lxfb: GEODE: Add framebuffer support for the AMD Geode LX

Add framebuffer support for the AMD Geode LX graphics engine.

Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Signed-off-by: Antonino Daplas <adaplas@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

various doc/kernel-parameters fixes

- tell what APIC (by request), MTD, & PARIDE mean
- correct some source file names
- remove IA64 "llsc*=" (seems to have been removed from source tree)
- removel SCSI "53c7xx=" (driver already removed)

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Jesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

xilinxfb: Don't bail if the xilinxfb platform device doesn't have any pdata

Lack of pdata is not a fatal omission.  The driver can still be used even
if we do not know the screen dimensions.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Andrei Konovalov <akonovalov@ru.mvista.com>
Signed-off-by: Antonino Daplas <adaplas@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

xilinxfb:Remove __initdata from xilinx_fb_fix and xilinx_fb_var

xilinxfb_drv_probe refers to both tables, but it cannot be initdata.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Andrei Konovalov <akonovalov@ru.mvista.com>
Signed-off-by: Antonino Daplas <adaplas@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

s3c2410fb: fix s3c2410 compilation

The implicit mapping has been removed from the arch
as this should be handled in the driver, this patch
fixes the s3c2410_fb driver to ioremap() the necessary
registers.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Acked-by: Arnaud Patard <arnaud.patard@rtp-net.org>
Signed-off-by: Antonino Daplas <adaplas@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fbdev: Fix logo if logo depth is less than framebuffer depth

The VGA16 logo (bpp = 4) and monochrome logo (bpp = 1) do not contain any
color information (no CLUT).  If the fb depth is > logo depth, these logo's
will not properly display. Fix by using the console palette instead of
creating a new one.

Signed-off-by: Antonino Daplas <adaplas@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fbcon: Fix typo and bogus logic in get_default_font

Reported in: Bugzilla Bug 8727

Fix typo and bogus logic in get_default_font(). The bug results in
get_default_font() returning a font that may not be displayed properly by a
framebuffer driver.

Signed-off-by: Antonino Daplas <adaplas@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

synclink_gt: fix transmit DMA stall

Fix transmit DMA stall when write() called in window after previous
transmit DMA completes but before previous serial transmission completes.

Signed-off-by: Paul Fulghum <paulkf@microgate.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

serial_txx9: fix printk format mismatch

Since the commit 4f640efb3170dbcf99a37a3cc99060647b95428c, "%lx" is not
suitable for p->mapbase (resource_size_t) in 32-bit.  This patch fixes a
compiler warning caused by the mismatch.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

x86-64: Calgary: fix section mismatch warnings in tce

Fix section mismatch warnings:
these functions are called only from __init functions.

WARNING: vmlinux.o(.text+0x1861c): Section mismatch: reference to .init.text:free_bootmem (between 'free_tce_table' and 'build_tce_table')
WARNING: vmlinux.o(.text+0x187e5): Section mismatch: reference to .init.text:__alloc_bootmem_low (between 'alloc_tce_table' and 'kretprobe_trampoline_holder')

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

oom: print points as unsigned long

In badness(), the automatic variable 'points' is unsigned long.  Print it
as such.

Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rtc-dev: Make RTC driver return ENOTTY instead of ENOIOCTLCMD

Prevent the RTC driver from returning ENOIOCTLCMD to userspace.

Signed-off-by: Thomas Hommel <thomas.hommel@gefanuc.com>
Acked-by: Alessandro Zummo <a.zummo@towertech.it>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Do not trigger OOM-killer for high-order allocation failures

out_of_memory() may be called when an allocation is failing and the direct
reclaim is not making any progress.  This does not take into account the
requested order of the allocation.  If the request if for an order larger
than PAGE_ALLOC_COSTLY_ORDER, it is reasonable to fail the allocation
because the kernel makes no guarantees about those allocations succeeding.

This false OOM situation can occur if a user is trying to grow the hugepage
pool in a script like;

#!/bin/bash
REQUIRED=$1
echo 1 > /proc/sys/vm/hugepages_treat_as_movable
echo $REQUIRED > /proc/sys/vm/nr_hugepages
ACTUAL=`cat /proc/sys/vm/nr_hugepages`
while [ $REQUIRED -ne $ACTUAL ]; do
echo Huge page pool at $ACTUAL growing to $REQUIRED
echo $REQUIRED > /proc/sys/vm/nr_hugepages
ACTUAL=`cat /proc/sys/vm/nr_hugepages`
sleep 1
done

This is a reasonable scenario when ZONE_MOVABLE is in use but triggers OOM
easily on 2.6.23-rc1. This patch will fail an allocation for an order above
PAGE_ALLOC_COSTLY_ORDER instead of killing processes and retrying.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

s3c2410: fixup after arch moves

Fixup the changes from moving around the arch support for s3c24xx based
systems.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Acked-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Make Bruce Fields (co-)maintainer of kNFSDd

Make Bruce Fields (co-)maintainer of kNFSDd

Signed-off-by: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Fix uninitialized local variable "covered" in i386 acpi-cpufreq driver

The local variable "covered" is used without initialization in i386
acpi-cpufreq driver.  The initial value of covered should be 0.  The bug
will cause memory leak when hit.  The following patch fixes this bug.

Signed-off-by: Fenghua Yu <fenghua.yu@intel.com>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

update dontdiff file

Updates based on recent .gitignore updates:

*.o.*: Says Alexey Dobriyan:
These are presumably temporary gcc files, which aren't interesting.

setup.bin, setup.elf: new x86 boot code files (from Matthew Wilcox)

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

.gitignore update

Somehow I ended up with the following in tree:

$ git status
...
# Untracked files:
#   (use "git add <file>..." to include in what will be committed)
#
#       fs/proc/root.o.FuMxJQ
#       net/ipv4/tcp_minisocks.o.geCDYR

These are presumably temporary gcc files, which aren't interesting.

Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

NOMMU: Fix SYSV IPC SHM

Fix the SYSV IPC SHM to work with the changes applied by the new fault handler
patches when CONFIG_MMU=n.

Signed-off-by: David Howells <dhowells@redhat.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[RTNETLINK]: Fix warning for !CONFIG_KMOD

replay label is unused otherwise.

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

[IPV4] ip_options.c: kmalloc + memset conversion to kzalloc

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>

[DECNET]: kmalloc + memset conversion to kzalloc

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>

[BBC_I2C]: kmalloc + memset conversion to kzalloc

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>

[BBC_ENVCTRL]: kmalloc + memset conversion to kzalloc

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: ethtool_perm_addr only has one implementation

All drivers implement ethtool get_perm_addr the same way -- by calling
the generic function.  So we can inline the generic function into the
caller and avoid going through the drivers.

Signed-off-by: Matthew Wilcox <matthew@wil.cx>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: ethtool ops are the only way

During the transition to the ethtool_ops way of doing things, we supported
calling the device's ->do_ioctl method to allow unconverted drivers to
continue working.  Those days are long behind us, all in-tree drivers
use the ethtool_ops way, and so we no longer need to support this.

The bonding driver is the biggest beneficiary of this; it no longer
needs to call ioctl() as a fallback if ethtool_ops aren't supported.

Also put a proper copyright statement on ethtool.c.

Signed-off-by: Matthew Wilcox <matthew@wil.cx>
Signed-off-by: David S. Miller <davem@davemloft.net>

[PPPOE]: Improve hashing function in hash_item().

The new code produces the same results as the old version and is
~ 3 to 6 times faster for 4-bit hashes on the CPUs I tested.

Signed-off-by: Florian Zumbiehl <florz@florz.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

[MIPS] Add smp_call_function_single()

In the other archs, there is more factoring of smp call code, and more care
in the use of get_cpu(). That can be a follow-up MIPS patch.

Signed-off-by: Peter Watkins <pwatkins@sicortex.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] thread_info.h: kmalloc + memset conversion to kzalloc

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Kexec: Fix several 64-bit bugs.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Kexec: Fix several warnings.

arch/mips/kernel/machine_kexec.c: In function 'machine_kexec':
arch/mips/kernel/machine_kexec.c:54: warning: assignment makes integer from pointer without a cast
arch/mips/kernel/machine_kexec.c:70: warning: assignment makes integer from pointer without a cast
arch/mips/kernel/machine_kexec.c:81: warning: format '%08x' expects type 'unsigned int', but argument 2 has type 'long unsigned int'

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] DDB5477: Remove support

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Fulong: Remove unneeded header file

Signed-off-by: Songmao Tian <tiansm@lemote.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Cobalt: Enable UART on RaQ1

Unlike the current code suggests, the RaQ1 actually has an UART.  Only
the Qube1 (Qube 2700) lacks one.

Signed-off-by: Martin Michlmayr <tbm@cyrius.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Remove unused GROUP_TOSHIBA_NAMES

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] remove some duplicate includes

This patch removes some duplicate includes from arch/mips/

Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Oprofile: Fix rm9000 performance counter handler

The new type of irq handler remove a parameter (struct pt_regs *),but
someone forgot to supply it.

Signed-off-by: Dajie Tan <jiankemeng@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Use -Werror on subdirectories which build cleanly.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Yosemite: Fix warning.

arch/mips/pmc-sierra/yosemite/smp.c: In function 'titan_mailbox_irq':
arch/mips/pmc-sierra/yosemite/smp.c:112: warning: 'status' may be used uninitialized in this function

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] PMON: Fix cpustart declaration.

This now matches how cpustart is actually being invoked from Linux.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Yosemite: Only build ll_ht_smp_irq_handler() if HYPERTRANSPORT.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Yosemite: Fix build error due to undeclared titan_mailbox_irq().

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Yosemite: Don't declare titan_mailbox_irq() as asmlinkage.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Yosemite: Fix warnings in i2c-yoesmite by deleting the unused code.

arch/mips/pmc-sierra/yosemite/i2c-yosemite.c: In function 'titan_i2c_xfer':
arch/mips/pmc-sierra/yosemite/i2c-yosemite.c:98: warning: 'data' may be used uninitialized in this function

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Delete unused arch/mips/gt64120/common/

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Fix build warning in unaligned load/store emulator.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] IP32: Don't ignore request_irq's return value.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] IP27: Fix warning.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] BCM1480: Include <linux/module.h>.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Remove Momentum Ocelot support.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Fix archhelp.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Cleanup default bootfile format rule mess.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Replace use of stext with _stext.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Fix build breakage due to duplicate cpu_clock definition

commit e436d80085133858bf2613a630365e8a0459fd58
Author: Ingo Molnar <mingo@elte.hu>
Date:   Thu Jul 19 21:28:35 2007 +0200

    [PATCH] sched: implement cpu_clock(cpu) high-speed time source

broke the build of several MIPS platforms which were already using the
symbol cpu_clock for the own purposes.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] ARC: Fix modpost warnings.

WARNING: vmlinux.o(.text+0x1608): Section mismatch: reference to .init.text:ArcRead (between 'indy_8254timer_irq' and 'indy_rtc_set_time')
WARNING: vmlinux.o(.text+0x1157ac): Section mismatch: reference to .init.text:ArcRead (between 'prom_getchar' and 'prom_putchar')
WARNING: vmlinux.o(.text+0x115808): Section mismatch: reference to .init.text:ArcWrite (between 'prom_putchar' and 'csum_partial')

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] SMTC: Fix modpost warning.

WARNING: vmlinux.o(.text+0xcf54): Section mismatch: reference to .init.text:smp_bootstrap (between 'smtc_boot_secondary' and 'ipi_interrupt')

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Fix modpost warning.

WARNING: vmlinux.o(.text+0x1718): Section mismatch: reference to .init.text:mipsmt_build_cpu_map (between 'plat_smp_setup' and 'prom_init_secondary')

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] rtlx: Fix modpost warning

WARNING: vmlinux.o(.text+0x11504): Section mismatch: reference to .init.data:register_chrdev_failed (between 'rtlx_module_init' and 'rtlx_dispatch')

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Add missing declaration for mips_ihb().

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] MTI: Add CoreFPGA4 ID.

Signed-off-by: Chris Dearman <chris@mips.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] RP: Sysfs interface for stopping RP program

The old method of attempting to load a invalid program was just too icky.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] RP: Pass number of TCs available to RP program in $2.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] MT: Enable coexistence of AP/SP with VSMP and SMTC.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] SMTC: Safety net for i8259A interrupts.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] SMTC: smtc_timer_broadcast ignores its arguments, make it void.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] SMTC: Declare static what should be static.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] SMTC: Statically initialize irq_ipi[].

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] rbtx4938: Fix warnings

linux/arch/mips/pci/fixup-tx4938.c:21:5: warning: symbol 'pci_get_irq' was not declared. Should it be static?
linux/arch/mips/pci/fixup-tx4938.c:76: warning: passing argument 1 of 'pci_get_irq' discards qualifiers from pointer target type

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Sort system types alphabetically

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Fix RBTX49x7 board name

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] 0 -> NULL

When comparing a pointer, it's clearer to compare it to NULL than to 0.

Here is an excerpt of the semantic patch:

@@
expression *E;
@@

  E ==
- 0
+ NULL

@@
expression *E;
@@

  E !=
- 0
+ NULL

Signed-off-by: Yoann Padioleau <padator@wanadoo.fr>
Cc: ralf@linux-mips.org
Cc: linux-mips@linux-mips.org
Cc: akpm@linux-foundation.org
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] IP27: Fix modpost warning.

  MODPOST vmlinux.o
WARNING: vmlinux.o(.text+0x3180): Section mismatch: reference to .init.text:per_cpu_init (between 'prom_init_secondary' and 'alloc_cpupda')

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Wire up the fallocate syscall.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Fixup secure computing stuff.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] tx49xx: Add mach specific headers

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] rbtx4927: Fix some gcc warnings and a section mismatch.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] rbtx4927: Fix -Werror-implicit-function-declaration build error.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[x86 setup] EDD: Fix the computation of the MBR sector buffer

Some BIOSes require that sector buffers not cross 64K
boundaries.  As a result, we compute a dynamic address on the
setup heap.  Unfortunately, this address computation was just
totally wrong.

Signed-off-by: H. Peter Anvin <hpa@zytor.com>

[x86 setup] Newline after setup signature failure message

End the "No setup signature found..." with a newline (the puts
routine will automatically add a carriage return.)

Signed-off-by: H. Peter Anvin <hpa@zytor.com>

x86 boot code comments typos

Fix comments typos in new x86 boot code.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>

[XFRM]: State selection update to use inner addresses.

This patch modifies the xfrm state selection logic to use the inner
addresses where the outer have been (incorrectly) used. This is
required for beet mode in general and interfamily setups in both
tunnel and beet mode.

Signed-off-by: Joakim Koskela <jookos@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Diego Beltrami <diego.beltrami@gmail.com>
Signed-off-by: Miika Komu <miika@iki.fi>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

[IPSEC]: Ensure that state inner family is set

Similar to the issue we had with template families which
specified the inner families of policies, we need to set
the inner families of states as the main xfrm user Openswan
leaves it as zero.

af_key is unaffected because the inner family is set by it
and not the KM.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TCP]: Bidir flow must not disregard SACK blocks for lost marking

It's possible that new SACK blocks that should trigger new LOST
markings arrive with new data (which previously made is_dupack
false). In addition, I think this fixes a case where we get
a cumulative ACK with enough SACK blocks to trigger the fast
recovery (is_dupack would be false there too).

I'm not completely pleased with this solution because readability
of the code is somewhat questionable as 'is_dupack' in SACK case
is no longer about dupacks only but would mean something like
'lost_marker_work_todo' too... But because of Eifel stuff done
in CA_Recovery, the FLAG_DATA_SACKED check cannot be placed to
the if statement which seems attractive solution. Nevertheless,
I didn't like adding another variable just for that either... :-)

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TCP]: Fix ratehalving with bidirectional flows

Actually, the ratehalving seems to work too well, as cwnd is
reduced on every second ACK even though the packets in flight
remains unchanged. Recoveries in a bidirectional flows suffer
quite badly because of this, both NewReno and SACK are affected.

After this patch, rate halving is performed for ACK only if
packets in flight was supposedly changed too.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>

[PPPOL2TP]: Add CONFIG_INET Kconfig dependency.

PPPOL2TP uses UDP so it obviously depends on CONFIG_INET.

Signed-off-by: James Chapman <jchapman@katalix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: Page offsets and lengths need to be __u32.

Based upon a report from Stephen Rothwell.

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

[AF_UNIX]: Make code static.

The following code can now become static:
- struct unix_socket_table
- unix_table_lock

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NETFILTER]: Make nf_ct_ipv6_skip_exthdr() static.

nf_ct_ipv6_skip_exthdr() can now become static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

[PKTGEN]: make get_ipsec_sa() static and non-inline

Non-static inline code usually doesn't makes sense.

In this case making is static and non-inline is the correct solution.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

[PPPoE]: move lock_sock() in pppoe_sendmsg() to the right location

and the last one for now: Acquire the sock lock in pppoe_sendmsg()
before accessing the sock - and in particular avoid releasing the lock
even though it hasn't been acquired.

Signed-off-by: Florian Zumbiehl <florz@florz.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

[PPPoX/E]: return ENOTTY on unknown ioctl requests

here another patch for the PPPoX/E code that makes sure that ENOTTY is
returned for unknown ioctl requests rather than 0 (and removes another
unneeded initializer which I didn't bother creating a separate patch for).

Signed-off-by: Florian Zumbiehl <florz@florz.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

[IPV6]: ipv6_addr_type() doesn't know about RFC4193 addresses.

ipv6_addr_type() doesn't check for 'Unique Local IPv6 Unicast
Addresses' (RFC4193) and returns IPV6_ADDR_RESERVED for that range.

SCTP uses this function and will fail bind() and connect() calls that
use RFC4193 addresses, SCTP will also ignore inbound connections from
RFC4193 addresses if listening on IPV6_ADDR_ANY.

There may be other users of ipv6_addr_type() that could also have
problems.

Signed-off-by: Dave Johnson <djohnson@sw.starentnetworks.com>
Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: Fix prio_tune() handling of root qdisc.

Fix the check in prio_tune() to see if sch->parent is TC_H_ROOT instead of
sch->handle to load or reject the qdisc for multiqueue devices.

Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: Fix sch_api to properly set sch->parent on the root.

Fix sch_api to correctly set sch->parent for both ingress and egress
qdiscs in qdisc_create().

Signed-off-by: Patrick McHardy <trash@kaber.net>
Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET_SCHED]: Fix prio/ingress classification logic error

Fix handling of empty or completely non-matching filter chains. In
that case -1 is returned and tcf_result is uninitialized, the
qdisc should fall back to default classification in that case.

Noticed by PJ Waskiewicz <peter.p.waskiewicz.jr@intel.com>.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

[IPV6]: Remove circular dependency on if_inet6.h

net/if_inet6.h includes linux/ipv6.h which also tries to include
net/if_inet6.h.  Since the latter only needs it for forward
declarations, we can fix this by adding the declarations.

A number of files are implicitly including net/if_inet6.h through
linux/ipv6.h.  They also use net/ipv6.h so this patch includes
net/if_inet6.h there.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

[IPV4/IPV6]: Fail registration if inet device construction fails

Now that netdev notifications can fail, we can use this to signal
errors during registration for IPv4/IPv6.  In particular, if we
fail to allocate memory for the inet device, we can fail the netdev
registration.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: Allow netdev REGISTER/CHANGENAME events to fail

This patch adds code to allow errors to be passed up from event
handlers of NETDEV_REGISTER and NETDEV_CHANGENAME.  It also adds
the notifier_from_errno/notifier_to_errnor helpers to pass the
errno value up to the notifier caller.

If an error is detected when a device is registered, it causes
that operation to fail.  A NETDEV_UNREGISTER will be sent to
all event handlers.

Similarly if NETDEV_CHANGENAME fails the original name is restored
and a new NETDEV_CHANGENAME event is sent.

As such all event handlers must be idempotent with respect to
these events.

When an event handler is registered NETDEV_REGISTER events are
sent for all devices currently registered.  Should any of them
fail, we will send NETDEV_GOING_DOWN/NETDEV_DOWN/NETDEV_UNREGISTER
events to that handler for the devices which have already been
registered with it.  The handler registration itself will fail.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET] loopback: Panic if registration fails

Because IPv4 and IPv6 both depend on the presence of the loopback
device to function, failure in registration the loopback device should
be fatal.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: Take dev_base_lock when moving device name hash list entry

When we added name-based hashing the dev_base_lock was designated as the
lock to take when changing the name hash list.  Unfortunately, because
it was a preexisting lock that just happened to be taken in the right
spots we neglected to take it in dev_change_name.

The race can affect calles of __dev_get_by_name that do so without taking
the RTNL.  They may end up walking down the wrong hash chain and end up
missing the device that they're looking for.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>