The wrappers aes_encrypt/aes_decrypt simply reverse the order of the
function arguments.  It's just as easy to get the actual assembly code
to read them in the opposite order.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>

This reverts commits

  3e3318de [PATCH] swsusp: x86_64 mark special saveable/unsaveable pages
  b6370d96 [PATCH] swsusp: i386 mark special saveable/unsaveable pages
  ce4ab001 [PATCH] swsusp: add architecture special saveable pages support

because not only do they apparently cause page faults on x86, the
infrastructure doesn't compile on powerpc.
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Copy the softirq bits in preempt_count from the current context into the
hardirq context when using 4K stacks to make the softirq_count macro work
correctly and thereby fix softirq cpu time accounting.
Signed-off-by: NBjörn Steinbrink <B.Steinbrink@gmx.de>
Acked-by: NArjan van de Ven <arjan@linux.intel.com>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

As described in a previous patch and documented in mm/filemap.h,
copy_from_user_inatomic* shouldn't zero out the tail of the buffer after an
incomplete copy.

This patch implements that change for i386.

For the _nocache version, a new __copy_user_intel_nocache is defined similar
to copy_user_zeroio_intel_nocache, and this is ultimately used for the copy.

For the regular version, __copy_from_user_ll_nozero is defined which uses
__copy_user and __copy_user_intel - the later needs casts to reposition the
__user annotations.

If copy_from_user_atomic is given a constant length of 1, 2, or 4, then we do
still zero the destintion on failure.  This didn't seem worth the effort of
fixing as the places where it is used really don't care.
Signed-off-by: NNeil Brown <neilb@suse.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: William Lee Irwin III <wli@holomorphy.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Add cpu_relax() to infinite loops in crash.c and doublefault.c.  This is
the safest change.
Signed-off-by: NChuck Ebbert <76306.1226@compuserve.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Add cpu_relax() to various smpboot.c init loops.  cpu_relax() always implies a
barrier (according to Arjan), so remove those as well.
Signed-off-by: NAndreas Mohr <andi@lisas.de>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Clean up and refactor i386 sub-architecture setup.

This change moves all the code from the
asm-i386/mach-*/setup_arch_pre/post.h headers, into
arch/i386/mach-*/setup.c.  mach-*/setup_arch_pre.h is renamed to
setup_arch.h, and contains only things which should be in header files.  It
is purely code-motion; there should be no functional changes at all.

Several functions in arch/i386/kernel/setup.c needed to be made non-static
so that they're visible to the code in mach-*/setup.c.  asm-i386/setup.h is
used to hold the prototypes for these functions.
Signed-off-by: NJeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: NChris Wright <chrisw@sous-sol.org>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Chris Wright <chrisw@sous-sol.org>
Cc: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Cc: Martin Bligh <mbligh@google.com>
Cc: James Bottomley <James.Bottomley@steeleye.com>
Cc: Andrey Panin <pazke@donpac.ru>
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Remove VM_LOCKED before remap_pfn range from device drivers and get rid of
VM_SHM.

remap_pfn_range() already sets VM_IO.  There is no need to set VM_SHM since
it does nothing.  VM_LOCKED is of no use since the remap_pfn_range does not
place pages on the LRU.  The pages are therefore never subject to swap
anyways.  Remove all the vm_flags settings before calling remap_pfn_range.

After removing all the vm_flag settings no use of VM_SHM is left.  Drop it.
Signed-off-by: NChristoph Lameter <clameter@sgi.com>
Acked-by: NHugh Dickins <hugh@veritas.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Convert a few stragglers over to for_each_possible_cpu(), remove
for_each_cpu().
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Considering that there isn't a lot of hw we can depend on during resume,
this is about as good as it gets.

This is x86-only for now, although the basic concept (and most of the
code) will certainly work on almost any platform.
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

list_splice_init(list, head) does unneeded job if it is known that
list_empty(head) == 1.  We can use list_replace_init() instead.
Signed-off-by: NOleg Nesterov <oleg@tv-sign.ru>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The hardirq_ctx and softirq_ctx variables are written to on init only,
Signed-off-by: NAndreas Mohr <andi@lisas.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Default values for boolean and tristate options can only be 'y', 'm' or 'n'.
This patch removes wrong default for SCHED_SMT.
Signed-off-by: NJean-Luc Leger <jean-luc.leger@dspnet.fr.eu.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Move do_suspend_lowlevel to correct segment.  If it is in the same hugepage
with ro data, mark_rodata_ro will make it unexecutable.
Signed-off-by: NShaohua Li <shaohua.li@intel.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

flush_tlb_all uses on_each_cpu, which will disable/enable interrupt.
In suspend/resume time, this will make interrupt wrongly enabled.
Signed-off-by: NShaohua Li <shaohua.li@intel.com>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Pages (Reserved/ACPI NVS/ACPI Data) below max_low_pfn will be saved/restored
by S4 currently.  We should mark 'Reserved' pages not saveable.

Pages (Reserved/ACPI NVS/ACPI Data) above max_low_pfn will not be
saved/restored by S4 currently.  We should save the 'ACPI NVS/ACPI Data'
pages.
Signed-off-by: NShaohua Li <shaohua.li@intel.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Nigel Cunningham <nigel@suspend2.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

New CPU flags for next generation of crypto engine as found in VIA C7
processors.
Signed-off-by: NMichal Ludvig <michal@logix.cz>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Sometimes thread_info and task_struct get out-of-sync with each other.
Printing task.thread_info in show_registers() can help spot this.  And when
task_struct is corrupt then task.comm can contain garbage, so only print as
many characters as it can hold.
Signed-off-by: NChuck Ebbert <76306.1226@compuserve.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Never allow int3 traps from V8086 mode to enter the kprobes handler.
Signed-off-by: NZachary Amsden <zach@vmware.com>
Cc: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Chuck Ebbert <76306.1226@compuserve.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

We need to check for vm86 mode first before looking at selector privilege
bits.

Segment limit is always base + 64k and only the low 16 bits of EIP are
significant in vm86 mode.
Signed-off-by: NChuck Ebbert <76306.1226@compuserve.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Rohit Seth <rohitseth@google.com>
Acked-by: NAnanth N Mavinakayanahalli <ananth@in.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Use proper defines instead of open-coded values.
Signed-off-by: NAndreas Mohr <andi@lisas.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

constify structs and add one __initdata.
Signed-off-by: NAndreas Mohr <andi@lisas.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAndreas Mohr <andi@lisas.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

PCI code was outside of CONFIG_PCI, add __initdata at cyrix_55x0 (since
accessed within __init function only).
Signed-off-by: NAndreas Mohr <andi@lisas.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAndreas Mohr <andi@lisas.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The i386 page fault handler does not allow enough slack when checking for
userspace access below the current stack pointer.  This prevents use of the
enter instruction by user code.  Fix this by allowing enough slack for
"enter $65535,$31" to execute.

Problem reported by Tomasz Malesinski <tmal@mimuw.edu.pl>

Tested using this program, based on the original from Tomasz:

	.file	"ovflow.S"
	.version	"01.01"
gcc2_compiled.:
.section	.rodata
.LC0:
	.string	"asdf\n"
.text
	.align 4
.globl main
	.type	 main,@function
main:
nest_level=0
.rept 30
	enter $0,$nest_level
nest_level=nest_level+1
.endr
	enter $65535,$30
	enter $65535,$31
	addl $-12,%esp
	pushl $.LC0
	call printf
	addl $16,%esp
.L2:
.rept 32
	leave
.endr
	ret
.Lfe1:
	.size	 main,.Lfe1-main
	.ident	"GCC: (GNU) 2.95.4 20011002 (Debian prerelease)"
Signed-off-by: NChuck Ebbert <76306.1226@compuserve.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

On i386, kernel irq balance doesn't work.

1) In function do_irq_balance, after kernel finds the min_loaded cpu but
   before calling set_pending_irq to really pin the selected_irq to the
   target cpu, kernel does a cpus_and with irq_affinity[selected_irq].
   Later on, when the irq is acked, kernel would calls
   move_native_irq=>desc->handler->set_affinity to change the irq affinity.
    However, every function pointed by
   hw_interrupt_type->set_affinity(unsigned int irq, cpumask_t cpumask)
   always changes irq_affinity[irq] to cpumask.  Next time when recalling
   do_irq_balance, it has to do cpu_ands again with
   irq_affinity[selected_irq], but irq_affinity[selected_irq] already
   becomes one cpu selected by the first irq balance.

2) Function balance_irq in file arch/i386/kernel/io_apic.c has the same
   issue.

[akpm@osdl.org: cleanups]
Signed-off-by: NZhang Yanmin <yanmin.zhang@intel.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAlexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

If CONFIG_FRAME_POINTERS is enabled, and one does a dump_stack() during
early SMP init, an infinite stackdump and a bootup hang happens:

 [<c0104e7f>] show_trace+0xd/0xf
 [<c0104e96>] dump_stack+0x15/0x17
 [<c01440df>] save_trace+0xc3/0xce
 [<c014527d>] mark_lock+0x8c/0x4fe
 [<c0145df5>] __lockdep_acquire+0x44e/0xaa5
 [<c0146798>] lockdep_acquire+0x68/0x84
 [<c1048699>] _spin_lock+0x21/0x2f
 [<c010d918>] prepare_set+0xd/0x5d
 [<c010daa8>] generic_set_all+0x1d/0x201
 [<c010ca9a>] mtrr_ap_init+0x23/0x3b
 [<c010ada8>] identify_cpu+0x2a7/0x2af
 [<c01192a7>] smp_store_cpu_info+0x2f/0xb4
 [<c01197d0>] start_secondary+0xb5/0x3ec
 [<c104ec11>] end_of_stack_stop_unwind_function+0x1/0x4
 [<c104ec11>] end_of_stack_stop_unwind_function+0x1/0x4
 [<c104ec11>] end_of_stack_stop_unwind_function+0x1/0x4
 [<c104ec11>] end_of_stack_stop_unwind_function+0x1/0x4
 [<c104ec11>] end_of_stack_stop_unwind_function+0x1/0x4
 [<c104ec11>] end_of_stack_stop_unwind_function+0x1/0x4
 [<c104ec11>] end_of_stack_stop_unwind_function+0x1/0x4
 [<c104ec11>] end_of_stack_stop_unwind_function+0x1/0x4
 [...]

Due to "end_of_stack_stop_unwind_function" recursing back to itself in the
EBP stackframe-walker.  So avoid this type of recursion when walking the
stack .
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

When multiple updates matching a given CPU are found in the update file, the
action taken by the microcode update driver was inappropriate:

- when lower revision microcode was found before matching or higher revision
  one, the driver would needlessly complain that it would not downgrade the
  CPU
- when microcode matching the currently installed revision was found before
  newer revision code, no update would actually take place

To change this behavior, the driver now concludes about possibly updates and
issues messages only when the entire input was parsed.

Additionally, this adds back (in different places, and conditionalized upon
a new module option) some messages removed by a previous patch.
Signed-off-by: NJan Beulich <jbeulich@novell.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Tigran Aivazian <tigran_aivazian@symantec.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Only drm, framebuffer, mtrr parts + misc files here and there.
Signed-off-by: NAlexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

- avoid expensive modulo (integer division) which happened
  since APM_MAX_EVENTS is 20 (non-power-of-2)
- kill compiler warnings by initializing two variables
- add __read_mostly to some important static variables that are read often
  (by idle loop etc.)
- constify several structures
Signed-off-by: NAndreas Mohr <andi@lisas.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAdrian Bunk <bunk@stusta.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Use the x86 cache-bypassing copy instructions for copy_from_user().

Some performance data are

Total of GLOBAL_POWER_EVENTS (CPU cycle samples)

2.6.12.4.orig    1921587
2.6.12.4.nt      1599424
1599424/1921587=83.23% (16.77% reduction)

BSQ_CACHE_REFERENCE (L3 cache miss)
2.6.12.4.orig      57427
2.6.12.4.nt        20858
20858/57427=36.32% (63.7% reduction)

L3 cache miss reduction of __copy_from_user_ll
samples  %
37408    65.1412  vmlinux                  __copy_from_user_ll
23        0.1103  vmlinux                  __copy_user_zeroing_intel_nocache
23/37408=0.061% (99.94% reduction)

Top 5 of 2.6.12.4.nt
Counted GLOBAL_POWER_EVENTS events (time during which processor is not stopped) with a unit mask of 0x01 (mandatory) count 100000
samples  %        app name                 symbol name
128392    8.0274  vmlinux                  __copy_user_zeroing_intel_nocache
64206     4.0143  vmlinux                  journal_add_journal_head
59746     3.7355  vmlinux                  do_get_write_access
47674     2.9807  vmlinux                  journal_put_journal_head
46021     2.8774  vmlinux                  journal_dirty_metadata
pattern9-0-cpu4-0-09011728/summary.out

Counted BSQ_CACHE_REFERENCE events (cache references seen by the bus unit) with a unit mask of 0x3f (multiple flags) count 3000
samples  %        app name                 symbol name
69755     4.2861  vmlinux                  __copy_user_zeroing_intel_nocache
55685     3.4215  vmlinux                  journal_add_journal_head
52371     3.2179  vmlinux                  __find_get_block
45504     2.7960  vmlinux                  journal_put_journal_head
36005     2.2123  vmlinux                  journal_stop
pattern9-0-cpu4-0-09011744/summary.out

Counted BSQ_CACHE_REFERENCE events (cache references seen by the bus unit) with a unit mask of 0x200 (read 3rd level cache miss) count 3000
samples  %        app name                 symbol name
1147      5.4994  vmlinux                  journal_add_journal_head
881       4.2240  vmlinux                  journal_dirty_data
872       4.1809  vmlinux                  blk_rq_map_sg
734       3.5192  vmlinux                  journal_commit_transaction
617       2.9582  vmlinux                  radix_tree_delete
pattern9-0-cpu4-0-09011731/summary.out

iozone results are

original 2.6.12.4 CPU time = 207.768 sec
cache aware       CPU time = 184.783 sec
(three times run)
184.783/207.768=88.94% (11.06% reduction)

original:
pattern9-0-cpu4-0-08191720/iozone.out:  CPU Utilization: Wall time   45.997    CPU time   64.527    CPU utilization 140.28 %
pattern9-0-cpu4-0-08191741/iozone.out:  CPU Utilization: Wall time   46.878    CPU time   71.933    CPU utilization 153.45 %
pattern9-0-cpu4-0-08191743/iozone.out:  CPU Utilization: Wall time   45.152    CPU time   71.308    CPU utilization 157.93 %

cache awre:
pattern9-0-cpu4-0-09011728/iozone.out:  CPU Utilization: Wall time   44.842    CPU time   62.465    CPU utilization 139.30 %
pattern9-0-cpu4-0-09011731/iozone.out:  CPU Utilization: Wall time   44.718    CPU time   59.273    CPU utilization 132.55 %
pattern9-0-cpu4-0-09011744/iozone.out:  CPU Utilization: Wall time   44.367    CPU time   63.045    CPU utilization 142.10 %
Signed-off-by: NHiro Yoshioka <hyoshiok@miraclelinux.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

sys_move_pages() support for 32bit (i386 plus x86_64 compat layer)

Add support for move_pages() on i386 and also add the compat functions
necessary to run 32 bit binaries on x86_64.

Add compat_sys_move_pages to the x86_64 32bit binary layer.  Note that it is
not up to date so I added the missing pieces.  Not sure if this is done the
right way.

[akpm@osdl.org: compile fix]
Signed-off-by: NChristoph Lameter <clameter@sgi.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Consolidate the various arch-specific implementations of pxm_to_node() and
node_to_pxm() into a single generic version.
Signed-off-by: NYasunori Goto <y-goto@jp.fujitsu.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Fix powernow-k8 doesn't load bug.
Reference: https://launchpad.net/distros/ubuntu/+source/linux-source-2.6.15/+bug/35145Signed-off-by: NBen Collins <bcollins@ubuntu.com>
Signed-off-by: NDave Jones <davej@redhat.com>

I haven't really maintained this driver for a while, and I'm not
keeping up with the latest in Intel power management.  I get a steady
stream of mail which I don't really do anything useful with; the
cpufreq list seems like a better destination, unless someone wants to
get the mail directly.

Also clean up a couple of ancient comments which don't really apply
anymore (as far as I know, nobody has ever damaged a CPU with this
driver).
Signed-off-by: NJeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: NDave Jones <davej@redhat.com>

On 15 Jun 2006 03:45:10 +0200, Andi Kleen wrote:

> Anyways I would say that if the BIOS can't get MCFG right then
> it's likely not been validated on that board and shouldn't be used.

According to Petr Vandrovec:

 ... "What is important (and checked) is address of MMCONFIG reported by MCFG
 table...  Unfortunately code does not bother with printing that address :-(

 "Another problem is that code has hardcoded that MMCONFIG area is 256MB large.
 Unfortunately for the code PCI specification allows any power of two between 2MB
 and 256MB if vendor knows that such amount of busses (from 2 to 128) will be
 sufficient for system.  With notebook it is quite possible that not full 8 bits
 are implemented for MMCONFIG bus number."

So here is a patch.  Unfortunately my system still fails the test because
it doesn't reserve any part of the MMCONFIG area, but this may fix others.

Booted on x86_64, only compiled on i386.  x86_64 still remaps the max area
(256MB) even though only 2MB is checked... but 2.6.16 had no check at all
so it is still better.

PCI: reduce size of x86 MMCONFIG reserved area check

1.  Print the address of the MMCONFIG area when the test for that area
    being reserved fails.

2.  Only check if the first 2MB is reserved, as that is the minimum.
Signed-off-by: NChuck Ebbert <76306.1226@compuserve.com>
Acked-by: NArjan van de Ven <arjan@linux.intel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>