Enables llvm dyn_cast for Assemblers.

Enables llvm dyn_cast for Assemblers.

IceCfg::getAssembler() is a template that simply static_casts the CFG's assembler. This could potentially be problematic in the future, so we enabled the (relatively) cheap llvm dyn_cast operator for Assemblers.

This CL also renames assembler_mips32.h to IceAssemblerMIPS32.h.

BUG=
R=stichnot@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=2da710ce9fe3bf20b47aacc16e2f6bccc45370c7

[John, 2015-06-26 23:57:55]

jpp@chromium.org changed reviewers:
+ jvoung@chromium.org, kschimpf@google.com, stichnot@chromium.org

[Jim Stichnoth, 2015-06-27 13:52:20]

otherwise lgtm

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerARM32.h
File src/IceAssemblerARM32.h (right):

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerARM32.h#new...
src/IceAssemblerARM32.h:33: AssemblerARM32(const AssemblerARM32 &) = delete;
Should there be
  AssemblerARM32() = delete;
?  Also in the X86 files.

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerMIPS32.h
File src/IceAssemblerMIPS32.h (right):

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerMIPS32.h#ne...
src/IceAssemblerMIPS32.h:1: //===- subzero/src/assembler_mips.h - Assembler for
MIPS --------*- C++ -*-===//
Update the filename.

BTW, Reed was doing this rename as part of
https://codereview.chromium.org/1176133004/ , but given that he's away for a few
weeks, this will be a drop in the bucket of rebase changes for him. :(

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerMIPS32.h#ne...
src/IceAssemblerMIPS32.h:22: #ifndef SUBZERO_SRC_ASSEMBLER_MIPS32_H
Update the guard to the new filename

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerMIPS32.h#ne...
src/IceAssemblerMIPS32.h:73: return Asm->getKind() == Asm_ARM32;
Asm_MIPS32

[John, 2015-06-29 14:56:02]

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerARM32.h
File src/IceAssemblerARM32.h (right):

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerARM32.h#new...
src/IceAssemblerARM32.h:33: AssemblerARM32(const AssemblerARM32 &) = delete;
On 2015/06/27 13:52:19, stichnot wrote:
> Should there be
>   AssemblerARM32() = delete;
> ?  Also in the X86 files.

Good question. I thought we didn't need one for we have

AssemblerARM32(bool = false);

and I did not know what the compiler would do. It turns out that, if we add

AssemblerARM32() = delete

we get errors that declarations like

AssemblerARM32 a;

are ambiguous (which makes sense -- do you want AssemblerARM32() or
AssemblerARM32(bool)?)

(PS: I am not a fan of default parameters for problems like these. :))

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerMIPS32.h
File src/IceAssemblerMIPS32.h (right):

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerMIPS32.h#ne...
src/IceAssemblerMIPS32.h:1: //===- subzero/src/assembler_mips.h - Assembler for
MIPS --------*- C++ -*-===//
On 2015/06/27 13:52:19, stichnot wrote:
> Update the filename.
> 
> BTW, Reed was doing this rename as part of
> https://codereview.chromium.org/1176133004/ , but given that he's away for a
few
> weeks, this will be a drop in the bucket of rebase changes for him. :(

Yeah, I thought about not renaming the file, but it seemed harmless enough. I am
sorry I caused him more pain. :(

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerMIPS32.h#ne...
src/IceAssemblerMIPS32.h:22: #ifndef SUBZERO_SRC_ASSEMBLER_MIPS32_H
On 2015/06/27 13:52:20, stichnot wrote:
> Update the guard to the new filename

Done.

https://codereview.chromium.org/1211863004/diff/1/src/IceAssemblerMIPS32.h#ne...
src/IceAssemblerMIPS32.h:73: return Asm->getKind() == Asm_ARM32;
On 2015/06/27 13:52:19, stichnot wrote:
> Asm_MIPS32

Done.

[John, 2015-06-29 14:57:06]

Committed patchset #2 (id:20001) manually as
2da710ce9fe3bf20b47aacc16e2f6bccc45370c7 (tree was closed).