make cryptohome check a magic file to discern between chrome-login and slim-login

make cryptohome check a magic file to discern between chrome-login and slim-login

Also, thanks to the magic of git, this CL also includes an edit to pull the chrome-login support binaries into the chromeos-chrome package.

[Will Drewry, 2009-12-10 15:15:10]

LGTM

[I can't wait to get cryptohome cleaned up :]

http://codereview.chromium.org/490001/diff/1/2
File src/platform/cryptohome/lib/common (right):

http://codereview.chromium.org/490001/diff/1/2#newcode22
src/platform/cryptohome/lib/common:22: CHROME_LOGIN_FILE=/tmp/doing-chrome-login
Another option is to check if chrome is already running:

if pkill -0 '^chrome$' &>/dev/null; then
  # doin' chrome login
fi

since no chrome will be running during a slim login until after mount.  Will a
machine with chrome-login ever support slim as well?

I'm assuming not, so it doesn't really matter how you notify cryptohome.

lgtm :)

[Chris Masone, 2009-12-10 16:56:34]

I added the thing to mount.  can you make sure I didn't botch it?

On 2009/12/10 15:15:10, Will Drewry wrote:
> LGTM
> 
> [I can't wait to get cryptohome cleaned up :]
> 
> http://codereview.chromium.org/490001/diff/1/2
> File src/platform/cryptohome/lib/common (right):
> 
> http://codereview.chromium.org/490001/diff/1/2#newcode22
> src/platform/cryptohome/lib/common:22:
CHROME_LOGIN_FILE=/tmp/doing-chrome-login
> Another option is to check if chrome is already running:
> 
> if pkill -0 '^chrome$' &>/dev/null; then
>   # doin' chrome login
> fi
> 
> since no chrome will be running during a slim login until after mount.  Will a
> machine with chrome-login ever support slim as well?
> 
> I'm assuming not, so it doesn't really matter how you notify cryptohome.
> 
> lgtm :)

[Will Drewry, 2009-12-10 17:03:13]

http://codereview.chromium.org/490001/diff/5001/6001
File src/platform/cryptohome/bin/mount (right):

http://codereview.chromium.org/490001/diff/5001/6001#newcode16
src/platform/cryptohome/bin/mount:16: /bin/mount -t tmpfs tmpfs
/home/$DEFAULT_USER
You'll need to do the same thing here.

Since second_chance is called after error_handler, you can set CHROME_LOGIN_FILE
and MOUNT_POINT up here, then just check them in second_chance - and use them if
they are there..

I'd say, just drop it in before the if statement, then have it and second_chance
use MOUNT_POINT (as you have below).

[Chris Masone, 2009-12-10 17:07:04]

http://codereview.chromium.org/490001/diff/5001/6001
File src/platform/cryptohome/bin/mount (right):

http://codereview.chromium.org/490001/diff/5001/6001#newcode16
src/platform/cryptohome/bin/mount:16: /bin/mount -t tmpfs tmpfs
/home/$DEFAULT_USER
On 2009/12/10 17:03:13, Will Drewry wrote:
> You'll need to do the same thing here.
> 
> Since second_chance is called after error_handler, you can set
CHROME_LOGIN_FILE
> and MOUNT_POINT up here, then just check them in second_chance - and use them
if
> they are there..
> 
> I'd say, just drop it in before the if statement, then have it and
second_chance
> use MOUNT_POINT (as you have below).

Done.

[Will Drewry, 2009-12-10 17:17:11]

http://codereview.chromium.org/490001/diff/5/6
File src/platform/cryptohome/bin/mount (right):

http://codereview.chromium.org/490001/diff/5/6#newcode17
src/platform/cryptohome/bin/mount:17: MOUNT_POINT=$MOUNT_POINT/user
Since we're depending on DEFAULT_USER here, we expect common to have been
loaded.  That means we'll have the mountpoint variable.  So this part can be
left off and the code below can use DEFAULT_MOUNT_POINT here and in second
chance.

[Chris Masone, 2009-12-10 17:20:35]

LGTY?

http://codereview.chromium.org/490001/diff/5/6
File src/platform/cryptohome/bin/mount (right):

http://codereview.chromium.org/490001/diff/5/6#newcode17
src/platform/cryptohome/bin/mount:17: MOUNT_POINT=$MOUNT_POINT/user
On 2009/12/10 17:17:11, Will Drewry wrote:
> Since we're depending on DEFAULT_USER here, we expect common to have been
> loaded.  That means we'll have the mountpoint variable.  So this part can be
> left off and the code below can use DEFAULT_MOUNT_POINT here and in second
> chance.

ah.  Well, that's much better.

[Will Drewry, 2009-12-10 19:14:31]

lgtm