Implement nacl_irt_filename for non-sfi mode.

Implement nacl_irt_filename for non-sfi mode.

This CL implements nacl_irt_filename for non-sfi mode.
Note that, in order to open a file, currently we need to set --no-sandbox
flag, otherwise EPERM will be returned.

BUG=https://code.google.com/p/nativeclient/issues/detail?id=3734
TEST=Tried to call a newly added function from plugin via AT_SYSINFO.

[hidehiko, 2014-01-09 14:13:06]

Could you take a look?

This CL is based on https://codereview.chromium.org/131133002, which is also
based on https://codereview.chromium.org/100373005/.
If you think code reviews for those CLs will affect to the this CL a lot, please
feel free to suspend this review.

Note: trybot looks somehow unstable. cros_daisy and linux_precise32 should run
later.

Thank you for your review in advance,
- hidehiko

[Mark Seaborn, 2014-01-10 13:41:09]

https://codereview.chromium.org/130753003/diff/1/components/nacl/loader/nonsf...
File components/nacl/loader/nonsfi/irt_interfaces.cc (right):

https://codereview.chromium.org/130753003/diff/1/components/nacl/loader/nonsf...
components/nacl/loader/nonsfi/irt_interfaces.cc:27:
NACL_INTERFACE_TABLE(NACL_IRT_FILENAME_v0_1, kIrtFileName),
This should be IRT_DEV_FILENAME.  IRT_FILENAME is not supported in the browser,
and we only have it queryable in existing ABIs for compatibility.  New ABIs
should not return it -- PNaCl doesn't.

So, if you want to add this, it should be behind a flag (as we discussed in
e-mail before).  But putting this behind a flag adds enough complexity that I
would want to have a test for that.  So can we put this change on hold until
PPAPI works under Bare Metal Mode and we have enabled tests for it on the bots?

[hidehiko, 2014-01-15 04:00:21]

https://codereview.chromium.org/130753003/diff/1/components/nacl/loader/nonsf...
File components/nacl/loader/nonsfi/irt_interfaces.cc (right):

https://codereview.chromium.org/130753003/diff/1/components/nacl/loader/nonsf...
components/nacl/loader/nonsfi/irt_interfaces.cc:27:
NACL_INTERFACE_TABLE(NACL_IRT_FILENAME_v0_1, kIrtFileName),
On 2014/01/10 13:41:09, Mark Seaborn wrote:
> This should be IRT_DEV_FILENAME.  IRT_FILENAME is not supported in the
browser,
> and we only have it queryable in existing ABIs for compatibility.  New ABIs
> should not return it -- PNaCl doesn't.
> 
> So, if you want to add this, it should be behind a flag (as we discussed in
> e-mail before).  But putting this behind a flag adds enough complexity that I
> would want to have a test for that.  So can we put this change on hold until
> PPAPI works under Bare Metal Mode and we have enabled tests for it on the
bots?

I see. I misunderstood. Let's keep this suspended for a while.

[hidehiko, 2014-10-29 09:34:38]

On 2014/01/15 04:00:21, hidehiko wrote:
>
https://codereview.chromium.org/130753003/diff/1/components/nacl/loader/nonsf...
> File components/nacl/loader/nonsfi/irt_interfaces.cc (right):
> 
>
https://codereview.chromium.org/130753003/diff/1/components/nacl/loader/nonsf...
> components/nacl/loader/nonsfi/irt_interfaces.cc:27:
> NACL_INTERFACE_TABLE(NACL_IRT_FILENAME_v0_1, kIrtFileName),
> On 2014/01/10 13:41:09, Mark Seaborn wrote:
> > This should be IRT_DEV_FILENAME.  IRT_FILENAME is not supported in the
> browser,
> > and we only have it queryable in existing ABIs for compatibility.  New ABIs
> > should not return it -- PNaCl doesn't.
> > 
> > So, if you want to add this, it should be behind a flag (as we discussed in
> > e-mail before).  But putting this behind a flag adds enough complexity that
I
> > would want to have a test for that.  So can we put this change on hold until
> > PPAPI works under Bare Metal Mode and we have enabled tests for it on the
> bots?
> 
> I see. I misunderstood. Let's keep this suspended for a while.

I think it's ok to close this CL.