This updates the PowerMac-only platinumfb driver to use the new mac-io
device infrastructure.  It also switch allocation to the new
framebuffer_alloc/release and fix a couple of bugs.

This adds a limit on how much of the framebuffer is ioremap'ed by
radeonfb, thus enabling it to work with 128Mb VRAM or more on an x86
with 900Mb of lowmem in the linear mapping.

It also adds a significant amount of debug messages and adds a CONFIG
option to enable the debugging output, that should help with diagnosing
new problems. Among others, it dumps the connector info as I understand
them (so far, they give "strange" informations on laptops, I need more
data on more various laptops to see if there's a pattern I can really use
to figure out on which connector the LVDS is)

Regarding the "lid closed at boot", ultimately, we may want to default
to the VGA output in those cases, though I'm not sure what logic to use
here. Maybe we could standardize some way for the platform to provide
this "environment" information to the driver, but i wouldn't rely on it.

More reliably, if we can find out that there is an LVDS output, and
LVDS is disabled, just ignore the flat panel...

We could assume any mobility chip has LVDS, which is true, but that would
still cause a problem for laptops with an additional DVI output (only
Macs so far afaik).

This should merge up the final piece of the ppc32/64 saga: the mac
PowerMac MACE and Airport network drivers.

Both of them are ported to the mac-io infrastructure, all probe code
rewritting & cleaned up, better error handling & resource management.

flat-panel displays.

From: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Fix proper detection of the "noaccel" command line argument for
new radeonfb so we can boot without acceleration. Useful when
diagnosing an accel-related problem.

Somebody has been watching lord of the rings a bit too
much.. "My precioussssss.."

From Peter Osterlund <petero2@telia.com>

This code only ends up being used when all else fails,
so probably very few people actually ever saw this.

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

The RFCOMM TTY code don't leak reference counting, because the TTY layer
will call the ->close() method even if open fails and the reference count
is decreased there.

Patch from David Woodhouse <dwmw2@infradead.org>

You can still build the old driver, but it doesn't work unless you also
enable it like this..

This makes fbcon ask for notification of events from fbdev to deal with
suspend/resume (stop cursor on suspend, refresh screen on resume).
Could probably do more (like dealing better with the cursor timer), but
this simple implementation works fine enough for now.

This adds some "state" information for power management to fbdev's,
along with a notifier mecanism to inform clients of state changes.  It
also "uses" this mecanism in the function fb_set_suspend() which was an
empty placeholder previously, and "shields" various places that access
the HW when state isn't running.  (It's best to not call them in the
first place, but the current state of fbcon makes that _very_ difficult)

This updates the aty128fb driver.  It adds more PCI IDs, uses the new
framebuffer alloc/release functions, make BIOS PLL data access more
reliable (using ROM whenever possible, with a fallback to RAM BIOS
image), cleanup the Power Management stuff (get rid of PowerMac specific
stuffs, use real PCI ones instead), along with some style cleanups

This removes the broken locking code in the pixmaps, and rewrite the
buffer access function to properly call fb_sync when needed.  The old
broken loocking is useless as we are covered by the console semaphore in
all cases hopefully (except if I missed one :)

This fixes the fbdev ioctl's and fbcon cursor management with the
console semaphore, which is the best we can do at this point in 2.6,
thus fixing a bunch of races where we could have, for example, tried to
blit while changing mode, etc..

Independently from the other fbdev updates I'm cooking (some of them
will be in your mailbox rsn), this fixes an error in parameter passing
to a function in rivafb (only used on ppc) that could cause an oops and
definitely causes a warning at compile time.

From: Keith M Wesolowski <wesolows@foobazco.org>

As of 2.6.3, restore_flags will no longer modify cwp on sparc.
Therefore you can apply this patch to the locking guide.

[ Indeed.  I'll also remove the atomic comments from Hacking
  Guide as part of my revision there when I get back to it.  --RR ]

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

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

arch/arm/kernel/time.c: Fix time_before type warning.
arch/arm/common/amba.c: Fix snprintf compiler warning.

Thus fixes the weird kconfig message "optimize || ?", it's an old debug
check and is triggered by the unusual dependency.  It's not incorrect,
but the solution below is better and it's the same FB_MATROX_I2C already
uses.

I'll send a fix for the kconfig message later.

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

This backs out James' sysfs support for fbdev again.  It introduces a
big, race for every driver not converted to framebuffer_{alloc,release}
(that is every driver but Ben's new radeonfb).

I've left in framebuffer_{alloc,release} as stubs so drivers can be
converted to it gradually and once all drivers are done it can be
enabled again.

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

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

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

Theres still more to do here, but at least the ifdef mess is gone. No
more checking for NULL before calling functions, that was playing with
fire. Oh yeah and lots more deletions :)

Clean up the debugger hooks, it was way too easy to screw up.
And we did. And Linus hit it.

- create CONFIG_DEBUGGER so we can enable kernel debugging options but not
  have any trace of debugger gunk.
- remove a bunch of xmon prototypes so no one gets the urge to call them
- Use die() instead of panic in a number of places, it gives us much better
  debug information.
- Get rid of the ifdef madness

Heres a patch I've had for a while, it removes a bunch of debugger code
which is good :) The next patch will sanitise it (and the rest of the
debugger hooks).

Various xmon cleanups

- recover from bad SPR read/write (we get a program check)
- remove some old code (bat and segment register stuff)
- update the help text to match reality
- add a "press ? for help" when xmon first appears to make rusty happy
- protect against flushing bad parts of memory from Milton
- dont print iseries specific stuff on pseries in SPR dump (S)
- add code to dump the segment table or SLB
- remove a number of functions that wouldnt work on LPAR

- Add thread_info to pointer, its a useful piece of information.
- Do the kallsyms lookup on the link register
- Remove extra newline on one call to die()

From: Paul Mackerras <paulus@samba.org>

Recent changes in include/linux/*.h meant that likely()
isn't defined here (since we don't set __KERNEL__), and thus
we don't get some prototypes and we can't use do_div.  This
fixes the resulting compile errors and warnings.

Remove %L handling from sprintf - we don't need it, and it
meant we needed do_div from asm/div64.h, which gives problems
when __KERNEL__ isn't defined.  Also add a prototype for
strlen to kill a warning.

into kernel.bkbits.net:/home/davem/sctp-2.6

into kernel.bkbits.net:/home/davem/sparc-2.6

into kernel.bkbits.net:/home/davem/netfix-2.6