Rearrange emit vs emitIAS. Wait till function is done before dumping text.

Rearrange emit vs emitIAS. Wait till function is done before dumping text.

Eventually, I wanted to have a flag "UseELFWriter" like:
https://codereview.chromium.org/678533005/diff/120001/src/IceCfg.cpp

Where the emit OStream would not have text, and only have
binary. This refactor hopefully means fewer places to
check for a flag to disable the text version of IAS,
and be able to write binary. Otherwise, there are some
text labels for branches that are still being dumped out.

BUG=none
R=stichnot@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=0faec4c9e067b974ec1d3f65c71af3abaca143aa

[jvoung (off chromium), 2014-11-05 23:06:32]

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

[Jim Stichnoth, 2014-11-06 00:02:13]

https://codereview.chromium.org/700263003/diff/1/src/IceInstX8632.cpp
File src/IceInstX8632.cpp (left):

https://codereview.chromium.org/700263003/diff/1/src/IceInstX8632.cpp#oldcode345
src/IceInstX8632.cpp:345: namespace {
Sweet!

https://codereview.chromium.org/700263003/diff/1/src/assembler.cpp
File src/assembler.cpp (right):

https://codereview.chromium.org/700263003/diff/1/src/assembler.cpp#newcode144
src/assembler.cpp:144: AssemblerFixup *LastFixup =
buffer_.GetLatestFixup(StartPosition);
The total work done in all the calls to GetLatestFixup() will be O(N^2) in the
number of fixups.  Can you maybe change the loop structure to loop forward over
the fixups, and emit the non-fixup bytes in between?

https://codereview.chromium.org/700263003/diff/1/tests_lit/llvm2ice_tests/ias...
File tests_lit/llvm2ice_tests/ias-multi-reloc.ll (left):

https://codereview.chromium.org/700263003/diff/1/tests_lit/llvm2ice_tests/ias...
tests_lit/llvm2ice_tests/ias-multi-reloc.ll:51: %dummy.bc = bitcast void (i32)*
@dummy to void ()*
FWIW, this construct, plus the curious addition of the i32 arg to @dummy(), was
the direct result of the build_pnacl_ir.py script.

[jvoung (off chromium), 2014-11-06 00:47:13]

https://codereview.chromium.org/700263003/diff/1/src/assembler.cpp
File src/assembler.cpp (right):

https://codereview.chromium.org/700263003/diff/1/src/assembler.cpp#newcode144
src/assembler.cpp:144: AssemblerFixup *LastFixup =
buffer_.GetLatestFixup(StartPosition);
On 2014/11/06 00:02:13, stichnot wrote:
> The total work done in all the calls to GetLatestFixup() will be O(N^2) in the
> number of fixups.  Can you maybe change the loop structure to loop forward
over
> the fixups, and emit the non-fixup bytes in between?

Ah right, I wasn't looking at how GetLatestFixup() is done too closely. Changed
to iterate over fixups and splat bytes as it iterates.

https://codereview.chromium.org/700263003/diff/1/tests_lit/llvm2ice_tests/ias...
File tests_lit/llvm2ice_tests/ias-multi-reloc.ll (left):

https://codereview.chromium.org/700263003/diff/1/tests_lit/llvm2ice_tests/ias...
tests_lit/llvm2ice_tests/ias-multi-reloc.ll:51: %dummy.bc = bitcast void (i32)*
@dummy to void ()*
On 2014/11/06 00:02:13, stichnot wrote:
> FWIW, this construct, plus the curious addition of the i32 arg to @dummy(),
was
> the direct result of the build_pnacl_ir.py script.

Hmm I wonder if it was a C oddness, where dummy() is different from dummy(void).

[Jim Stichnoth, 2014-11-06 01:18:14]

lgtm

https://codereview.chromium.org/700263003/diff/1/tests_lit/llvm2ice_tests/ias...
File tests_lit/llvm2ice_tests/ias-multi-reloc.ll (left):

https://codereview.chromium.org/700263003/diff/1/tests_lit/llvm2ice_tests/ias...
tests_lit/llvm2ice_tests/ias-multi-reloc.ll:51: %dummy.bc = bitcast void (i32)*
@dummy to void ()*
On 2014/11/06 00:47:13, jvoung wrote:
> On 2014/11/06 00:02:13, stichnot wrote:
> > FWIW, this construct, plus the curious addition of the i32 arg to @dummy(),
> was
> > the direct result of the build_pnacl_ir.py script.
> 
> Hmm I wonder if it was a C oddness, where dummy() is different from
dummy(void).

Ah, good point, that's probably it.

[jvoung (off chromium), 2014-11-06 01:30:01]

Committed patchset #3 (id:40001) manually as
0faec4c9e067b974ec1d3f65c71af3abaca143aa (presubmit successful).