fix-glamo-crank-memory-to-90MHz.patch blocks resuming from suspend

[Andy Green]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Somebody in the thread at some point said:
| Hello, in order to try to fix my huge issues with glamo [1], I've tried
| to apply the latest andy kernel patches for glamo itself:
|
|  - fix-glamo-turbo-host-interface.patch [2]
|  - fix-glamo-crank-memory-to-90MHz.patch [3]
|
| Well, after applying them I was finally able to use my Xglamo server
| without any freeze, but I wasn't more able to suspend and resume (phone
| go to sleep but it never woke-up [4]).
|
| That's why I've tried to unapply the turbo-host-interface patch [2] and
| fortunately I got all I needed: phone was able to suspend and resume and
| Xglamo ran smoothly.
|
| So, I'd like to thank andy for the glamo-memory patch that seems to fix
| all my (and of few other users) glamo issues and warn the list that the
| turbo-host patch seems to compromise the resuming of the phone (also if
| it seems to give a fps speedup).
| Andy, do you have another fix that is not disabling that patch?

Hm it's tough to know if we change timing behaviour in suspend / resume
and change race outcomes, or if the different results we see are driven
by the change we made directly.  It's easy to get fooled by apparently
direct causality when it's really indirect.

What I would do if I had ten minutes to rub together is write glamo
register at +0x200 to the pre-patch value of 0xef0 in suspend function
for it and write it back to 0xe03 at end of resume, see if that gets
around it.

- -Andy
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iEYEARECAAYFAki9dx4ACgkQOjLpvpq7dMo3QQCdFhd1cffOdc1032domh3YgcXi
ryQAn35uhoyYiVPYUqqIyJ8X+pU6hLMD
=tGec
-----END PGP SIGNATURE-----