the sorted result as it is being
built.

This also changes the direction of the topological sort
from bottom-up to top-down. The AllNodes list starts out in
roughly top-down order, so this reduces the amount of
reordering needed. Top-down is also more convenient for
Legalize, and ISel needed only minor adjustments.

llvm-svn: 56867

and X86FastISel.cpp into X86MachineFunction.h, so that it
can be shared, instead of having each selector keep track
of its own.

llvm-svn: 56825

llvm-svn: 56727

Implement "punpckldq  %xmm0, $xmm0" as "pshufd  $0x50, %xmm0, %xmm" unless optimizing for code size.

llvm-svn: 56711

consistancy with ConstantInt, and re-implement it in terms
of ConstantInt's getSExtValue.

llvm-svn: 56700

require RIP-relative addressing and use it to fix a bug
in X86FastISel in x86-64 PIC mode, where it was trying to
use base/index registers with RIP-relative addresses. This
fixes a bunch of x86-64 testsuite failures.

llvm-svn: 56676

Properly handle 'm' inline asm constraints. If a GV is being selected for the addressing mode, it requires the same logic for PIC relative addressing, etc.

llvm-svn: 56526

llvm-svn: 56495

X86ISelDAGToDAG.cpp and into X86InstrInfo.cpp. This will allow
it to be reused by FastISel.

llvm-svn: 56494

up some new ascii art to illustrate what it does. This change
currently has no effect on generated code.

llvm-svn: 56270

Apologies for the thrashing.

llvm-svn: 56251

- Add linkage to SymbolSDNode (default to external).
- Change ISD::ExternalSymbol to ISD::Symbol.
- Change ISD::TargetExternalSymbol to ISD::TargetSymbol

These changes pave the way to allowing SymbolSDNodes with non-external linkage.

llvm-svn: 56249

with ConstantInt. This led to fixing a bug in TargetLowering.cpp
using getValue instead of getAPIntValue.

llvm-svn: 56159

llvm-svn: 55588

llvm-svn: 55504

llvm-svn: 55394

Try approach to moving call address load inside of callseq_start. Now it's done during the preprocess of x86 isel. callseq_start's chain is changed to load's chain node; while load's chain is the last of callseq_start or the loads or copytoreg nodes inserted to move arguments to the right spot.

llvm-svn: 55338

process up to a higher level. This allows FastISel to leverage
more of SelectionDAGISel's infastructure, such as updating Machine
PHI nodes.

Also, implement transitioning from SDISel back to FastISel in
the middle of a block, so it's now possible to go back and
forth. This allows FastISel to hand individual CallInsts and other
complicated things off to SDISel to handle, while handling the rest
of the block itself.

To help support this, reorganize the SelectionDAG class so that it
is allocated once and reused throughout a function, instead of
being completely reallocated for each block.

llvm-svn: 55219

from all targets.

llvm-svn: 55124

out of X86ISelDAGToDAG.cpp C++ code and into tablegen code.
Among other things, using tablegen for these things makes them
friendlier to FastISel.

Tablegen can handle the case of i8 subregs on x86-32, but currently
the C++ code for that case uses MVT::Flag in a tricky way, and it
happens to schedule better in some cases. So for now, leave the
C++ code in place to handle the i8 case on x86-32.

llvm-svn: 55078

Fix a (u)comiss intrinsic lowering bug. It was using anyext which can return junk in higher bits. Patch by Nate Begeman.

llvm-svn: 54903