There were a lot of unnecessary copying and resizing and sloppiness in the use of BitVector for liveness analysis.  This tries to reduce the amount of copying and associated memory allocation.

Also, works around a ConstantRelocatable hashing problem that was causing a huge slowdown in MINIMAL mode for the vector_shuffle scons test.

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1746613002 .

BUG= None
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1741733003 .

The problem is that when switch lowering decides to use a JumpTable and the switch variable is i64 on x86-64, the lowering tries to movzx the i64 variable into an i32 variable, and the Movzx ctor asserts.

This happens when translating pnacl-llc.pexe, but luckily it is also triggered in the existing adv-switch.ll test.

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1743133002 .

One is rarely interested in seeing dump output regarding global initializers, so we add an additional verbose flag, and don't include it in the "-verbose most" set.

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1744913002 .

Along with the nexe equivalents of pnacl-sz, also build the browser-integrated versions that can be dropped in as replacements.  Note that these are not exact replacements, as they likely support all targets.

In "make presubmit", instead build sandboxed translators in DEBUG=1 mode, so they can build in parallel with the assembler unit tests.

Add a "make bloat-sb" target that does the "make bloat" equivalent for the sandboxed translators.

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1746593002 .

This reverts commit 352db935 from https://codereview.chromium.org/1740033002/ .

That commit broke "make -f Makefile.standalone check-lit check-xtest".

BUG= none
R=smklein@chromium.org

Review URL: https://codereview.chromium.org/1745783002 .

Adds "-verbose mem" to enable printing the amount of CfgLocalAllocator memory allocated, each time Cfg::dump() is called.

"-verbose mem,status" is a good way to get summary info.

BUG= none
R=kschimpf@google.com

Review URL: https://codereview.chromium.org/1743043002 .

This CL updates "isPNaClABIExternalName" -- Subzero checked for the symbol
"__pnacl_pso_root" as a function, but it is a declaration, and should be
checked as one. Additionally, when the PNaClTranslator is verifying the linkage
of declarations, allow "__pnacl_pso_root" to be flipped to external as a special
case.

TEST=pnacl-translate -pso --use-sz -arch x86-32-nonsfi test_pll.final.pso
BUG=https://bugs.chromium.org/p/nativeclient/issues/detail?id=4351
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1740033002 .

BUG=

Review URL: https://codereview.chromium.org/1742833002 .

1) Clones llvm::BitVector, and makes it Allocator aware (using the
CfgLocalAllocator<>)

2) Uses mallopt to set the malloc granularity. The default granularity
is too small, which forces too many mmap calls.

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

Review URL: https://codereview.chromium.org/1738683003 .

See failed build at https://build.chromium.org/p/tryserver.nacl/builders/nacl-toolchain-win7-pnacl-x86_64/builds/3441 .

BUG= none
TBR=jpp@chromium.org

Review URL: https://codereview.chromium.org/1732233002 .

Instead of "./pnacl-sz <args>", one can run:

  ../../../run.py ./pnacl-sz.x8632.nexe <args>
or
  ../../../run.py ./pnacl-sz.x8664.nexe <args>

Hopefully the translator performance characteristics are close to that of the browser hookup.

Adds sb builds to two of the presubmit configurations.

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1738633002 .

Introduces Ice::SmallBitVector.

Modifies ConstantRelocatables so that known offsets (i.e.,
not offsets to the code stream) do no require GlobalContext
allocations.

Modifies Cfg-local containers to use the CfgLocalAllocator.

BUG=
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1738443002 .

TBR=jpp@chromium.org

BUG= none

Review URL: https://codereview.chromium.org/1730263002 .

This is part of ARM patch:
https://codereview.chromium.org/1151663004/

BUG=
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1724643002 .

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

Implement 64 bit multiply in mips32 and, in addition, add the lo/hi registers which are also used for other 64 bit math such as div, rem.

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

Review URL: https://codereview.chromium.org/1716483003 .

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

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4076
R=eholk@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1708753002 .

This removes most of the #ifndef ARM32 directives so we get more
thorough testing. It still uses fewer iterations due to running on
Qemu, but I did manually check to be sure that the tests pass without
any special casing at all.

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

Review URL: https://codereview.chromium.org/1708903002 .

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

Review URL: https://codereview.chromium.org/1685253003 .

With this change, we pass the select crosstest.

Since this would have introduced a three-argument version of
scalarizeInstruction, I decided to generalize it using templates.

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

Review URL: https://codereview.chromium.org/1683243003 .

X8664_STACK_HACK was an intrusive way to get the crosstests to run
while x32 support was not implemented in llvm. The hack is longer
needed.

R=eholk@chromium.org

Review URL: https://codereview.chromium.org/1706883003 .

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

Review URL: https://codereview.chromium.org/1697263007 .

It's troublesome that we have to type (this) in several places
but I don't see a away around it yet and it does not add to the line
of code count at all to have to do this; it's just not aesthetic to me.

If this idea is okay, I can make all the similar changes
in the Mips Subzero port.

BUG=

Review URL: https://codereview.chromium.org/1661233002 .

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

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

Review URL: https://codereview.chromium.org/1409863006 .

The original implementation hard-coded the target to x86-32.

The new implementation mirrors pnacl-llc by probing the architecture it is running on.

Continues the plumbing work in https://codereview.chromium.org/1698523003/ , toward the goal of enabling sandboxed translator tests for x86-64, particularly on the spec bots.

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4077
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1696123003 .

There are many occurrences of if (RegNum == RegNumT::NoRegister).

This patch eliminates NoRegister and provides a simpler mechanism for declaring and testing RegNumT values to see if they are undefined.

BUG= none
R=stichnot@chromium.org

Review URL: https://codereview.chromium.org/1691193002 .

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

Implements a vector move as a vector orr on the corresponding Q
registers.

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

Review URL: https://codereview.chromium.org/1694533002 .

Adds nonsfi support to the ARM32 backend.

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

Review URL: https://codereview.chromium.org/1665263003 .

Adds class name for the "QtoS" register class.

BUG=None
R=eholk@chromium.org

Review URL: https://codereview.chromium.org/1687163002 .

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

Review URL: https://codereview.chromium.org/1679023008 .

This is part of a sequence of patches to quickly fill out vector
support by scalarizing the remaining operations. Later we can work to
generate better code.

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

Review URL: https://codereview.chromium.org/1683153003 .

Originally, register numbers were represented explicitly as int32_t, particularly so that -1 (or negative values in general) could be used as a NoRegister sentinel value.  This created two problems:

1. It would be better to use a unique name for the type, to distinguish from other explicit int32_t uses such as stack offsets.

2. Apart from NoRegister, register number values ultimately come from unsigned sources like enum values and bitvector positions.  This results in a number of clumsy casts to remove compiler warnings.

This creates a simple RegNumT class to manage this.  It also deletes ordered comparison operators to help catch errors where particular register number orderings are assumed (as opposed to orderings of the encoded register values).

In addition, it creates a RegNumBitVector wrapper class that makes it much cleaner to do range-based for loops over bit vectors that represent RegNumT sets.

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

Review URL: https://codereview.chromium.org/1676123002 .

CL https://codereview.chromium.org/1687553002 introduced a bug when
running:

   make -f Makefile.standalone check-lit FORCEASM=1

The cause of the problem is the way options "--asemble --disassemble"
work in run-pnacl-sz.py. When compiling using "--filetype=asm", the
assembler writes:

   .word 0xe7fedef0

The output after assembly/disassembly is the same as above.

On the other hand, when compiling using "--filetype=iasm", the assembler
writes:

   .byte 0xe7
   .byte 0xfe
   .byte 0xde
   .byte 0xf0

While the same sequence of bytes is assembled, the dissassembly for the
latter generates assembly instruction:

    udf #60896 ; 0xede0

The fix is to not check the generated disassembled instructions. Rather,
have it check if the same word is associated with the assembly
instruction.

Longer term, we should fix the several different ways --filetype=asm
introduces this instruction to match the "udf ..." output.

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

Review URL: https://codereview.chromium.org/1682253003 .

Enables vector division by scalarization.

Also, removed an assert as suggested by Karl in a previous CL:
https://codereview.chromium.org/1646033002/diff/1/src/IceInstARM32.cpp#newcode717

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

Review URL: https://codereview.chromium.org/1681003002 .

To view the non-whitespace changes in this CL:

  git cl patch -b testbranch 1678133003
  git diff -w --ignore-blank-lines -b master

Such changes are only in gen_arm32_reg_tables.py and IceInst.cpp.

There are lots of tab characters in .ll files that shouldn't be there, but fixing them would require some thought about how to do consistent formatting, so that's left for later.

BUG= none
R=eholk@chromium.org, kschimpf@google.com

Review URL: https://codereview.chromium.org/1678133003 .

This is done by putting far pointers (for calls) into a
register and then do an indirect call.

This was done to guarantee that we aren't putting a size limit on ARM32 executables.

BUG= https://bugs.chromium.org/p/nativeclient/issues/detail?id=4076
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1687553002 .

This change allows pnacl-sz to randomly insert NOPs into the generated
code, as is already done with X86.

BUG=None
R=eholk@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1670413002 .

BUG=None
R=jpp@chromium.org, stichnot@chromium.org

Review URL: https://codereview.chromium.org/1669973002 .

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

Review URL: https://codereview.chromium.org/1655313002 .

Some casts to size_t for use as array indexes are simply unnecessary.

Some explicit declaration types are changed to "auto" to avoid redundancy with the static_cast type.

A few llvm::dyn_cast<> operations are changed to llvm::cast<>, and vice versa.

A few explicit declaration types are changed to "auto" when used with llvm::cast<> and llvm::dynamic_cast<>.  Some of these were missed during an earlier cleansing because of multi-line issues.

There are still a few opportunities related to Variable register numbers, but they are ignored for now because they are being addressed in another CL.

BUG= none
R=jpp@chromium.org

Review URL: https://codereview.chromium.org/1674033002 .