Consider PlatformThread::Join() to be blocking IO.

Consider PlatformThread::Join() to be blocking IO.

Marks PlatformThread::Join() as blocking IO using ThreadRestrictions.
Whitelists existing spots where we join on the UI/IO threads.

Also noteworthy is I allow blocking IO on shutdown.

BUG=65530, 66077, 66082
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=68893

[Evan Martin, 2010-12-10 21:43:04]

LGTM, but please add comments :)

http://codereview.chromium.org/5693003/diff/2001/base/platform_thread_posix.cc
File base/platform_thread_posix.cc (right):

http://codereview.chromium.org/5693003/diff/2001/base/platform_thread_posix.c...
base/platform_thread_posix.cc:222: base::ThreadRestrictions::AssertIOAllowed();
Perhaps stick a comment in here about how this is intended to protect joining a
non-IO thread to an IO one.

http://codereview.chromium.org/5693003/diff/2001/chrome/browser/browser_proce...
File chrome/browser/browser_process_impl.cc (right):

http://codereview.chromium.org/5693003/diff/2001/chrome/browser/browser_proce...
chrome/browser/browser_process_impl.cc:255: // of it on shutdown for valid
reasons.
Heh, I hope this is the right place for this code...

http://codereview.chromium.org/5693003/diff/2001/chrome/browser/geolocation/g...
File chrome/browser/geolocation/geolocation_provider.cc (right):

http://codereview.chromium.org/5693003/diff/2001/chrome/browser/geolocation/g...
chrome/browser/geolocation/geolocation_provider.cc:51: // http://crbug.com/66077
Can you include a short summary here?  (So people don't need to click this link
every time they look at this code...)

http://codereview.chromium.org/5693003/diff/2001/chrome/browser/printing/prin...
File chrome/browser/printing/printer_query.cc (right):

http://codereview.chromium.org/5693003/diff/2001/chrome/browser/printing/prin...
chrome/browser/printing/printer_query.cc:100: // http://crbug.com/66082
Also include a note here about the bug.  (E.g. it's nice to be able to know
whether this reference is here to note that the next line is here for a good
reason, or whether we intend to remove it...)

[willchan no longer on Chromium, 2010-12-10 21:54:28]

I've added said comments.

http://codereview.chromium.org/5693003/diff/2001/chrome/browser/browser_proce...
File chrome/browser/browser_process_impl.cc (right):

http://codereview.chromium.org/5693003/diff/2001/chrome/browser/browser_proce...
chrome/browser/browser_process_impl.cc:255: // of it on shutdown for valid
reasons.
On 2010/12/10 21:43:04, Evan Martin wrote:
> Heh, I hope this is the right place for this code...

What, you don't trust me? :)  Try it for yourself if you'd like.  But yeah, from
my extensive experience trawling through the morass of shutdown code nonsense,
this is the right place to do it.  It's interesting to note that APP_TERMINATING
is not the right signal to watch, because the automation provider can trigger
shutdown in ways that bypass APP_TERMINATING.  But this is the way we stop the
UI thread, so it's guaranteed to always work.

[willchan no longer on Chromium, 2010-12-13 21:00:48]

Turns out I shouldn't have been trusted and this broke Windows.  I'm going to
reland after I figure out what Windows is doing.

On 2010/12/10 21:54:28, willchan wrote:
> I've added said comments.
> 
>
http://codereview.chromium.org/5693003/diff/2001/chrome/browser/browser_proce...
> File chrome/browser/browser_process_impl.cc (right):
> 
>
http://codereview.chromium.org/5693003/diff/2001/chrome/browser/browser_proce...
> chrome/browser/browser_process_impl.cc:255: // of it on shutdown for valid
> reasons.
> On 2010/12/10 21:43:04, Evan Martin wrote:
> > Heh, I hope this is the right place for this code...
> 
> What, you don't trust me? :)  Try it for yourself if you'd like.  But yeah,
from
> my extensive experience trawling through the morass of shutdown code nonsense,
> this is the right place to do it.  It's interesting to note that
APP_TERMINATING
> is not the right signal to watch, because the automation provider can trigger
> shutdown in ways that bypass APP_TERMINATING.  But this is the way we stop the
> UI thread, so it's guaranteed to always work.