Add an IPC for PNaCl to check if the session is incognito, before deciding

Add an IPC for PNaCl to check if the session is incognito,
before deciding to cache a translated nexe (current cache
mechanism won't work in incognito). Use this to decide
whether or not to cache translations for now.

For now, the linker always writes to a self-deleting temp
file, instead of to the cache directly.  We then copy the
data from the temp file to the cache as a separate step.

Took about 60ms to copy a 4.6MB nexe in debug mode with a
128kb buffer and 15ms with a 1mb buffer...


BUG= http://code.google.com/p/nativeclient/issues/detail?id=2683
TEST= try compiling an app that doesn't use the File system
in incognito mode -- we should add a test though...
we should add a test though

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=149749

[jvoung (off chromium), 2012-08-01 23:14:46]

http://codereview.chromium.org/10826107/diff/2001/ppapi/native_client/src/tru...
File ppapi/native_client/src/trusted/plugin/pnacl_translate_thread.h (right):

http://codereview.chromium.org/10826107/diff/2001/ppapi/native_client/src/tru...
ppapi/native_client/src/trusted/plugin/pnacl_translate_thread.h:45: TempFile*
nexe_file,
Or we can allow RunTranslate to take the cache file (or temp file) directly to
write to.

If we later use the browser cache instead of the html FS, I'm not sure writing
to the cache directly would be okay or not...

[Derek Schuff, 2012-08-02 16:33:28]

LGTM on the coordinator stuff

On 2012/08/01 23:14:46, jvoung (chromium) wrote:
>
http://codereview.chromium.org/10826107/diff/2001/ppapi/native_client/src/tru...
> File ppapi/native_client/src/trusted/plugin/pnacl_translate_thread.h (right):
> 
>
http://codereview.chromium.org/10826107/diff/2001/ppapi/native_client/src/tru...
> ppapi/native_client/src/trusted/plugin/pnacl_translate_thread.h:45: TempFile*
> nexe_file,
> Or we can allow RunTranslate to take the cache file (or temp file) directly to
> write to.
> 
> If we later use the browser cache instead of the html FS, I'm not sure writing
> to the cache directly would be okay or not...

[jvoung - send to chromium..., 2012-08-02 16:55:47]

+brettw, if you could take a look at the ppb_nacl interface stuff again --
thanks!

[brettw, 2012-08-02 19:50:47]

It looks like the incognito bit is available in-process in the
ChromeRenderProcessObserver (in chrome/renderer). And that's a static function.
Can you see if this will work instead of the sync message?

http://codereview.chromium.org/10826107/diff/2003/chrome/common/render_messag...
File chrome/common/render_messages.h (right):

http://codereview.chromium.org/10826107/diff/2003/chrome/common/render_messag...
chrome/common/render_messages.h:596: 
Got extra blank line.

http://codereview.chromium.org/10826107/diff/2003/chrome/renderer/pepper/ppb_...
File chrome/renderer/pepper/ppb_nacl_private_impl.cc (right):

http://codereview.chromium.org/10826107/diff/2003/chrome/renderer/pepper/ppb_...
chrome/renderer/pepper/ppb_nacl_private_impl.cc:314: if (is_off_the_record) {
This should just be:
  return PP_FromBool(is_off_the_record);

[jvoung (off chromium), 2012-08-02 22:26:07]

Yup the renderer bit works -- simplified.  Thanks!

Also had to fix an error path in pnacl_coordinator to make the error-checking
test work.

http://codereview.chromium.org/10826107/diff/2003/chrome/common/render_messag...
File chrome/common/render_messages.h (right):

http://codereview.chromium.org/10826107/diff/2003/chrome/common/render_messag...
chrome/common/render_messages.h:596: 
On 2012/08/02 19:50:47, brettw wrote:
> Got extra blank line.

Removed.

http://codereview.chromium.org/10826107/diff/2003/chrome/renderer/pepper/ppb_...
File chrome/renderer/pepper/ppb_nacl_private_impl.cc (right):

http://codereview.chromium.org/10826107/diff/2003/chrome/renderer/pepper/ppb_...
chrome/renderer/pepper/ppb_nacl_private_impl.cc:314: if (is_off_the_record) {
On 2012/08/02 19:50:47, brettw wrote:
> This should just be:
>   return PP_FromBool(is_off_the_record);

Done.  Also used the renderer bit.

[sehr (please use chromium), 2012-08-02 22:37:48]

Some nits.

https://chromiumcodereview.appspot.com/10826107/diff/4001/chrome/common/rende...
File chrome/common/render_messages.h (right):

https://chromiumcodereview.appspot.com/10826107/diff/4001/chrome/common/rende...
chrome/common/render_messages.h:593: // Send by the renderer to check if the
current tab is off the record.
s/Send/Sent/

https://chromiumcodereview.appspot.com/10826107/diff/4001/ppapi/native_client...
File ppapi/native_client/src/trusted/plugin/pnacl_coordinator.cc (right):

https://chromiumcodereview.appspot.com/10826107/diff/4001/ppapi/native_client...
ppapi/native_client/src/trusted/plugin/pnacl_coordinator.cc:27: const uint32_t
kCopyBufSize = 131072;
I usually prefer something like 128 << 10 for these constants.

https://chromiumcodereview.appspot.com/10826107/diff/4001/ppapi/native_client...
File ppapi/native_client/src/trusted/plugin/temporary_file.cc (right):

https://chromiumcodereview.appspot.com/10826107/diff/4001/ppapi/native_client...
ppapi/native_client/src/trusted/plugin/temporary_file.cc:74: // only plain host
descs.  It's probably good hygene to separate the
hygiene

[jvoung - send to chromium..., 2012-08-02 22:44:29]

https://chromiumcodereview.appspot.com/10826107/diff/4001/chrome/common/rende...
File chrome/common/render_messages.h (right):

https://chromiumcodereview.appspot.com/10826107/diff/4001/chrome/common/rende...
chrome/common/render_messages.h:593: // Send by the renderer to check if the
current tab is off the record.
On 2012/08/02 22:37:49, sehr wrote:
> s/Send/Sent/

Done. (code removed)

https://chromiumcodereview.appspot.com/10826107/diff/4001/ppapi/native_client...
File ppapi/native_client/src/trusted/plugin/pnacl_coordinator.cc (right):

https://chromiumcodereview.appspot.com/10826107/diff/4001/ppapi/native_client...
ppapi/native_client/src/trusted/plugin/pnacl_coordinator.cc:27: const uint32_t
kCopyBufSize = 131072;
On 2012/08/02 22:37:49, sehr wrote:
> I usually prefer something like 128 << 10 for these constants.

Done.

https://chromiumcodereview.appspot.com/10826107/diff/4001/ppapi/native_client...
File ppapi/native_client/src/trusted/plugin/temporary_file.cc (right):

https://chromiumcodereview.appspot.com/10826107/diff/4001/ppapi/native_client...
ppapi/native_client/src/trusted/plugin/temporary_file.cc:74: // only plain host
descs.  It's probably good hygene to separate the
On 2012/08/02 22:37:49, sehr wrote:
> hygiene

Done.

[brettw, 2012-08-02 23:52:18]

lgtm