Partly change IRT load-skipping to be explicit.

Partly change IRT load-skipping to be explicit.

From sel_ldr_launcher_standalone, revert to having the IRT load
when -B is given.  This reverts part of:
https://codereview.chromium.org/10681005/

From chrome, accept irt_fd == -1 as a way to indicate that IRT
loading will be skipped. Wait for this to roll into chrome,
then this CL https://codereview.chromium.org/11761025/ will actually
set that up for PNaCl's translator nexes.  Once the chrome CL
is in, we can come back and delete the other case where IRT loads
can be skipped (when a segment gap is missing).

BUG= http://code.google.com/p/nativeclient/issues/detail?id=3241
TEST= manual (patch this into Chrome, along with chrome-side
CL and run PNaCl w/ TOOLS_REVISION at 10509).

Committed: https://src.chromium.org/viewvc/native_client?view=rev&revision=10533

[Mark Seaborn, 2013-01-03 23:36:40]

LGTM

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
File src/trusted/service_runtime/sel_main_chrome.c (right):

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
src/trusted/service_runtime/sel_main_chrome.c:311: * TODO(mseaborn): Instead of
looking for the absence of a segment gap,
You can change this to TODO(jvoung) if you're going to do it. :-)

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
src/trusted/service_runtime/sel_main_chrome.c:319: "Main executable has no
segment gap or irt_fd == -1; "
You should probably drop this log message, or change the log level to >=1.  We
don't want this to be printed when PNaCl is working fine.

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
File src/trusted/service_runtime/sel_main_chrome.h (right):

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
src/trusted/service_runtime/sel_main_chrome.h:37: * descriptors are emulated by
the C runtime library).  Optional; may be -1.
Nit: can you word-wrap to roughly the same length as the rest of the paragraph?

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
src/trusted/service_runtime/sel_main_chrome.h:38: * This is only optional for
non-ABI stable nexes.
"optional when loading nexes that don't follow NaCl's stable ABI, such as the
PNaCl translator" - just to be more explicit?

[jvoung (off chromium), 2013-01-04 00:05:31]

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
File src/trusted/service_runtime/sel_main_chrome.c (right):

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
src/trusted/service_runtime/sel_main_chrome.c:311: * TODO(mseaborn): Instead of
looking for the absence of a segment gap,
On 2013/01/03 23:36:40, Mark Seaborn wrote:
> You can change this to TODO(jvoung) if you're going to do it. :-)

Done.

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
src/trusted/service_runtime/sel_main_chrome.c:319: "Main executable has no
segment gap or irt_fd == -1; "
On 2013/01/03 23:36:40, Mark Seaborn wrote:
> You should probably drop this log message, or change the log level to >=1.  We
> don't want this to be printed when PNaCl is working fine.

Done.

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
File src/trusted/service_runtime/sel_main_chrome.h (right):

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
src/trusted/service_runtime/sel_main_chrome.h:37: * descriptors are emulated by
the C runtime library).  Optional; may be -1.
On 2013/01/03 23:36:40, Mark Seaborn wrote:
> Nit: can you word-wrap to roughly the same length as the rest of the
paragraph?

Done.

https://codereview.chromium.org/11743028/diff/1/src/trusted/service_runtime/s...
src/trusted/service_runtime/sel_main_chrome.h:38: * This is only optional for
non-ABI stable nexes.
On 2013/01/03 23:36:40, Mark Seaborn wrote:
> "optional when loading nexes that don't follow NaCl's stable ABI, such as the
> PNaCl translator" - just to be more explicit?

Done.