[PATCH] __cmpxchg() must really always be inlined on alpha

With the latest 2.6.15 kernel builds for alpha on Debian, we ran into a
problem with undefined references to __cmpxchg_called_with_bad_pointer() in
a couple of kernel modules (xfs.ko and drm.ko; see
http://bugs.debian.org/347556).

It looks like people have been trying to out-clever each other wrt the
definition of "inline" on this architecture :), with the result that
__cmpxchg(), which must be inlined so the compiler can see its argument is
const, is not guaranteed to be inlined.  Indeed, it was not being inlined
when building with -Os.

The attached patch fixes the issue by adding an
__attribute__((always_inline)) explicitly to the definition of __cmpxchg()
instead of relying on redefines of "inline" elsewhere to make this happen.

Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ide: cast arguments to pr_debug() properly

This does not show up unless you #define DEBUG in the file, which most
people wouldn't do.  On PPC405, at least, "sector_t" is unsigned long,
which doesn't match %llx/%llu.  Since sector# may well be >32 bits, promote
the value to match the format.

Signed-off-by: Michael Richardson <mcr@xelerance.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] knfsd: fix nfs4_open lock leak

I just noticed that my patch "don't create on open that fails due to
ERR_GRACE" (recently commited as fb553c0f17444e090db951b96df4d2d71b4f4b6b)
had an obvious problem that causes a deadlock on reboot recovery.  Sending
in this now since it seems like a clear 2.6.16 candidate.--b.

We're returning with a lock held in some error cases.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] hugetlbpage: return VM_FAULT_OOM on oom

Remove wrong and misleading comments.

Return VM_FAULT_OOM if the hugetlbpage fault handler cannot allocate a
page.  do_no_page will end up doing do_exit(SIGKILL).

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] Hugepages need clear_user_highpage() not clear_highpage()

When hugepages are newly allocated to a file in mm/hugetlb.c, we clear them
with a call to clear_highpage() on each of the subpages.  We should be
using clear_user_highpage(): on powerpc, at least, clear_highpage() doesn't
correctly mark the page as icache dirty so if the page is executed shortly
after it's possible to get strange results.

Signed-off-by: David Gibson <dwg@au1.ibm.com>
Acked-by: William Lee Irwin III <wli@holomorphy.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] x86_64: Fix the node cpumask of a cpu going down

Currently, x86_64 and ia64 arches do not clear the corresponding bits in
the node's cpumask when a cpu goes down or cpu bring up is cancelled.  This
is buggy since there are pieces of common code where the cpumask is checked
in the cpu down code path to decide on things (like in the slab down path).
 PPC does the right thing, but x86_64 and ia64 don't (This was the reason
Sonny hit upon a slab bug during cpu offline on ppc and could not reproduce
on other arches).  This patch fixes it for x86_64.  I won't attempt ia64 as
I cannot test it.

Credit for spotting this should go to Alok.

(akpm: this was applied, then reverted.  But it's OK now because we now use
for_each_cpu() in the right places).

Signed-off-by: Alok N Kataria <alokk@calsoftinc.com>
Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Shai Fultheim <shai@scalex86.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] swsusp: kill unneeded/unbalanced bio_get

- Remove unneeded bio_get() which would cause a bio leak

- Writing doesn't dirty pages.  Reading dirties pages.

- We should dirty the pages after the IO completion, not before

(Busy-waiting for disk I/O completion isn't very polite.)

Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] drivers/base/bus.c warning fixes

drivers/base/bus.c:166: warning: `driver_attr_unbind' defined but not used
drivers/base/bus.c:194: warning: `driver_attr_bind' defined but not used

Looks like these two attributes and supporting functions want to be
#ifdef HOTPLUG'd

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Merge /pub/scm/linux/kernel/git/mchehab/v4l-dvb

[PATCH] cciss: softirq handler needs to save interrupt flags

The softirq rq completion handler needs to save/restore interrupt flags
appropriately.

Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

[PATCH] Fix bad apic fix on i386

Fix wrong '!' in bad apic fix

I forgot to remove the ! when moving the code from x86-64 to i386 x86-64
tested !disable_apic, but of course for cpu_has_apic it shouldn't be
negated.

Credit goes to Jan Beulich for spotting it with eagle eyes.

Cc: Jan Beulich <jbeulich@novell.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Merge branch 'upstream-fixes' of /linux/kernel/git/jgarzik/netdev-2.6

Merge git://git./linux/kernel/git/sam/kbuild-bugfix

[MIPS] Clear ST0_RE on bootup.

The reset state is undefined and some firmware doesn't clear this bit
possibly resulting in crashes on entry into userland.

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

[MIPS] TX49x7: Fix reporting of the CPU name and PCI clock

I've noticed that PCI clock was incorrectly reported as 66 MHz while being
mere 33 MHz on RBTX4937 board -- this was due to the different encoding of
the PCI divisor field in CCFG register between TX4927 and TX4937 chips...

Also, RBTX49x7 was printed out as a CPU name (e.g., "CPU is RBTX4937");
and some debug printk() were duplicating each other...

Signed-off-by: Konstantin Baydarov <kbaidarov@ru.mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] TX49 MFC0 bug workaround

If mfc0 $12 follows store and the mfc0 is last instruction of a
page and fetching the next instruction causes TLB miss, the result
of the mfc0 might wrongly contain EXL bit.

ERT-TX49H2-027, ERT-TX49H3-012, ERT-TX49HL3-006, ERT-TX49H4-008

Workaround: mask EXL bit of the result or place a nop before mfc0.  It
doesn't harm to always clear those bits, so we change the code to do so.

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

[MIPS] Fix linker script to work for non-4K page size.

Very much to my surprise Fuxin Zhang reports this is all it takes to get
the kernel to work for page sizes larger than 4kB.  This also paves the
way for support for the R6000 and R8000 which don't support 4kB page size.

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

[MIPS] Oprofile: Support for 34K UP kernels.

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

[MIPS] Fix dump_tlb.c warning and cleanup.

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

[MIPS] Fix minor sparse warnings

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

[MIPS] MIPS R2 optimized endianess swapping.

From Franck Bui-Huu <vagabon.xyz@gmail.com> with modifications by me.

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

[MIPS] Remove buggy inline version of memscan.

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

[MIPS] Sparse: Add some __user tags to signal functions.

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

[MIPS] Shrink Qemu configuration to the bare minimum that is need and tested.

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

[MIPS] Support /proc/kcore for MIPS

I'm pretty sure that the CKSEG0 bits are wrong, but I did need to
cover that region - because the SB-1 kernel links at 0xffffffff80100000
or so, disassembly and printing static variables don't work unless the
debugger can read that region.

Signed-off-by: Daniel Jacobowitz <dan@codesourcery.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] IP22: Fix serial console detection

From: Kaj-Michael Lang <milang@tal.org>

In ip22-setup.c the checks for serial/graphics console logic does
not check if ARCS console=g but the machine is using serial console, as
it does if no keyboard is attached.

This patch adds a check if ConsoleOut is serial. There might also be
support for other graphics than Newport soon...

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

[MIPS] ieee754[sd]p_neg workaround

It looks glibc's pow() assumes an unary '-' operation for any number
(including NaNs) always inverts its sign bit (though IEEE754 does not
specify the sign bit for NaNs).  This patch make the kernel math-emu
emulates real MIPS neg.[ds] instruction.

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

[MIPS] Remove wrong __user tags.

This fixes sparse warnings 'dereference of noderef expression'.

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

[MIPS] Cleanup fls implementation.

fls was the only called of flz, so fold flz into fls, same for the
__ilog2 call.  Delete the now unused flz function.

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

[MIPS] local_irq_restore wasn't safe to be used in other macros mode.

It always left the assembler in reorder mode possibly causing disaster.

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

[MIPS] Remove commented out code to add -mmad for Nevada.

Adding -mmad is not usable since over half a decade in gcc and when
fixed the proper -march option values should enable the use of the
mad, madu and mul instructions of the R5500, RM5200, RM7000 and RM9000
families.

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

[MIPS] Fix Cobalt PCI cache line sizes

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

[MIPS] Sparse: Add _MIPS_SZINT and _MIPS_ISA to CHECKFLAGS to fix sparse warnings.

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

[MIPS] Build blast_cache routines from template

Build blast_xxx, blast_xxx_page, blast_xxx_page_indexed from template.
Easier to maintaina and saves 300 lines.  Generated code should be
unchanged.

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

[MIPS] Sparse: Fix some compiler/sparse warnings in ptrace32.c

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

[MIPS] Reformat to 80 columns.

Patch courtesy of Emily Postnews ;-)

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

[MIPS] Sibyte: Make all setup functions __init.

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

[MIPS] Get rid of CONFIG_SB1_PASS_1_WORKAROUNDS #ifdef crapola.

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

[MIPS] Au1200: Make KGDB compile

AMD Au1200 SOC just doesn't have UART3, so KGDB won't even compile for it
as is, here's the fix to make KGDB use UART1.

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Au1xx0: really set KSEG0 to uncached on reboot

Fix a really old buglet in AMD Au1xx0 restart code: instead of
modifying the whole CP0 Config.K0 field to 010b (meaning KSEG0 uncached)
before flushing the caches and resetting a board, it only sets bit 1 of that
reg. which is effectively a NOP since Config.K0 == 011b as the kernel sets it
up (which is also its default value for Au1xx0).

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] TX49x7: Fix timer register #define's

Fix the #define's for TX4927/37 timer reg's to match the datasheets (those

Signed-off-by: Konstantin Baydarov <kbaidarov@mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@mvista.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Bullet proof uaccess.h against 4.0.1 miss-compilation.

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

[MIPS] SB1: Add oprofile support.

Signed-off-by: Mark Mason <mason@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] BCM1480: Cleanup debug code left behind in the PCI driver.

Signed-off-by: Mark Mason <mason@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] BCM1125 PCI fixes

Make BCM1125 targets to link again.

Signed-off-by: Mark Mason <mason@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

[MIPS] Nevada support for SGI O2.

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

[MIPS] CPU definitions for Cobalt.

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

[MIPS] Rename include/asm-mips/cobalt to include/asm-mips/mach-cobalt.

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

[MIPS] Rename _machine_power_off to pm_power_off so the kernel builds again.

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

[MIPS] Check function pointers are non-zero before calling.

Several boards don't initialize the pointers, so let's play safe.

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

[MIPS] Revert "mips: add pm_power_off"

pm_power_off duplicates the functionality of _machine_restart.

This reverts b142159fa5ffbad73b6927fafa5440148030f3f1 commit.

[MIPS] RTLX compile fixes.

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

[MIPS] Get rid of unnecessary prototypes.  Fixes and optimizations for HZ > 100.

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

[MIPS] Fix C version of ssnop to use the right opcode.

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

[MIPS] Remove stray .set mips3 resulting in 64-bit instruction in 32-bit kernels.

Only the NMI handler was affected so this is a low impact bug.

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

Merge branch 'work-fixes'

Merge branch 'origin'

V4L/DVB (3325): Disabled debug on by default in tvp5150

disabled debug on by default in tvp5150

Signed-off-by: Markus Rechberger <mrechberger@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3318e): DVB: remove the at76c651/tda80xx frontends

The at76c651 and tda80xx frontends are currently completely unused, IOW
their only effect is making the kernel larger for people accitentially
enabling them.

The current in-kernel drivers differ from the drivers at cvs.tuxbox.org,
and re-adding them when parts of the dbox2 project get merged should be
trivial.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3318c): fix saa7146 kobject register failure

Whoops.

kobject_register failed for hexium HV-PCI6/Orion (-13)
[<c01d3eb6>] kobject_register+0x31/0x47
[<c023a996>] bus_add_driver+0x4a/0xfd
[<c01de3c1>] __pci_register_driver+0x82/0xa4
[<d083400a>] hexium_init_module+0xa/0x47 [hexium_orion]
[<c013bdae>] sys_init_module+0x167b/0x1822
[<c01633f7>] do_sync_read+0xb8/0xf3
[<c0133fa3>] autoremove_wake_function+0x0/0x2d
[<c0145390>] audit_syscall_entry+0x118/0x13f
[<c0106ae2>] do_syscall_trace+0x104/0x14a
[<c0103d21>] syscall_call+0x7/0xb

slashes in kobject names aren't allowed.

Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3318a): Makes Some symbols static.

Some symbols at cx88-alsa were global. Making those static.

Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3313): FIX: Check if FW was downloaded or not + new firmware file

- When a firmware was downloaded dvb_usb_device_init returns NULL for the
  dvb_usb_device, then nothing should be done with that pointer and device,
  because it will re-enumerate.
- A new firmware should be used with digitv devices.
- It should make "slave"-devices work and others, too.

Signed-off-by: Patrick Boettcher <pb@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3312): FIX: Multiple usage of VP7045-based devices

Reassigning function pointers in a static led to infinite loops when using
multiple VP7045-based device at the same time on one system. Using kmalloc'd
copies for reassignments is better.

Signed-off-by: Patrick Boettcher <pb@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3310): Use MT352 parallel transport function for all Bluebird FusionHDTV DVB-T boxes.

Use the parallel transport function of the MT352 demodulator in
TH7579 and LGZ201 -based FusionHDTV Bluebird usb boxes.

Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3308): Use parallel transport for FusionHDTV Dual Digital USB

Use the parallel transport function of the MT352 in USB demodulator of the
Dual Digital board.

Signed-off-by: Chris Pascoe <c.pascoe@itee.uq.edu.au>
Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3307): Support for Galaxis DVB-S rev1.3

support for Galaxis DVB-S rev1.3 (subsystem 13c2:0004)

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3306): Fixed i2c return value, conversion mdelay to msleep

fixed i2c return value, conversion mdelay to msleep

Signed-off-by: Markus Rechberger <mrechberger@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3300): Add standard for South Korean NTSC-M using A2 audio.

South Korea uses NTSC-M but with A2 audio instead of BTSC. Several audio
chips need this information in order to set the correct audio processing
registers.

Acked-by: Mauro Carvalho Chehab <mauro_chehab@yahoo.com.br>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3297): Add IR support to KWorld DVB-T (cx22702-based)

add IR support to KWorld DVB-T (cx22702-based)

Signed-off-by: Marco Manenti <marco_manenti@colman.it>
Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3294): Fix [Bug 5895] to correct snd_87x autodetect

With DVB drivers enabled snd_87x (ALSA) don't detect.

Signed-off-by: Manu Abraham <manu@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3281): Added signal detection support to tvp5150

- added signal detection support to tvp5150

Signed-off-by: Markus Rechberger <mrechberger@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3266): Fix NICAM buzz on analog sound

Apparently, having the number of lines fixed at 4 reduces (or even kills)
the buzz found in NICAM stereo with analog sound.

Signed-off-by: Marcin Rudowski <mar_rud@poczta.onet.pl>
Signed-off-by: Ricardo Cerqueira <v4l@cerqueira.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

V4L/DVB (3299): Kconfig: DVB_USB_CXUSB depends on DVB_LGDT330X and DVB_MT352

- rename DVB_USB_CXUSB one-liner description to:
  Conexant USB2.0 hybrid reference design support.
- with the addition of bluebird support to dvb-usb-cxusb,
  it now depends on lgdt330x and mt352 modules.

Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

[PATCH] Kbuild menu - hide empty NETDEVICES menu when NET is disabled

Make the whole netdevices menu depend on NET, rather than having an empty
submenu when networking is disabled.

Indeed, almost the whole body of the menu was surrounded by if NETDEVICES,
and what was outside depended on NETCONSOLE which is inside the menu.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] uli526x warning fix

drivers/net/tulip/uli526x.c: In function `__check_mode':
drivers/net/tulip/uli526x.c:1693: warning: return from incompatible pointer type

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

Merge branch 'for-jeff' of git://electric-eye.fr.zoreil.com/home/romieu/linux-2.6

[PATCH] gianfar: Fix sparse warnings

Fixed sparse warnings mainly due to lack of __iomem.

Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] dscc4: fix dscc4_init_dummy_skb check

It returns a pointer.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] bonding: Sparse warnings fix

drivers/net/bonding/bond_sysfs.c:263:27: warning: Using plain integer as NULL pointer
drivers/net/bonding/bond_sysfs.c:998:26: warning: Using plain integer as NULL pointer
drivers/net/bonding/bond_sysfs.c:1126:26: warning: Using plain integer as NULL pointer

Signed-off-by: Luiz Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] bonding: allow bond to use TSO if slaves support it

Add NETIF_F_TSO (NETIF_F_UFO) to BOND_INTERSECT_FEATURES so that it can
be used by a bonding device iff all its slave devices support TSO (UFO).

Signed-off-by: Arthur Kepner <akepner@sgi.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sky2: version 0.15 update

Increase version, and get rid of out-dated comment.
Speed setting has worked for quite a while.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sky2: support msi interrupt (revised)

This hardware supports Message Signaled interrupts.
When setting up, use software interrupt to check for bad hardware.

Signed-off-by: Stephen Hemminger <shemminger @osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sky2: add irq to entropy pool

The sky2 interrupt can be used to add entropy.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sky2: clear irq race

Move the interrupt clear to before processing, this avoids a
possible races with status delaying.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sky2: set mac address fix

Using the sky2 driver with bonding can result in oopses related to
reinitializing the PHY when the MAC address is changed (which bonding
is wont to do).  This patch changes sky2_set_mac_address to take less
drastic measures.

This is analagous to the skge patch here:

http://lkml.org/lkml/2005/9/29/399

which fixed the issue here:

http://bugzilla.kernel.org/show_bug.cgi?id=5271

Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sky2: ethtool rx_coalesce settings fix

This fixes setting rx_coalesce_usecs_irq via ethtool in sky2.
The write was directed to the wrong register.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sky2: pci config space checking

There were bugs in mmconfig access to PCI space, up to and
include 2.6.16-rc1. These prevented the sky2 driver from being
able to clear PCI express errors.

This patch makes the driver check (during probe), for errors
in PCI config access and fail.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sky2: power management fix

Fix suspend/resume for sky2. The status ring was getting reallocated
and a bunch of other mistakes. Also, check return from power_state
on resume.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

[PATCH] sis900: remove cfgpmcsr I/O space register define

sis900 defines 'cfgpmcsr' as an I/O space register, but CFGPMCSR is
in fact a config space register, and there is no register at offset
0x44 in I/O space, so delete the enum.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

e100: remove init_hw call to fix panic

e100 seems to have had a long standing bug where e100_init_hw was being
called when it should not have been.  This caused a panic due to recent
changes that rely on correct set up in the driver, and more robust error
paths.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

[IA64-SGI] Shub2 BTE address fix

After converting the cpu physical address to shub2 physical
addressing, the address was run through TO_PHYS() which
clobbered a high node offset bit causing the BTE to fail
on shub2 nodes with large memory.  This fix corrects
that problem.

Signed-off-by: Russ Anderson (rja@sgi.com)
Signed-off-by: Tony Luck <tony.luck@intel.com>

Pull sn-recursive-flags-for-select-builds into release branch

Merge branch 'origin'

Merge branch 'origin'

[PPP]: Fixed hardware RX checksum handling

When we pull the PPP protocol off the skb, we forgot to update the
hardware RX checksum.  This may lead to messages such as

dsl0: hw csum failure.

Similarly, we need to clear the hardware checksum flag when we use
the existing packet to store the decompressed result.

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

[PATCH] USB: Fix GPL markings on usb core functions.

I thought we had fixed up all non-gpl USB drivers, and was wrong to do
this.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

mm/slab.c (non-NUMA): Fix compile warning and clean up code

The non-NUMA case would do an unmatched "free_alien_cache()" on an alien
pointer that had never been allocated.

It might not matter from a code generation standpoint (since in the
non-NUMA case, the code doesn't actually _do_ anything), but it not only
results in a compiler warning, it's really really ugly too.

Fix the compiler warning by just having a matching dummy allocation.
That also avoids an unnecessary #ifdef in the code.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Merge /pub/scm/linux/kernel/git/davem/sparc-2.6