Remove nonsfi token workaround.

Remove nonsfi token workaround.

Previously, file tokens were exposed to nonsfi plugins. Now that this is fixed
(and we have test coverage for a nonsfi packaged app), we can remove the
workaround that allowed nonzero token values.

BUG=394130
TEST=browser_tests on linux_rel_precise32

Committed: https://crrev.com/4923bbcd251b3cf3519bc1251e26f8243db5ea02
Cr-Commit-Position: refs/heads/master@{#309916}

[teravest, 2014-12-30 21:12:28]

Patchset #1 (id:1) has been deleted

[teravest, 2014-12-30 21:13:25]

teravest@chromium.org changed reviewers:
+ mseaborn@chromium.org

[Mark Seaborn, 2014-12-31 03:21:54]

LGTM, thanks

[teravest, 2015-01-05 16:29:08]

The CQ bit was checked by teravest@chromium.org

[commit-bot: I haz the power, 2015-01-05 16:30:26]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/807193006/20001

[commit-bot: I haz the power, 2015-01-05 17:17:29]

Committed patchset #1 (id:20001)

[commit-bot: I haz the power, 2015-01-05 17:18:24]

Patchset 1 (id:??) landed as
https://crrev.com/4923bbcd251b3cf3519bc1251e26f8243db5ea02
Cr-Commit-Position: refs/heads/master@{#309916}

[Shuhei Takahashi, 2015-01-07 08:13:29]

A revert of this CL (patchset #1 id:20001) has been created in
https://codereview.chromium.org/836273002/ by nya@chromium.org.

The reason for reverting is: This change caused huge ARC boot time regression on
nonsfi mode (~600ms -> 900ms).

See: go/arcperf
.

[Mark Seaborn, 2015-01-12 21:25:45]

Checking I understand the cause of the performance regression...

https://codereview.chromium.org/807193006/diff/20001/components/nacl/renderer...
File components/nacl/renderer/ppb_nacl_private_impl.cc (right):

https://codereview.chromium.org/807193006/diff/20001/components/nacl/renderer...
components/nacl/renderer/ppb_nacl_private_impl.cc:688: if
(load_manager->nonsfi())
So this caused a fallback to a slow path in the caller, DownloadNexe()...

https://codereview.chromium.org/807193006/diff/20001/components/nacl/renderer...
components/nacl/renderer/ppb_nacl_private_impl.cc:1240: // The fast path didn't
work, we'll fetch the file using URLLoader and write
So the reason for the performance regression was that this change causes Non-SFI
Mode to fall back to this code path, where open_resource() makes a copy of each
file?

Oops.  I should have realised that.  I actually didn't realise that we still
*have* a code path that makes a copy of each file on startup.

[teravest, 2015-01-12 22:00:53]

On 2015/01/12 21:25:45, Mark Seaborn wrote:
> Checking I understand the cause of the performance regression...
> 
>
https://codereview.chromium.org/807193006/diff/20001/components/nacl/renderer...
> File components/nacl/renderer/ppb_nacl_private_impl.cc (right):
> 
>
https://codereview.chromium.org/807193006/diff/20001/components/nacl/renderer...
> components/nacl/renderer/ppb_nacl_private_impl.cc:688: if
> (load_manager->nonsfi())
> So this caused a fallback to a slow path in the caller, DownloadNexe()...
> 
>
https://codereview.chromium.org/807193006/diff/20001/components/nacl/renderer...
> components/nacl/renderer/ppb_nacl_private_impl.cc:1240: // The fast path
didn't
> work, we'll fetch the file using URLLoader and write
> So the reason for the performance regression was that this change causes
Non-SFI
> Mode to fall back to this code path, where open_resource() makes a copy of
each
> file?

That's correct.

> 
> Oops.  I should have realised that.  I actually didn't realise that we still
> *have* a code path that makes a copy of each file on startup.