On Fri, Jan 06, 2012 at 03:49:27PM -0800, Mark Brown wrote:
On Fri, Jan 06, 2012 at 11:44:45PM +0000, Russell King - ARM Linux wrote:
On Fri, Jan 06, 2012 at 03:41:39PM -0800, Mark Brown wrote:
This only helps this specific device model bit of things, if anything is still actually holding a reference to the device and tries to use it after we tore away the resource underneath it we'll still explode (never mind the fact that we're backing this stuff up with some global pointers to other devices which may or may not actually be there...).
Err what? Explain showing the code where you think this is the case with the patch I proposed please.
These aren't new problems being introduced, they're preexisting problems which aren't fixed by this (hence why I say it "only helps with this specific device model side of things") but can come up in pretty much the same circumstances.
Anything which helps reduce the abuses of the driver model is a plus - it helps remove the possibility of kernel oopses.