To see the main effect of this, look at the doxygen for file
IceGlobalContext.h to see the effect of turning this on.
Without it, the nested classes are not available because they
are private.

My guess is that often doxygen is used to generate API documents and for that reason, exporting of private info is excluded by default.

BUG=
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1411123019 .

Creates a special OperandARM32 for representing imm5 used when
performing a shift operation.

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4076
R=kschimpf@google.com

Review URL: https://codereview.chromium.org/1449263003 .

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1458523002 .

1. Include a test build (but no link) using g++, to identify errors and warnings before they hit the Windows bots.

2. Move "git diff --quiet" to the end so that the presubmit tests can easily be run before committing locally, if desired.

3. Add "make presubmit" as an alias for "make check-presubmit".

4. Document the individual steps.

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1453713002 .

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4076
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1438773004 .

BUG= none
TEST= make -j32 -f Makefile.standalone CXX=g++ LLVM_EXTRA_WARNINGS="-Wno-unknown-pragmas -Wno-unused-parameter -Wno-comment -Wno-enum-compare -Wno-strict-aliasing" STDLIB_FLAGS=
(this command will compile but fail to link)

R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1452993002 .

The register allocator does not need to be considering rematerializable variables at all.  When it does, there tends to be a big performance cost because the live range of a rematerializable variable tends to be large, leading to lots of expensive overlap computations against register allocation candidates.

BUG= none
R=jpp@chromium.org, sehr@chromium.org

Review URL: https://codereview.chromium.org/1450233002 .

Rather than bumping the stack pointer around the scalar return sequence in
_fld, ensure the prolog allocates enough space.

BUG=
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1442753008 .

Also removes reliance on lowerAlloca entirely for the fixed allocations.

BUG=
R=jpp@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1435363002 .

No functional changes, as measured by identical spec2k asm output.

1. Use early "return" and "continue" to reduce "if" nesting.

2. Reflow comments to 80 columns (instead of presumably 79).

3. Add some BuildDefs::dump() tests to reduce translator code size.

BUG= none
R=kschimpf@google.com

Review URL: https://codereview.chromium.org/1448773002 .

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1452553002 .

Fixes case where the ARM integrated assembler for class InstARM32Mov
did not revert to using the stand-alone assembler (method emit) to
generate the corresponding assembly instructions(s).

This fixes last known problem with method emitIAS (other than
reverting to the stand-alone assembler if not implemented). Removes
use of workaround flag "-unsafe-ias".

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1448783004 .

Lots of definitions of Target outside of asserts.

BUG=

Review URL: https://codereview.chromium.org/1448673002 .

This involves changing AdjustStack to grow/shrink the stack, and to use that
operation exclusively to move the StackAdjustment variable in lowering, rather
than in call emission as before.

BUG=
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1449523002 .

There are situations where a variable is assigned as the result of a rematerializable alloca instruction, and then another variable is assigned as essentially a known-offset interior pointer into the alloca space.  In this case, the secondary variable is also rematerializable.

We add a pass, after alloca analysis, to find these derived variables and mark them transitively as rematerializable.  Because we lack use-def chains (or in fact any map to variable use locations), we need to iterate over the CFG until convergence.  Fortunately, this is pretty cheap, and not even done unless the alloca analysis seeds it with an initial set of rematerializable variables.

This analysis is only really needed for arithmetic instructions, but we also need to apply it to assignments and pointer-type bitcasts that are added when the IceConverter directly parses a .ll file rather than a .pexe file.

BUG= none
R=jpp@chromium.org, sehr@chromium.org

Review URL: https://codereview.chromium.org/1441793002 .

BUG= none
R=jpp@chromium.org, sehr@chromium.org

Review URL: https://codereview.chromium.org/1439363002 .

Remove a bit of complex repeated template naming.

BUG=
R=jpp@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1438933002 .

getOffset() was being dereferenced even when it was nullptr.

BUG= none
R=kschimpf@google.com

Review URL: https://codereview.chromium.org/1435283003 .

This just prints the function being translated, once per function.  This is useful if there is a crash or fatal error somewhere, and you want to quickly discover which function to set -verbose-focus on.

Also, fixes some warnings/errors in the MINIMAL build.

BUG= none
R=kschimpf@google.com

Review URL: https://codereview.chromium.org/1439983002 .

These two methods introduce multiple instructions. Between each
instruction a newline must be inserted, and a call to startNextInt().
This CL makes sure both cases are met.

BUG=None
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1441023002 .

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1432413003 .

This CL passes in context from the ARM target lowering, so that
it can figure out what register (SP or FP) and offset to use
when loading/storing variables.

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1414043015 .

Partition allocas that occur in the entry block into two categories.  The first is those whose size is fixed and alignment are less than or equal to the stack alignment.  These are emitted relative to a pointer, either in increasing offset relative to the stack pointer or decreasing offset relative to the frame pointer.  (Actually, we are not enabling this optimization for frame pointer frames yet)  The second category is allocas whose size is dynamic or alignment is creater than the stack alignment.  These are emitted relative to a user variable in increasing offset order.  This optimization is only enabled for x86 at O2.

BUG=
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1411583007 .

Extends assembler movw method to handle integer constants in addition
to relocatable constants, since (32-bit) integer constants are
frequently used in movw instructions.

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1440693002 .

Improves the bool folding logic so that branches are short circuited.

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4076
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1417393003 .

Removes blank line after push instructions.

BUG=None
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1439683002 .

Fold and/or followed by branch to eliminate cmp.  Also, fold fcmp instructions into branches similarly to what was done for icmp instructions.

BUG=
R=jpp@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1436623002 .

Originally, for each non-deleted instruction, CfgNode::emit() would call the virtual Inst::emit() and then print a newline (also printing end-of-live-range info as necessary).  This resulted in clumsy blank lines in the asm output, corresponding to non target specific pseudo instructions such as FakeDef, FakeUse, FakeKill.

We change this so that CfgNode::emit() only prints a newline for an InstTarget subclass, or if any end-of-live-range text was printed.

If a high-level instruction still wants to emit something in a comment, it's responsible for printing its own newline.

BUG= none
TEST= ./pydir/szbuild_spec2k.py -O2 --force -v --filetype=asm --sz=--asm-verbose=0
TEST= ./pydir/szbuild_spec2k.py -O2 --force -v --filetype=asm --sz=--asm-verbose=1
TEST= ./pydir/szbuild_spec2k.py -O2 --force -v --filetype=asm --target=arm32
R=kschimpf@google.com

Review URL: https://codereview.chromium.org/1431353003 .

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=jpp@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1433743002 .

This enables the following workflow:

1. Run "szbuild.py --filetype=asm ..." to generate a .s file.
2. Edit the .s file to mock up a Subzero change.
3. Run "szbuild.py --filetype=asm --no-sz ..." to restart the build post pnacl-sz.

This workflow is good for trying out localized changes to hot basic blocks, though it isn't really appropriate for global optimizations like register allocation.

BUG= none
R=kschimpf@google.com

Review URL: https://codereview.chromium.org/1417003005 .

BUG=
R=jpp@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1430273004 .

Patch from Reed Kotler <rkotlerimgtec@gmail.com>.

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4076
R=kschimpf@google.com, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1422753010 .

Also formatted IceCfg.cpp, since it needed it.

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1422253003 .

First, the width was not being correctly defined for non-vector
instructions.

Second, the order of the width/condition was incorrect when the
instruction was prefixed with a V. That is, for V prefixed instructions,
the order is predicate/width while for non-V prefixed instructions the
order is width/predicate.

Also fixes bug in target lowering that did not always convert results
of a compare to i1.

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1415953007 .

Sets up additional register attributes, plus the notion of register classes, to enable robust usage of the high 8-bit GPRs (ah/bh/ch/dh), for both x86-32 and x86-64.  (Note that the x86-64 changes are currently untested.)

We add a Register Class field to the Variable class.  The default register class is a value corresponding to the variable's type, but the target can extend the set of register class values, and the target lowering can assign different register classes as needed.  The register allocator uses the register class instead of the type to determine the set of registers to draw from.

For x86-64, the high 8-bit registers are not included in the general register allocation pool, but there are explicit references to ah for lowering the div/rem instructions.

The target lowering is modified as needed to make sure types are appropriate and register use in instructions is legalized.

Some other fixes and cleanups are included in this CL:

* Makefile.standalone changes.  Source files are reordered so that the more expensive compiles are done earlier, speeding up parallel builds by decreasing fragmentation.  A dependency error is fixed for check-spec.

* A bug is fixed in advanced phi lowering.  When a temporary is introduced to break a cycle, we were neglecting to updated the predecessor count for one of the operands, leading to an assertion failure.  (Applying that fix to master resulted in no changes to spec2k code generation.)  A consistency check is added to help find future problems like this.  Also, refactored iteration over the Phi descriptor array to use range-based for loops and avoid directly indexing the array.

* Removed most of the "IceType_" prefixes in x-macro tables for brevity.

* Fix a correctness TODO in the register allocator.  This had no effect on spec2k code generation in master or in this CL, so we were probably just lucky.

* Made some much-needed s/Dest->getType()/Ty/ changes for brevity, in the target lowering sections that needed other changes.

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4095
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1427973003 .

When a temporary is introduced to break a cycle, we neglected to update the predecessor count for one of the operands, leading to a possible assertion failure.

This problem isn't currently seen in master, but it arises when we enable register aliases, as in https://codereview.chromium.org/1427973003/ .  No changes are seen in spec2k code generation as a result of this fix.

A consistency check is added to help find future problems like this.

Also, refactored iteration over the Phi descriptor array to use range-based for loops and avoid directly indexing the array.

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1435543002 .

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1416113007 .

Split allocas in the entry block into two categories.  The first has alignment <= stack alignment and constant size.  The second violates one or both of those conditions.  Sort both of these lists in descending alignment order and emit.  Also, compute the need for a frame pointer during the pass.

BUG=
R=jpp@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1414343010 .

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1412963008 .

The integrated ARM assembler was incorrectly assuming that an
(instruction) label defines a corresponding assembler instruction.
Therefore, placement into the buffer was incorrect.

This CL fixes this mistake. This fixes assembler translator problems
with ARM branch, movw, and movt instructions.

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4334
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1430973003 .