Endianness fixes

- rtl8169_rx_csum() forgot to convert the descriptor to cpu order,
- same thing for rtl8169_rx_vlan_skb() but this one is more tricky as
  the layout of the (u32) word in the 8169 descriptor calls for a
  second level of swap at the vlan tag level (u16). The old code only
  did the second part (in an endian-dependant way, how fun).
- rtl8169_tx_vlan_tag(): this time the (u32 descriptor level) cpu_to_le32
  is issued in rtl8169_start_xmit but the (u16 vlan tag level) cpu_to_be16
  is not right any more.

Summary: avoid even calls to cpu_to_{b/l}eXX on a given piece of data.

The change has no effect on x86. Now sparc64 talks to x86.
Pinpointed-by: Jon Mason <jdmason@us.ibm.com>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

The dynamic clock code in radeonfb comes almost as-is from X.org (where
it was contributed by ATI). It has a few typos (wrong register access
macros) that this patch fixes.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

A simple 'grep memset.*\<0);' shows argument order errors in several
uses of memset.

This grep was inspired by Al Viro's recent patch, megaraid_mbox fix,
which fixed this problem in the megaraid driver.

This removes the CRs from the kernel file of Eicon ISDN driver.

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

Patch from Ben Dooks

Off by 4 error on S3C2440_CAMDVIN register address, and
re-indented the pll calculation code to proper C style.
Thanks to Guillaume Gourat for spotting the register mistake

Signed-off-by: Guillaume Gourat

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

Patch from Ben Dooks

GPIOJ12 alternate function is CAMRESET, not CAMCLKOUT
as mis-copied on creation
Patch from Guillaume Gourat

Signed-off-by: Guillaume Gourat

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

Patch from Lennert Buytenhek

The alignment trap handler in arch/arm/mm/alignment.c performs halfword
and word accesses by translating them into byte accessing and then
combining the pieces.  However, it assumes little-endian byte ordering
in doing so, which causes unaligned accesses on big-endian platforms
to erroneously load or store byteswapped data.
This patch was pulled from an old 2.4 -ds tree, and appears to fix the
issue.  It was submitted for 2.4 at the time (ARM patch ID 1205/1206),
but rejected since big-endian ARM platforms are not supported in 2.4.

Signed-off-by: Lennert Buytenhek
Signed-off-by: Russell King

For device memory, just return a value that will never
cause pte_pfn() to return true.

Noticed by Tom Callaway.
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: David S. Miller <davem@davemloft.net>

This fixes the raid5 and raid6 prolems that crept in with the recent
introduction of "bi_max_vecs".

Since raid5/raid6 allocate their own bio's, they need to make sure
that bi_max_vecs is updated along with bi_vcnt.

wrong order of arguments in memset().

This, BTW, shows why cross-builds are useful - the only indication of
problem had been a new warning showing up in sparse output on alpha
build (number exceeding 256 got truncated).
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk>

The vmap vmalloc rework in 2.5 had a unintended side effect.  vmalloc uses
kmalloc now to allocate an array with a list of pages.  kmalloc has a 128K
maximum.  This limits the vmalloc maximum size to 64MB on a 64bit system
with 4K pages.  That limit causes problems with other subsystems, e.g.
iptables relies on allocating large vmallocs for its rule sets.

This is a bug IMHO - on 64bit platforms there shouldn't be such a low limit
on the vmalloc size.  And even on 32bit it's too small for custom kernels
with enlarged vmalloc area.

Another problem is that this makes vmalloc unreliable.  After the system
has been running for some time it is unlikely that kmalloc will be able to
allocate >order 2 pages due to memory fragmentation.

This patch takes the easy way out for fixing this by just allocating this
array with vmalloc when it is larger than a page.  While more complicated
and intrusive solutions would be possible they didn't use vmalloc
recursively they didn't seem it worth to handle this very infrequent case.

Please note that the vmalloc recursion is strictly bounded because each
nested allocation will generate a much smaller stack frame.  Also the
kernel stack can handle even a few recursion steps easily because vmalloc
has only a small stack frame.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This patch removes excess '~' before the bit masks.
Signed-off-by: Andrei Konovalov <akonovalov@ru.mvista.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

dma_declare_coherent_memory() is calling kmalloc with wrong arguments.
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This fixes an issue related to the second PCI host controller working on
MPC8555/41 systems.

We track the last PCI bus number on PCI1 so that the PCI2 host
controller can properly exclude itself at the right time, exclusion
should occur after initial setup so that the early pci config cycles in
setting PCI2 actually get to the controller. 
Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
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

Cleanups (initially sent by Adrian Bunk):
 - make some needlessly global code static
 - removed obsolete #define OLD_MAX_DESCRIPTORS
 - removed more platform independend code not used in linux
 - removed dos-<CR> at end of lines 
 - fix indentation in already modified files
Signed-off-by: Armin Schindler <armin@melware.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

convert from pci_module_init to pci_register_driver
Signed-off-by: Christophe Lucas <c.lucas@ifrance.com>
Signed-off-by: Armin Schindler <armin@melware.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

My recent do_munmap hugetlb fix has proved inadequate.  There are
other places (madvise, mbind, mlock, mprotect) where split_vma is
called.  Only mprotect excludes a hugetlb vma: the others are in
danger of splitting at a misaligned address, causing later BUGs.

So move the ~HPAGE_MASK check from do_munmap to split_vma itself;
and fix up those places (madvise and mlock) which expect split_vma
can fail only with -ENOMEM, and wish to convert that to -EAGAIN.
(It appears genuine that some of these syscalls should be failing
with -ENOMEM and some with -EAGAIN, so respect those behaviours.)

madvise_dontneed doesn't use split_vma, but is equally in danger
of causing a hugetlb BUG via zap_page_range.  Whereas elsewhere the
patch is permissive (allowing the operation on a hugetlb vma even when
pointless, so long as it doesn't missplit it), here we must use -EINVAL
on any hugetlb vma, since a page fault would hit the BUG in its nopage.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Acked-by: William Irwin <wli@holomorphy.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Signed-off-by: David S. Miller <davem@davemloft.net>

This patch fixes the problem reported in:

http://marc.theaimsgroup.com/?l=linux-kernel&m=110798711911645&w=2


The 5700 link problem was caused by reading uninitialized values in sram and
causing capacitive coupling mode to be enabled by mistake. This patch fixes
the problem by properly validating the sram contents.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This fixes the mincore syscall in three ways:

 (1) It moves as much argument checking outside of the semaphore-holding
     region as possible.

 (2) It checks the region parameters against TASK_SIZE so that a 32-bit binary
     on a 64-bit platform will get the right error when calling this syscall
     on a region that overlaps the end of the 32-bit address space.

 (3) It tidies up the VMA checking loop a little.
Signed-Off-By: David Howells <dhowells@redhat.com>
Signed-Off-By: Linus Torvalds <torvalds@osdl.org>

The attached patch fixes sigaltstack handling for RT signal return. It was
reading a userspace struct into kernel space and then passing the kernel copy
to a generic signalling routine which then assumed it had been passed a
userspace pointer...
Signed-Off-By: David Howells <dhowells@redhat.com>
Signed-Off-By: Alexander Viro <aviro@redhat.com>
Signed-Off-By: Linus Torvalds <torvalds@osdl.org>

do_i2c_rdwr_ioctl() does two compat_alloc_user_space().  That doesn't
work; no state is kept and second allocation will ignore the first one
(i.e.  give overlapping chunk of user stack).

Fixed by doing allocation at once, slightly cleaned up.
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

dm-stripe does do_div() on potentially 32bit data.  do_div()
implementation has every right to be Not Happy(tm) with that...

The fix is obvious - that's what sector_div() is for...
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This patch removes TANBAC_TB0219 doubly registered in kernel config.
Signed-off-by: Yoichi Yuasa <yuasa@hh.iij4u.or.jp>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The mcd driver drives only very old hardware (some single and double speed
CD drives that were connected either via the soundcard or a special ISA
card), and the mcdx driver offers more functionality for the same hardware.

My plan is to mark MCD as broken in 2.6.11 and if noone complains 
completely remove this driver some time later.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

It adds the sleep support for newer powermacs, improve power saving on some
laptops, makes use of the new fbdev modelist management routines, and fixes
a few backlight related issues.

I tested it on a thinkpad T30 and a few PPC boxes with success.  It should
be less invasive than the previous one (I don't try to restore the mode on
exit, that is what breaks the thinkpad and possibly other stuffs that boot
in VGA text mode), plus fixed a couple of bugs in the mode detection code. 
I also reverted the memory map fix on ppc since it doesn't work properly on
some recent laptops where the firmware sets a tiled display.  I'll rework
that completely to update the memory map as part of the mode setting later.
That should fix various issues when switching with X/DRI on x86.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This patch updates aty128fb power management code to the changes of the
powermac sleep mecanism.  It makes the driver use the new hook for early
wakeup, adds the call to the arch code indicating wether it can wakeup the
chip, etc...  This patch shouldn't break non-ppc, but this should be
tested.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The new PowerMac sleep code provides an arch hook that can be used by video
drivers on laptops to bring back the screen very early, pretty much before
anything else.  This basically turns reports of the style "my laptop doesn't
wakeup" to "I get this or this oops/error/panic on wakeup", making fixing the
PM related bugs possible on a whole range of them.

However, the fbdev wakeup code triggers WARN_ON's in the VT subsystem if
called without the console semaphore when redrawing the screen (I added those
warnings a couple of kernel versions ago), and we can't call
acquire_console_sem() since we are so early in the wakeup process that we are
considered as in_atomic() (we hold irqs off too).

This patch addds a try_acquire_console_sem() function that can be used by
those video drivers that implement this early wakeup hook.  If the acquire
fails (which should never happen in practice), wakeup is delayed to the normal
PCI callback which does a blocking acquire_console_sem().
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The 1999 version of the DMI spec had a different configuration than the
newer versions for the IPMI configuration information.  This patch handles
the differences between the two.
Signed-off-by: Corey Minyard <minyard@acm.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

In linux 2.6, PTRACE_SETOPTIONS is redefined to 0x4200, while the old 2.4
value (21) is still available as PTRACE_OLDSETOPTIONS.

So, if UML uses PTRACE_SETOPTIONS, an UML-kernel built on a 2.6 won't run
on a 2.4 host.  Hence we must use PTRACE_OLDSETOPTIONS.

For cases when PTRACE_OLDSETOPTIONS does not exists (i.e.  2.4 host or
archs which miss it because they don't have a "deprecated" value), we
fallback this macro to PTRACE_SETOPTIONS.
Signed-off-by: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Fix a typo in the Makefile cleanup merged earlier, which causes compile
failures in some edge cases.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

into nuts.davemloft.net:/disk1/BK/net-2.6

into nuts.davemloft.net:/disk1/BK/sparc-2.6