[ARM] Fix broken sl82c105 DMA prevention
authorRussell King <rmk@dyn-67.arm.linux.org.uk>
Sat, 12 Nov 2005 16:57:29 +0000 (16:57 +0000)
committerRussell King <rmk+kernel@arm.linux.org.uk>
Sat, 12 Nov 2005 16:57:29 +0000 (16:57 +0000)
commit9648f552f9e08548a3979643b99f14c21c7d8f5b
tree56ad9baae4679e1fecdb9ba5a754a57b79332b56
parent6460177f42358a744e110cc583adbe8a36f6aa08
[ARM] Fix broken sl82c105 DMA prevention

We must _never_ _ever_ on pain of death enable IDE DMA on SL82C105
chipsets where the southbridge revision is <= 5, otherwise data
corruption will occur.

Strangely this used to work, but something has changed in the upper
echelons of the IDE layer to break the hosts decision to deny DMA.
Let's make it crystal clear to the IDE layer that we know best.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
drivers/ide/pci/sl82c105.c