Re: msk(4) require to sync status buffer

[David Young <dyoung%pobox.com@localhost>]

On Tue, Feb 26, 2008 at 07:56:36AM +0900, Izumi Tsutsui wrote:
> Furthermore, on some architecbure invalidate can't be done
> without writeback. Even if your architecbure supports
> invalidate without writeback, you can't use invalidate
> (and have to use invalidate and writeback) if buffer isn't
> cacheline aligned. In that case, if you defer PREREAD calls
> (which might do writeback and invalidate), you might lost
> some DMA'ed data by CPU writeback.

IMO, the bus back-end should KASSERT or panic when a driver may lose data
by calling bus_dmamap_sync() on a region that is either non-cacheline
aligned, or smaller than a cacheline.  We will soon detect and squash
the bugs as broken drivers panic the kernel.

Dave

-- 
David Young             OJC Technologies
dyoung%ojctech.com@localhost      Urbana, IL * (217) 278-3933 ext 24