Fix --filetype=iasm non-pc-rel fixup offsets (double counted).

Fix --filetype=iasm non-pc-rel fixup offsets (double counted).

For pc-rel fixups, we have a ConstantRelocatable referring
to Foo+0, and and the offset "-4" is encoded in the code
buffer (but not the ConstantRelocatable object). Thus we
need to load from the code buffer in order to
get that "-4" instead of just taking the +0 from Foo+0.

For non-pc-rel fixups, we have the ConstantRelocatable
with a true offset, and we also write that offset into the
code buffer (for ELF REL and not RELA, it expects the
offset in the code buffer). In this case, we want to choose
one and not double-count.

BUG=none
176.gcc seemed to be failing when compiled with --filetype=iasm...
load address for 64-bit pointers were +8 instead of +4

R=stichnot@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=b7db1a521f932c2b9d0b09694112ff29a51a212c

[jvoung (off chromium), 2015-07-20 23:26:00]

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

[jvoung (off chromium), 2015-07-20 23:29:05]

https://codereview.chromium.org/1241313002/diff/1/src/IceELFSection.h
File src/IceELFSection.h (right):

https://codereview.chromium.org/1241313002/diff/1/src/IceELFSection.h#newcode370
src/IceELFSection.h:370: Rela.r_addend = Fixup.offset();
This doesn't consult the buffer's contents, so we'll probably run into a bug if
the ConstantRelocatable object for a function call doesn't hold the true offset
(and only the buffer holds that offset), even if the offset is kind of silly
(e.g., the -4 you would get from pcrel call).

[Jim Stichnoth, 2015-07-21 13:14:51]

Wow, I thought this used to work fine.  Any idea how it broke?

LGTM

[jvoung (off chromium), 2015-07-21 15:09:38]

On 2015/07/21 13:14:51, stichnot wrote:
> Wow, I thought this used to work fine.  Any idea how it broke?
> 
> LGTM

I bisected the introduction of the add to
https://codereview.chromium.org/1027593002/diff/20001/src/IceFixups.cpp.  Prior
to that, it was using a hardcoded "FixupSize = 4" instead of loading from the
buffer, and only did that for PCRel, which was hiding the fact that sometimes
the fixup object's ConstantRelocatable's offset isn't accurate (the -4 from
pcrel).

[jvoung (off chromium), 2015-07-21 16:39:05]

Committed patchset #1 (id:1) manually as
b7db1a521f932c2b9d0b09694112ff29a51a212c (presubmit successful).