This changes PCI resource allocation algorithm to 3 passes vs.
current 2 passes. Extra pass is used for calculation of required
size and alignment of PCI buses behind PCI-PCI bridges. After
that, in the pass #3, these buses get allocated like regular
PCI devices. This gives tighter PCI IO and memory packing -
for instance, this fixes allocation problems on certain alphas
with very small (112Mb) PCI memory range. Also, the new code
- will allow mixed approach to resource allocation:
  architecture can keep BIOS settings for some devices,
  and re-allocate resources for others, including improperly
  initialized bridges;
- makes prefetchable ranges support much simpler;
- allows sizing of IO and memory ranges for the host
  bridges, which might be very useful in some situations.

It was tested on various alphas; I haven't heard any complaints
from rmk and rth, so probably all of this is ok. :-)

Part 1:
- for all archs, 4th argument (align) added to
  pcibios_align_resource (and its callers).
  It's necessary because this function will be called for
  bus resources as well, and in this case size != alignment.
- for several archs, dead/bogus code removed from
  pcibios_fixup_pbus_ranges().

This allows user to select 'default' encoding, which has defined mapping
for each of 255 byte values, so one can use it as a fallback when it
finds filesystem with unknown encoding (EBCDIC for example) just to get
characters through filesystem.

				Petr Vandrovec

During my lurking around NLS code I found that KOI8-RU returns character
code instead of length of character for two characters.

					Petr Vandrovec

- kdev_t fixes
- bio fixes
- other cleanups and compile fixes

- small compile fixes
- id update
- add asm-ppc/zorro.h

extract show_trace()/show_trace_task() out of dump_stack().

Move several tlb functions into tlbflush.h.

update some proc functions to use the seq_file interface

Update INIT_RLIMITS.

These function were removed by a "cleanup", but they are still used.

Rename p_pptr into parent.

Compile fixes

This patch renames KTHREAD_SIZE into THREAD_SIZE, similiar to all other archs.

Add missing include files.

Fix idle functions.

- replace hwclk_time with rtc_time
- use hwclk instead of gettod to set initial time

compile fix

Add missing license tags.

Update to the latest config.in.

This moves several cache flush function into cacheflush.h.

Add __clear_bit/__ffs/sched_find_first_bit.

Some compile fixes.

- sizeof changes from Kernel Janitor Project
- several bug fixes found with fsx

Here is one fix and one enhancement to the scsi generic (sg)
driver in 2.5.14
Changelog:
    - off by one fix for last scatter gather element
    - if possible compact kiobuf_map into scatter gather list

This is the OnStream variant of the st driver that
has been in Dave's tree for some time.

The change looks like it is from Willem Riede (osst@riede.org)
and that it has been forwarded ported from lk 2.4 .

This ISA scsi adapter still seems to be popular
given the number of people that supply patches
for it collected in Dave's tree.

The latest patch was from William Stinson (first item):
  - request_region cleanup
  - removal of old scsi error handling (leaving newer version)
  - scatterlist::address replaced
  - host_lock replacing io_request_lock


Doug Gilbert

P.S. This is a transfer from Dave's tree. It runs fine for
me on a dual Celeron SMP box.

after the manual merge of the PCI re-organization
removed it.

The largest part of the OSS update from 2.4
Various changelog entries scattered throughout the patch.

From 2.4
- Lindented
- Add Docbook comments
- Remove duplicate MODULE_LICENSE tag

The PNPBIOS driver and the floppy driver both fight over reservation
of the floppy port. This patch makes things friendly again.
As the comment suggests, this fix could be done better, but until
someone steps forward to fix this, this is the best we have.

copy_to_user already does the relevant checking.

Done by William Stinson.
Adds error handling to request_region() calls,
and converts some old check_region() calls too.

This patch from 2.4 makes sure we don't prefetch past the
end of a range to be copied (in case its at the end of a memrange)
i386 case looks safe already, we just weren't optimal for the last
chunk to be copied.

Andi. same change needed for x86-64.

    Dave.

Syncs up with latest from Stelian

Brings up to date with the quirk in 2.4
This makes sure we only apply it on the right chipsets,
and adjusts the bits that get poked depending on which we find.

Done by William Stinson.
Adds error handling to request_region() calls,
and converts some old check_region() calls too.