This fixes CONFIG_UID16 problems on x86-64 as discussed earlier.

CONFIG_UID16 now only selects the inclusion of kernel/uid16.c, all
conversions are triggered dynamically based on type sizes.  This allows
x86-64 to both include uid16.c for emulation purposes, but not truncate
uids to 16bit in sys_newstat. 

- Replace the old macros from linux/highuid.h with new SET_UID/SET_GID
  macros that do type checking. Based on Linus' proposal.

- Fix everybody to use them.

- Clean up some cruft in the x86-64 32bit emulation allowed by this
  (other 32bit emulations could be cleaned too, but I'm too lazy for 
  that right now) 

- Add one missing EOVERFLOW check in x86-64 32bit sys_newstat while
  I was at it.

Needs a forward declaration of struct pt_regs

Current match_token recognize "foo=" as "foo=%s". So this change the
nul string does not recognize as "%s".

If original code uses the following,

	opts->value = simple_strtoul(value, &value, 0);
	if (*value)
		return error;

It doesn't recognize negative value. So option format should use "%u".

This is just misc cleanup of fat option.

vfat doesn't recognize the msdos option. Also msdos doesn't recognize
vfat option.

And this separates fat/vfat/msdos option structures. I think this
looks more good as document.

This uses more fine granularity ->d_lock than ->dcache_lock for
->d_parent with d_move() race.

From René Scharfe <l.s.r@web.de>

there's one call to strncpy() in vfat that really should be a memcpy().
The source, msdos_name, is a space-filled char array, not a
NUL-terminated string.

The rest of the patch eliminates duplicate occurencies of the
CURRENT_TIME macro inside functions. This shrinks vfat.o by a few
bytes.

It seems that we're attempting to work out the free stack size using two
unrelated pointers for the lowest address of the stack.  Fix this to use
the correct pointer.

into home.osdl.org:/home/torvalds/v2.5/linux

into samba.org:/home/paulus/kernel/for-linus-ppc

From: Felipe W Damasio <felipewd@terra.com.br>

Found by smatch.

Remove outdated URLs from the x86-64 Kconfig.

Originally from the trivial patch monkey / Petri Koistinen <petri.koistinen@iki.fi>

Make linux32 ... on x86-64 to do proper uname emulation again.

Also changes i386 to i686 because that's more near the truth.

The zfcp scsi host adapater.

 - Remove read, write and data device pointers from card structure. Use
   ccw group array to get the device pointers.

- Add dummy release function to iucv bus.

Kernel Janitors: remove unnecessary calls to verify_area.

 - Remove dynamic allocation of major numbers. Just use static major 94.
 - Use bus_id instead of device number where possible.
 - Don't check open_count in dasd_generic_set_offline.

 - Remove atomic_read hack to test for presence of ccw device directory.
 - Add dummy release function to channel path object and cu3088 group devices.
 - Don't rely on the chaining bit of the channel report word.
 - Don't call schedule while holding a lock in read_dev_chars/read_conf_data.

 - Makefile fix: default make target builds the kernel image.
 - Add export statement for cpcmd.
 - Add tgkill system call.
 - Reserve system call number for vserver.

- Maintainers email address and copyright messages updated
- Scanning for EISA and ISA controllers disabled due to a resource conflict
  with the Adaptec aic driver
- Switch "probe_eisa_isa" added to give the possibility to enable the
  EISA/ISA scan if required
- Driver version increased to 2.08

idle_cpu had the same botched move from kernel/ksyms.c to kernel/sched.c
that __wake_up_sync() had.

From: Alexey Dobriyan <adobriyan@mail.ru>

Don't set process state to TASK_RUNNING after a schedule_timeout(), since
there's no way we will return from the scheduler if we're not running.

from kernel/ksyms.c to kernel/sched.c.

Noted by Richard Henderson <rth@twiddle.net>

into home.osdl.org:/home/torvalds/v2.5/linux

into flint.arm.linux.org.uk:/usr/src/bk/linux-2.6-rmk

update.

The socket is "s", not "skt".

into home.osdl.org:/home/torvalds/v2.5/linux

into flint.arm.linux.org.uk:/usr/src/bk/linux-2.6-pcmcia

into home.osdl.org:/home/torvalds/v2.5/linux

I'm trying to remove all the legacy UART cruft from ia64.  In theory
we should be able to discover all UARTs on ia64 via ACPI and PCI
enumeration, so we shouldn't need SERIAL_PORT_DFNS.

But SERIAL_PORT_DFNS currently does two things:
        1) it tells you about legacy devices you can't discover
           via a standard enumeration method, and
        2) it sizes old_serial_port[], which determines UART_NR,
           which is the maximum number of ports the driver will
           support

So here's a proposal to make SERIAL_PORT_DFNS optional and
provide another mechanism for configuring the number of ports
to support.

Patch from Bjorn Helgaas.

RS_TABLE is defined by lots of architectures, but only referenced
in arch/{mips,ppc}.

This patch removes most of the unused definitions from 2.6.

into flint.arm.linux.org.uk:/usr/src/bk/linux-2.6-serial