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

The timex/time code requires CLOCK_TICK_RATE to be constant.  We
assume that the platform picks an appropriate clock source such
that we generate an exact HZ value, and set CLOCK_TICK_RATE to
a value where ACTHZ == HZ.

The device_driver suspend/resume methods are no longer used.  Instead,
the bus_type contains the suspend/resume methods.  Fix the SA1111 bus
support for this change.

We place the probe/remove/suspend/resume methods inside struct
sa1111_driver and call them from the SA1111 bus driver (ie, how
Pat wants this stuff done.)

We leave the parent bus device suspend/resume methods in the device
driver until power management for platform devices works again.
However, we adjust these methods so they run only once, like the
other PM methods.

The device_driver suspend/resume methods are no longer used.  Instead,
the bus_type contains the suspend/resume methods.  Fix the AMBA bus
support for this change.

Old binutils (without .incbin) had the idea that a certain assembler
instruction was illegal.  binutils has since been fixed to allow it.

This ensures that we release reset on devices behind a PCI bridge,
and that we have error reporting enabled behind bridges.

These calls no longer take "level" arguments, so there's no need to
call them multiple times.

Since we're now 32-bit "armv" only, we don't need to select the
linker script in vmlinux.lds.S.  Also, whoever moved vmlinux.lds.S
into arch/arm/kernel forgot to move the other scripts.

This cset replaces arch/arm/kernel/vmlinux.lds.S with
arch/arm/vmlinux-armv.lds.in, and deletes the obsolete scripts.

Patch from Nicolas Pitre

In both 2.5.70-rmk1 and 2.4.19-rmk7 the syscall macros are clobering
the returned error value when building library code.

Example code:

#include <linux/unistd.h>
#include <errno.h>
extern int fake_syscall(int x, int y, int z);
_syscall3(int, fake_syscall, int, x, int, y, int, z)

Current generated code:

fake_syscall:
        @ args = 0, pretend = 0, frame = 0
        @ frame_needed = 0, uses_anonymous_args = 0
        str     lr, [sp, #-4]!
        swi     __NR_fake_syscall
        cmn     r0, #126
        ldrls   pc, [sp], #4
        bl      __errno_location
        rsb     r3, r0, #0
        str     r3, [r0, #0]
        mvn     r0, #0
        ldr     pc, [sp], #4

In the code above, whenever the return value is an error code, it
is lost due to the call to __errno_location. And because of the
asm("r0") constraint on the variable __res the compiler continues
using r0 for it even if it's now a pointer value.  errno ends up with
a totally bogus value.

With the patch below the above code becomes:

fake_syscall:
        @ args = 0, pretend = 0, frame = 0
        @ frame_needed = 0, uses_anonymous_args = 0
        stmfd   sp!, {r4, lr}
        swi     __NR_fake_syscall
        cmn     r0, #126
        mov     r4, r0
        bls     .L3
        bl      __errno_location
        rsb     r3, r4, #0
        str     r3, [r0, #0]
        mvn     r4, #0
.L3:
        mov     r0, r4
        ldmfd   sp!, {r4, pc}

which is correct.

Oh and added a small estetic change for generated code too.

Patch from Frank Becker

GPIO for 48 MHz clock output is 7 not 8.

Patch from Deepak Saxena

This is an update to patch 1529/1 that cleans up the code so we don't need 
#ifdef's for little vs. big-endian systems. Tested on both systems with
various network apps (ping, ftp, tftp, ssh, telnet, NFS root, http) with no
issues.

Patch from Deepak Saxena

Patch from Deepak Saxena

Patch from Deepak Saxena

Removes extraneous bits that belong to separate IOP3xx PCI cleanup patch
Supersedes 1618/1

Patch from Deepak Saxena

Patch from Deepak Saxena

Small cleanups for ADIFCC and IOP3XX machine types to support
ATAG parameters. Working with Intel and ADI to get updated 
bootloaders that pass the tags. 

Also, all known IOP3xx boards have memory starting at 0xa0000000,
so we can remove the redundant ARCH_IQ* zreladdr values.

Patch from Deepak Saxena

Fix IOP321 and IQ80310 timer interrupts to return IRQ_HANDLED

Patch from Deepak Saxena

This is required for usr/initramfs_data.o to build properly when
CONFIG_CPU_BIG_ENDIAN is enabled.

Patch from Holger Freyther

Patch from Holger Freyther

Patch from Holger Freyther

see content

sound/pcmcia/vx/vx_entry.c:

This patch removes the PCMCIA timer release functionality which is no
longer required. Without this the module does not compile.

Too much cut-and-paste, noticed by Dmitry Torokhov <dtor_core@ameritech.net>

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

On PPC32 we reorder the arguments so they fit into 6 registers.  Glibc will
need a two-line stub to change them from the standard order to the ordering
used by the system call: (fd, advice, offset, len).

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

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

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

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

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

In both 2.4 and in 2.6, error handling for bad cdrom media is
wrong.  And it is my fault I'm afraid, since I botched an earlier
fix for the problem by putting the fix in the wrong spot.

My kids have a "Jumpstart Toddlers" cd they have long since
completely killed, which makes a great test disc.  Without this
fix, the best time projection I can get for completing a dd type
sector copy is about 2 years...  Most of that is spent thrashing
about in kernel space trying to re-read sectors we already know
are not correctable....  After the fix, I was able to rip a copy
the CD (or rather muddle through it getting lots of EIO errors)
in about 15 minutes.

Attached is the fix for 2.6.x,

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