Wait, wait... so the fix is to take a 2.6.18 driver which the maintainer has deliberately changed and slam it into 2.6.20/22? Seriously? Was this problem reported upstream? (Google suggests 'no'.)
Wait, wait... so the fix is to take a 2.6.18 driver which the maintainer has deliberately changed and slam it into 2.6.20/22? Seriously? Was this problem reported upstream? (Google suggests 'no'.)