Introduce OS_NACL_SFI and OS_NACL_NONSFI macros.

Introduce OS_NACL_SFI and OS_NACL_NONSFI macros.

PNaCl toolchain starts to support "Non-SFI build binary" mode.
Some sources are built both for SFI binary and for Non-SFI binary, but, in some
cases, there are small diffs which need to be guarded by #ifdef.
In "Non-SFI build binary" mode, __native_client_nonsfi__ is defined.
This CL introduce OS_NACL_SFI and OS_NACL_NONSFI which can be used among the
chrome code, to slightly simplify the #ifdef conditions.

This CL is the follow up of crrev.com/659243002

BUG=358465
TEST=Ran trybot.

Committed: https://crrev.com/f7295f2e7b0c17a7784f23e06b924ab966fea3a8
Cr-Commit-Position: refs/heads/master@{#301603}

[hidehiko, 2014-10-27 09:43:36]

hidehiko@chromium.org changed reviewers:
+ dmichael@chromium.org, mseaborn@chromium.org, phajdan.jr@chromium.org,
uekawa@google.com

[hidehiko, 2014-10-27 09:43:37]

This CL is follow up for crrev.com/659243002.
Mark, Junichi, could you review from project points of view?
Dave, could you review as IPC owner of NaCl related code?
Pawel, could you review as PRESUBMIT.py owner?

Thank you for your review in advance,
- hidehiko

[Paweł Hajdan Jr., 2014-10-27 11:45:53]

PRESUBMIT LGTM

[Junichi Uekawa, 2014-10-27 12:13:56]

lgtm

thanks, the conditions look much more parse-able.

[Junichi Uekawa, 2014-10-27 12:14:08]

https://codereview.chromium.org/664373003/diff/1/PRESUBMIT.py
File PRESUBMIT.py (right):

https://codereview.chromium.org/664373003/diff/1/PRESUBMIT.py#newcode269
PRESUBMIT.py:269: 'OS_NACL_NONSFI',
sort ?

https://codereview.chromium.org/664373003/diff/1/build/build_config.h
File build/build_config.h (right):

https://codereview.chromium.org/664373003/diff/1/build/build_config.h#newcode23
build/build_config.h:23: // build is for Non-SFI binary or SFI binary.
This file is read by many people when they port things.
When a reader reads this document, I would expect to be able to distinguish what
they are.
Since it's something that we've defined as opposed to someone else, it might be
our duty to try to explain it succinctly.

ala "OS_NACL comes in two sandboxing technology flavors, SFI or Non-SFI. See XXX
for details."

[hidehiko, 2014-10-27 13:11:53]

Thank you for review. PTAL.

https://codereview.chromium.org/664373003/diff/1/PRESUBMIT.py
File PRESUBMIT.py (right):

https://codereview.chromium.org/664373003/diff/1/PRESUBMIT.py#newcode269
PRESUBMIT.py:269: 'OS_NACL_NONSFI',
On 2014/10/27 12:14:08, Junichi Uekawa wrote:
> sort ?

Done.

https://codereview.chromium.org/664373003/diff/1/build/build_config.h
File build/build_config.h (right):

https://codereview.chromium.org/664373003/diff/1/build/build_config.h#newcode23
build/build_config.h:23: // build is for Non-SFI binary or SFI binary.
On 2014/10/27 12:14:08, Junichi Uekawa wrote:
> This file is read by many people when they port things.
> When a reader reads this document, I would expect to be able to distinguish
what
> they are.
> Since it's something that we've defined as opposed to someone else, it might
be
> our duty to try to explain it succinctly.
> 
> ala "OS_NACL comes in two sandboxing technology flavors, SFI or Non-SFI. See
XXX
> for details."
> 

Indeed. Updated the comment.

[Mark Seaborn, 2014-10-27 15:53:59]

Can you change the other conditionals using __native_client_nonsfi__ too? 
(Either in this change or in a follow-up.)  I think Chromium style generally
prefers using the OS_* macros over "__foo__" macros.

LGTM

https://codereview.chromium.org/664373003/diff/20001/build/build_config.h
File build/build_config.h (right):

https://codereview.chromium.org/664373003/diff/20001/build/build_config.h#new...
build/build_config.h:24: //
https://developer.chrome.com/native-client/reference/sandbox_internals/x86-64...
This link isn't very relevant since it doesn't explain the difference between
SFI and Non-SFI.  I'd just leave it out (and the sentence before).

[dmichael (off chromium), 2014-10-27 15:57:28]

lgtm

[Junichi Uekawa, 2014-10-28 07:18:05]

lgtm

[hidehiko, 2014-10-28 08:07:21]

Thank you for review. Submitting...

Mark, I'm planning to send another follow-up CL to clean up
__native_client_nonsfi__, which is only in base/.
Note: we use it in nacl_helper_linux.cc, too, but it will be edited very soon,
so I'll leave it as is for now.

https://codereview.chromium.org/664373003/diff/20001/build/build_config.h
File build/build_config.h (right):

https://codereview.chromium.org/664373003/diff/20001/build/build_config.h#new...
build/build_config.h:24: //
https://developer.chrome.com/native-client/reference/sandbox_internals/x86-64...
On 2014/10/27 15:53:59, Mark Seaborn wrote:
> This link isn't very relevant since it doesn't explain the difference between
> SFI and Non-SFI.  I'd just leave it out (and the sentence before).

Done.

[hidehiko, 2014-10-28 08:07:25]

The CQ bit was checked by hidehiko@chromium.org

[commit-bot: I haz the power, 2014-10-28 08:08:03]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/664373003/60001

[commit-bot: I haz the power, 2014-10-28 11:57:29]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2014-10-28 11:58:10]

Patchset 4 (id:??) landed as
https://crrev.com/f7295f2e7b0c17a7784f23e06b924ab966fea3a8
Cr-Commit-Position: refs/heads/master@{#301603}