significant clean up and optimizations:
- don't call bioq_disksort() on every command, the hardware will do that for
  us.
- remove all of the complicated bio deferral code.  bio's that can't be
  serviced immediately can just wait on the bioq.
- Only reserve one command object for doing control commands to the card.
  This simplifies a lot of code and significantly reduces the size of the
  command struct.
- Allocate commands out of a slab instead of embedding them into the softc.
- Call the command action method directly instead of having ips_get_free_cmd()
  call it indirectly.

MFC After: 1 week

Also some minor style cleanups.

Add two new arguments to bus_dma_tag_create(): lockfunc and lockfuncarg.
Lockfunc allows a driver to provide a function for managing its locking
semantics while using busdma.  At the moment, this is used for the
asynchronous busdma_swi and callback mechanism.  Two lockfunc implementations
are provided: busdma_lock_mutex() performs standard mutex operations on the
mutex that is specified from lockfuncarg.  dftl_lock() is a panic
implementation and is defaulted to when NULL, NULL are passed to
bus_dma_tag_create().  The only time that NULL, NULL should ever be used is
when the driver ensures that bus_dmamap_load() will not be deferred.
Drivers that do not provide their own locking can pass
busdma_lock_mutex,&Giant args in order to preserve the former behaviour.

sparc64 and powerpc do not provide real busdma_swi functions, so this is
largely a noop on those platforms.  The busdma_swi on is64 is not properly
locked yet, so warnings will be emitted on this platform when busdma
callback deferrals happen.

If anyone gets panics or warnings from dflt_lock() being called, please
let me know right away.

Reviewed by:	tmm, gibbs

Found by:       FlexeLint

series.  This driver was generously developed and released by David
Jeffreys and Adaptec.  I've updated it to work with 5.x and fixed a
few bugs.

MFC After:	1 week