On Wed, Apr 13, 2011 at 9:44 AM, hch@xxxxxxxxxxxxx <hch@xxxxxxxxxxxxx> wrote: > On Tue, Apr 12, 2011 at 11:32:29PM -0700, Dan Williams wrote: >> Thanks. We can do this a bit simpler and just move ram initialization >> completely out of the start path, where it does not really belong. >> >> I'll fold the fix below into the patch. > > That looks much better than the initial version. I'm still not sure > there actually is much of a point of doing a lot of work in the > start state handler. Is there any fundamental reason for delaying work > from the normal init path? > Probably everything above "/* Start all of the ports on this controller */" can be moved to the normal init path. "Start" should just be "turn on the phys". However, we'll need to revisit this allocation code when adding power-management support (post 2.6.39) since the allocation is tied to the register re-initialization. Probably just add a devres group to handle all the dma host memory areas for the controller, rather than track the allocations individually. But the simple patch is sufficient for now. -- Dan -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
