RFC Starting to pull out the I/O operations

This is an untested WIP but as it was coming in discussion about device-tree as well I figured it will be useful to post the prototype code at this point.

There is lots left to do - in particular resource claim/release should be an ops method, and it's going to need people to update various embedded 8250 drivers when we finally go for a version of it.

On the bright side once we have an ops structure we have the basis for avoiding further mayhem in the arch code.

---

Alan Cox (5):
8250: extract out the "odd" port types
8250: evict the hub6 I/O operations
8250: register upio users
8250: introduce the idea of an operations struct for an 8250 port 8250: Make serial_dl_write a port operation

 drivers/char/mwave/mwavedd.c    |    2 
 drivers/tty/serial/8250.c       |  450 ++++++++++++++-------------------------
 drivers/tty/serial/8250.h       |   27 ++
 drivers/tty/serial/8250_au.c    |   90 ++++++++
 drivers/tty/serial/8250_dwapb.c |  120 ++++++++++
 drivers/tty/serial/8250_hub6.c  |   39 +++
 drivers/tty/serial/8250_pci.c   |    2 
 drivers/tty/serial/8250_pnp.c   |    2 
 drivers/tty/serial/8250_rm9k.c  |   95 ++++++++
 drivers/tty/serial/Kconfig      |    9 +
 drivers/tty/serial/Makefile     |    3 
 drivers/tty/serial/of_serial.c  |    2 
 drivers/tty/serial/serial_cs.c  |    2 
 include/linux/serial_8250.h     |   36 ++-
 include/linux/serial_core.h     |    5 
 15 files changed, 579 insertions(+), 305 deletions(-)
 create mode 100644 drivers/tty/serial/8250_au.c
 create mode 100644 drivers/tty/serial/8250_dwapb.c
 create mode 100644 drivers/tty/serial/8250_rm9k.c

Words without body bare winter barren trees much unspoken