Warn once for each unknown command sent through SG_IO or the old SCSI
command ioctl.
Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Current BK doesn't compile with the G5 defconfig because the last
libata updated missed to change sata_svw to the new conventions.
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

Correct MODULE_PARM to module_param (somehow I changed the
macro parameters but not the macro invocation).
Signed-off-by: Randy Dunlap <rddunlap@osdl.org>
Signed-off-by: Randy Dunlap <torvalds@osdl.org>

Stuff marked initdata that is referenced in non-init context.

Referenced in non-init context.

The below removes an __initdata
(for initarray_generic that is referenced in non-init code).

The i386 timers use a struct timer_opts that has a field init
pointing at a __init function. The rest of the struct is not __init.

Nothing is wrong, but if we want to avoid having references to init stuff
in non-init sections, some reshuffling is needed.

The below replaces

struct timer_opts {
	int (*init)(char *override);
	... more ...
};

by the two structs

struct timer_opts {
	... more ...
};

struct init_time_opts {
	int (*init)(char *override);
	struct timer_opts *opts;
};

where the second is __initdata.

Routines referred to from __devinitdata.

In do_mounts_rd.c and initramfs.c there are many references
from .text to .text.init because of the inclusion of lib/inflate.c.
The below adds INIT markup in lib/inflate.c.

It is forbidden to refer from __devinit to __init.  (Unless one argues
that in the given situation hotplugging is actually impossible.  Too
ugly.)

However, es1371_probe() is __devinit and referred to __initdata initvol[]
and to __init src_init().

Many people are being bitten by the fact that if they select
CONFIG_PARTITION_ADVANCED in order to get some additional support,
they suddenly lose support for the MSDOS_PARTITION type.

Moreover, everybody needs MSDOS_PARTITION - also people not on i386,
since it is the type used on smart media and compact flash and
similar cards.

So, the below advises people "Say Y here" for MSDOS_PARTITION,
and does not change the default choices when PARTITION_ADVANCED
is selected.

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Fix up cpufreq Makefile merge manually

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Patch from Ben Dooks

Add PORT_S3C2440 to include/linux/serial_core.h

Signed-off-by: Ben Dooks
Signed-off-by: Russell King

Patch from Alan Cox, expanded to cover ARM drivers by Russell King.

- define SUPPORT_SYSRQ early
- ensure consistent include ordering
- use proper helper functions for flip buffer handling
- remove deadlock on overrun

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Patch from Ben Dooks

Add machine entry and support for the HP iPAQ rx3715
into arch/arm/mach-s3c2410, as well as updating the
default configuration for the s3c2410 to include it.

Signed-off-by: Ben Dooks
Signed-off-by: Russell King

Patch from Ben Dooks

Remove s3c2410_pm_init() if CONFIG_PM is not set.

Signed-off-by: Ben Dooks
Signed-off-by: Russell King

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Also remove dep on CONFIG_EXPERIMENTAL for libata itself,
and several libata drivers.

If CMSPAR doesn't exist, just #define it to zero. That makes the
code automatically DTRT.

device power state is in "dev.power.power_state" now, rather than
in "dev.power_state".

I had another look at the ugly rivafb memory ordering requirements.

It seems P{V,C}IO areas are only accessed using VGA_{RD,WR}8 macros. 
NV_{RW,WR}08 are never actually used directly.  So this patch makes at
least usage consistent: VGA_{RD,WR}8 to access "I/O areas" in an ordered
way.  NV_* for the rest.
Signed-off-by: Guido Guenther <agx@sigxcpu.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Bingo...  edd_devices[] was too short.  When we keep more than 6
signatures, it overruns the end.  Also, I rewrote edd_num_devices to be
clearer about its goal. 

This patch is necessary even after the last edd.S patch was reverted.

Thanks to Christian Kujau for testing and narrowing it down enough to
debug this. 

It still doesn't explain why Christian's BIOS reports more devices than
he has, that's still up in the air, so don't re-apply the edd.S patch
just reverted. 
Signed-off-by: Matt Domsch <Matt_Domsch@dell.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>