icside.c:
- Use C99 structure initialisers.
- Use generic DMA API.
- Use new hwif->hwif_data rather than hwif->hw.priv.
- Add DMA supported/dma supported masks.
- Remember the card type for the remove/shutdown methods.
- Add shutdown method.
rapide.c
- ecard_{claim,release} are no longer required.

- Fix build warnings.
- Don't modify region in copyarea method.
- Remove FBCON_HAS_CFBx preprocessor tests.
- Use runtime test for netwinder mclk parameters.

Move Acorn and ARM SCSI and net drivers to drivers/{net,scsi}/arm.
This also allows us to find a reasonable home for other ARM specific
net drivers.

These were broken by two changes - the removal of the old device model
class code, and when scsi device lists appeared.  This cset allows
these drivers to build again.  We also drop some unnecessary code
from one of the drivers.

These drivers got missed when the i2c subsystem was converted to the
device model.

This avoids a problem with the original ptrace code using a system
call (SWI) to implement single stepping; programs such as ltrace
do not expect to receive system call trace traps when breakpoints
are hit.

Request the memory region used for the keyboard and mouse ports,
and ioremap.

Also remove uninitialised variable warning and update mach-types.

Patch from Nicolas Pitre

From: Nicolas Pitre.

(manual entry since bk openlogging crapped out again)

Patch from Nicolas Pitre

Of course, the PLD macro is always defined even if it's empty.
Without this fix anything below ARMv5 is broken.

into flint.arm.linux.org.uk:/usr/src/linux-bk-2.5/linux-2.5-rmk

This patch adds dynamic allocation of request structures. Right now we
are reserving 256 requests per initialized queue, which adds up to quite
a lot of memory for even a modest number of queues. For the quoted 4000
disk systems, it's a disaster.

Instead, we mempool 4 requests per queue and put an upper limit on the
number of requests that we will put in-flight as well. I've kept the 128
read/write max in-flight limit for now. It is trivial to experiement
with larger queue sizes now, but I want to change one thing at the time
(the truncate scenario doesn't look all that good with a huge number of
requests, for instance).

Patch has been in -mm for a while, I'm running it here against stock 2.5
as well. Additionally, it actually kills quite a bit of code as well

Limit deadline hash to 32 entries instead of 1024. This has been benched
and profiled extensively and shows no increased system time.

Also, move "hot" entries to the front of the list.

Resend of the ide-cd buggy debug check removal.

From der.eremit@email.de.

increment bi_sector in bio_endio() so make_request_fn drivers don't
have to do this accounting themselves.

into home.transmeta.com:/home/torvalds/v2.5/linux

into home.transmeta.com:/home/torvalds/v2.5/linux

into home.transmeta.com:/home/torvalds/v2.5/linux

into kroah.com:/home/greg/linux/BK/gadget-2.5

into kroah.com:/home/greg/linux/BK/class-2.5

into kroah.com:/home/greg/linux/BK/i2c-2.5

send_sig_info() has been broken since 2.5.60.

The function can be invoked from a the time interrupt (timer_interrpt ->
do_timer -> update_process_times -> -> update_one_process -> (
do_process_times, do_it_prof, do_it_virt ) -> -> send_sig ->
send_sig_info) but it uses spin_unlock_irq instead of the correct
spin_unlock_irqrestore. 

This enables interrupts, and later scheduler_tick() locks runqueue
(without disabling interrupts).  And if we are unlucky, a new interrupt
comes at this point.  And if this interrupt tries to do wake_up() (like
RTC interrupt does), we will deadlock on runqueue lock :-(

The bug was introduced by signal-fixes-2.5.59-A4, which split the
original send_sig_info into two functions, and in one branch it started
using these unsafe spinlock variants (while the "group" variant uses
irqsave/restore correctly). 

Thanks to Jonathan Corbet for pointing this out.

Thanks to Mike Anderson for the idea for this.