Subzero: Assemble calls to constant addresses.

Assemble calls to constant addresses.

Finally address this TODO in the assembler. This will help translate
non-IRT using programs (no ABI stability). The default scons testing
mode is non-IRT, so this helps with that. I haven't actually tested
this against scons yet, but I'm filling in the tests based on how
LLVM translates the same bitcode.

The filetype=asm is adjusted to omit the "*" and the "$".

The filetype=obj is adjusted to check for fixups with NullSymbols,
and also fill the assembler buffer at the instruction's immediate
field w/ the right constant.

The filetype=iasm is still TODO (hits an new assert in the Fixup's emit() function).

Reverts 7ad1bed99d058199a3ba246a5383458518596fbc:
"Allow stubbing of called constant addresses using command line argument."
since this is now handled (except for iasm).

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4080
R=stichnot@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=f644a4b3a36d707b32dd4b7855774a346a575e96

[jvoung (off chromium), 2015-03-19 00:37:53]

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

[Jim Stichnoth, 2015-03-19 01:09:55]

Nice!  lgtm

I guess it would be nice to get filetype=iasm working for this at some point. 
Currently, filetype=asm doesn't completely work because llvm-mc doesn't get the
lock and 16-bit prefixes in the right order for the validator, leaving us with
only filetype=obj.

Also, I have a CL in progress (not yet uploaded) that modifies
pnacl-translate.py and driver_env.py to plumb in pnacl-sz.  Trying to get the
Subzero bugs and omissions solved before submitting it, but I can upload early
if it would help with the scons tests.

[jvoung (off chromium), 2015-03-19 16:40:36]

jvoung@chromium.org changed reviewers:
+ mcgrathr@chromium.org

[jvoung (off chromium), 2015-03-19 16:50:36]

On 2015/03/19 01:09:55, stichnot wrote:
> Nice!  lgtm
> 
> I guess it would be nice to get filetype=iasm working for this at some point. 
> Currently, filetype=asm doesn't completely work because llvm-mc doesn't get
the
> lock and 16-bit prefixes in the right order for the validator, leaving us with
> only filetype=obj.
> 
> Also, I have a CL in progress (not yet uploaded) that modifies
> pnacl-translate.py and driver_env.py to plumb in pnacl-sz.  Trying to get the
> Subzero bugs and omissions solved before submitting it, but I can upload early
> if it would help with the scons tests.

+ Roland, in case you know a trick to handle our "filetype=iasm" case.

filetype=iasm is where we use the integrated assembler to get the instruction
encodings into a buffer, but then we write out a .s file that looks like:

# Some two byte instruction
.byte N
.byte M
# A function call:
.byte 0xe8
.long Function63 - (. + 4)

So it uses an assembler to generate ELF, while still testing the instruction
encoder.

However, I don't know how to get a PCRel relocation for "call IMM" using those
.byte and .long.

With our filetype=obj integrated ELF writer, I could have the relocation refer
to the 0/null symbol, and have the type be pcrel directly.

[jvoung (off chromium), 2015-03-19 18:58:01]

Committed patchset #4 (id:60001) manually as
f644a4b3a36d707b32dd4b7855774a346a575e96 (presubmit successful).