[CFT] PCMCIA patches

[CFT] PCMCIA patches

[Russell King]

Ok guys,

Here's another set of PCMCIA patches to keep people occupied for a while.
Tested sa11xx and yenta here.  Please report successes/failures.

Note: if you are using modules, you will only be able to remove the
socket driver when the cards are ejected (by either physically removing
the card or via cardctl eject.)  Some init scripts may get upset with
this on shutdown; this will eventually be noted in davej's 2.6 changes
document.  The script needs to run cardctl eject before rmmoding the
pcmcia modules.

 drivers/pcmcia/cs.c           |  103 ++++++++++++++++++++++++------------------
 drivers/pcmcia/i82092.c       |   46 ++----------------
 drivers/pcmcia/i82092aa.h     |    1
 drivers/pcmcia/i82365.c       |   75 +++---------------------------
 drivers/pcmcia/sa11xx_core.c  |   66 +++++---------------------
 drivers/pcmcia/sa11xx_core.h  |    3 -
 drivers/pcmcia/tcic.c         |   48 ++-----------------
 drivers/pcmcia/yenta_socket.c |   39 +--------------
 drivers/pcmcia/yenta_socket.h |    5 --
 include/pcmcia/ss.h           |    3 -
 10 files changed, 96 insertions(+), 293 deletions(-)

http://patches.arm.linux.org.uk/pcmcia/pcmcia-event-20030623-1.diff

	Move ->owner field from socket operations to pcmcia_socket.
	(This change is mainly for the SA11xx drivers, which use
	a core driver for the chip, and a separate module for all
	the machine specific bits.)

http://patches.arm.linux.org.uk/pcmcia/pcmcia-event-20030623-2.diff

	Get/Put module when we insert and remove a card.  This avoids
	a potential deadlock when socket drivers are unloaded, and we
	have a cardbus card known to the system.

http://patches.arm.linux.org.uk/pcmcia/pcmcia-event-20030623-3.diff

	Remove original module use accounting in register_callback.

http://patches.arm.linux.org.uk/pcmcia/pcmcia-event-20030623-4.diff

	Add work-around for i82365-based socket drivers to the core
	PCMCIA code.  Since insert processing is not a time critical
	event, we can afford to delay (by sleeping) these for everyone.

http://patches.arm.linux.org.uk/pcmcia/pcmcia-event-20030623-5.diff

	Remove register_callback methods; allow socket drivers to call
	pcmcia_parse_events() directly.

http://patches.arm.linux.org.uk/pcmcia/pcmcia-event-20030623-6.diff

	Remove now obsolete work queues, spinlocks, and code from
	socket drivers.


-- 
Russell King (rmk@arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html

[CFT] PCMCIA patches

[Russell King]

Hi,

I'm about to send a set of 6 PCMCIA patches, which I'd like people to
test.  They're against base current 2.5.64 BK (as of this morning GMT,)
but they apply with some offset to plain .64.

The patch numbering starts at 2 - the first set (which was 1a to 1h)
was ARM stuff, which Linus has already accepted.

pcmcia-2.diff

  get_io_map and get_mem_map PCMCIA socket methods are never called
  by the PCMCIA core code.  They are therefore dead code, and can be
  removed.

pcmcia-3.diff

  Remove the bus_* abstractions.

pcmcia-4.diff

  Cardbus uses socket->cb_config to detect when the cardbus card has
  been initialised.  Since cb_config will eventually die, we need a
  solution - introduce the SOCKET_CARDBUS_CONFIG flag, which is set
  once we have initialised the cardbus socket.

pcmcia-5.diff

  Add an element of locking to the resource manager - don't allow
  the PCMCIA resource lists to be changed while the pcmcia code is
  scanning them.

pcmcia-6.diff

  Remove the dependence of the PCMCIA layer on CONFIG_ISA - introduce
  CONFIG_PCMCIA_PROBE to determine whether we need the resource
  handling code.  This prevents oopsen on SA11x0 and similar platforms
  which use statically mapped, non-windowed sockets.

pcmcia-7.diff

  Remove support for the old PCMCIA cardbus clients - all cardbus
  drivers should be converted to be full-class PCI citizens.


-- 
Russell King (rmk@arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html

Re: [CFT] PCMCIA patches

[Paul Mackerras]

Russell King writes:

> I'm about to send a set of 6 PCMCIA patches, which I'd like people to
> test.  They're against base current 2.5.64 BK (as of this morning GMT,)
> but they apply with some offset to plain .64.

Just tried them on a G3 powerbook, and it all seems fine.  I plugged
in a CF card and a wireless ethernet card (both 16-bit pcmcia) and
they both worked.  I don't have any cardbus cards to try.

Paul.