07 Jun, 2008
40 commits
-
When posting:
[PATCH 1/8] Scaling msgmni to the amount of lowmem
(see http://lkml.org/lkml/2008/2/11/171), I have added a KERN_INFO message
that is output each time msgmni is recomputed.In http://lkml.org/lkml/2008/4/29/575 Tony Luck complained that this
message references an ipc namespace address that is useless.I first thought of using an audit_log instead of a printk, as suggested by
Serge Hallyn. But unfortunately, we do not have any other information
than the namespace address to provide here too. So I chose to move the
message and output it only at boot time, removing the reference to the
namespace.Signed-off-by: Nadia Derbey
Cc: Pierre Peiffer
Cc: Manfred Spraul
Acked-by: Tony Luck
Cc: "Serge E. Hallyn"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
When posting:
[PATCH 1/8] Scaling msgmni to the amount of lowmem
(see http://article.gmane.org/gmane.linux.kernel/637849/) I changed the
MSGPOOL value to make it fit what is said in the man pages (i.e. a size
in bytes).But Michael Kerrisk rightly complained that this change could affect the
ABI. So I'm posting this patch to make MSGPOOL expressed back in Kbytes.
Michael, on his side, has fixed the man page.Signed-off-by: Nadia Derbey
Cc: Pierre Peiffer
Cc: Manfred Spraul
Acked-by: Michael Kerrisk
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Original report: """I used to force my console to black-on-white by the
command `setterm -inversescreen on`. In 2.6.26-rc4, I get lots of black
background characters."""Another addendum to commit c9e587ab. This was previously missed out since
I was not aware of what vc_decscnm was for.Signed-off-by: Jan Engelhardt
Reported-by:
Tested-by:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
If cpu specific cpufreq driver(i.e. longrun) has "setpolicy" function,
governor object isn't set into cpufreq_policy object at "__cpufreq_set_policy"
function in driver/cpufreq/cpufreq.c .This causes a null object access at "store_scaling_setspeed" and
"show_scaling_setspeed" function in driver/cpufreq/cpufreq.c when reading or
writing through /sys interface (ex. cat
/sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed)Addresses:
http://bugzilla.kernel.org/show_bug.cgi?id=10654
https://bugzilla.redhat.com/show_bug.cgi?id=443354Signed-off-by: CHIKAMA Masaki
Cc: Dave Jones
Cc: Chuck Ebbert
Acked-by: Dominik Brodowski
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The interlaced and double line mode bits should not be copied to new
console when the console is switched. Otherwise, the new console may be
set to incorrect refresh rate.Also, the x and y offsets does not need to be copied.
Signed-off-by: Krzysztof Helt
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This semaphore doesn't appear to be used, so remove it.
Signed-off-by: Daniel Walker
Cc: David Howells
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Adding a nonexistent cpu to a cpuset will be omitted quietly. It should
return -EINVAL.Example: (real_nr_cpus < NR_CPUS or cpu#4 was just offline)
# cat cpus
0-1
# /bin/echo 4 > cpus
# /bin/echo $?
0
# cat cpus#
The same occurs when add a nonexistent mem.
This patch will fix this bug.
And when *buf == "", the check is unneeded.Signed-off-by: Lai Jiangshan
Acked-by: Paul Jackson
Cc: David Rientjes
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix a bug in add_to_pagemap. Previously, since pm->out was a char *,
put_user was only copying 1 byte of every PFN, resulting in the top 7
bytes of each PFN not being copied. By requiring that reads be a multiple
of 8 bytes, I can make pm->out and pm->end u64*s instead of char*s, which
makes put_user work properly, and also simplifies the logic in
add_to_pagemap a bit.[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Thomas Tuttle
Cc: Matt Mackall
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Consider you added a 'c foo:bar r' permission to some cgroup and then (a
bit later) 'c'foo:bar w' for it. After this you'll see thec foo:bar r
c foo:bar wlines in a devices.list file.
Another example - consider you added 10 'c foo:bar r' permissions to some
cgroup (e.g. by mistake). After this you'll see 10 c foo:bar r lines in
a list file.This is weird. This situation also has one more annoying consequence.
Having many items in a white list makes permissions checking slower, sine
it has to walk a longer list.The proposal is to merge permissions for items, that correspond to the
same device.Signed-off-by: Pavel Emelyanov
Acked-by: Serge Hallyn
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Currently even if a task sits in an all-denied cgroup it can still mount
any block device in any mode it wants.Put a proper check in do_open for block device to prevent this.
Signed-off-by: Pavel Emelyanov
Acked-by: Serge Hallyn
Tested-by: Serge Hallyn
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Two functions, that need to get a device_cgroup from a task (they are
devcgroup_inode_permission and devcgroup_inode_mknod) make it in a strange
way:They get a css_set from task, then a subsys_state from css_set, then a
cgroup from the state and then a subsys_state again from the cgroup.
Besides, the devices_subsys_id is read from memory, whilst there's a
enum-ed constant for it.Optimize this part a bit:
1. Get the subsys_stats form the task and be done - no 2 extra
dereferences,
2. Use the device_subsys_id constant, not the value from memory
(i.e. one less dereference).Found while preparing 2.6.26 OpenVZ port.
Signed-off-by: Pavel Emelyanov
Acked-by: Serge Hallyn
Acked-by: Paul Menage
Cc: Balbir Singh
Cc: James Morris
Cc: Chris Wright
Cc: Stephen Smalley
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This is just picking the container_of out of cgroup_to_devcgroup into a
separate function.This new css_to_devcgroup will be used in the 2nd patch.
Signed-off-by: Pavel Emelyanov
Acked-by: Serge Hallyn
Cc: Paul Menage
Cc: Balbir Singh
Cc: James Morris
Cc: Chris Wright
Cc: Stephen Smalley
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch introduces memory_read_from_buffer().
The only difference between memory_read_from_buffer() and
simple_read_from_buffer() is which address space the function copies to.simple_read_from_buffer copies to user space memory.
memory_read_from_buffer copies to normal memory.Signed-off-by: Akinobu Mita
Cc: Al Viro
Cc: Doug Warzecha
Cc: Zhang Rui
Cc: Matt Domsch
Cc: Abhay Salunke
Cc: Greg Kroah-Hartman
Cc: Markus Rechberger
Cc: Kay Sievers
Cc: Bob Moore
Cc: Thomas Renninger
Cc: Len Brown
Cc: Benjamin Herrenschmidt
Cc: "Antonino A. Daplas"
Cc: Krzysztof Helt
Cc: Geert Uytterhoeven
Cc: Martin Schwidefsky
Cc: Heiko Carstens
Cc: Peter Oberparleiter
Cc: Michael Holzheu
Cc: Brian King
Cc: James E.J. Bottomley
Cc: Andrew Vasquez
Cc: Seokmann Ju
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Bluetooth will be able to use this.
Signed-off-by: Harvey Harrison
Cc: Marcel Holtmann
Cc: Dave Young
Cc: Akinobu Mita
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Change the name of the device from "rtc-ds1374" to just "ds1374", to match
what all other RTC drivers do. I seem to remember that this name was
chosen to avoid possible confusion with an older ds1374 driver, but that
driver was removed 3 months ago.Signed-off-by: Jean Delvare
Signed-off-by: Alessandro Zummo
Acked-by: Kumar Gala
Cc: Paul Mackerras
Cc: Benjamin Herrenschmidt
Cc: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Although if people have questions about ARCnet, perhaps it's _better_
for them to be mailing dwmw2@cam.ac.uk about it...Signed-off-by: David Woodhouse
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Include limits.h to get a definition of PATH_MAX.
Signed-off-by: Ingo Molnar
Signed-off-by: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We lost the marking of SIGWINCH as being OK to receive during stub
execution, causing a panic should that happen.Cc: Benedict Verheyen
Signed-off-by: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
x86_64 defines either memcpy or __memcpy depending on the gcc version, and
it looks like UML needs to follow that in its exporting.Cc: Gabriel C
Signed-off-by: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch makes os_get_task_size locate the bottom of the address space,
as well as the top. This is for systems which put a lower limit on mmap
addresses. It works by manually scanning pages from zero onwards until a
valid page is found.Because the bottom of the address space may not be zero, it's not
sufficient to assume the top of the address space is the size of the
address space. The size is the difference between the top address and
bottom address.[jdike@addtoit.com: changed the name to reflect that this function is
supposed to return the top of the process address space, not its size and
changed the return value to reflect that. Also some minor formatting
changes]
Signed-off-by: Tom Spink
Signed-off-by: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Removed duplicated include file "kern_util.h" in
arch/um/drivers/ubd_kern.c.Signed-off-by: Huang Weiyi
Signed-off-by: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Protection against the host's time going backwards (eg, ntp activity on
the host) by keeping track of the time at the last tick and if it's
greater than the current time, keep time stopped until the host catches
up.Cc: Nix
Signed-off-by: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Update status and URL for the "Gary's Encyclopedia" entry.
Signed-off-by: Paul Jackson
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Dan Williams
Cc: Jonathan Woithe
Cc: Len Brown
Cc: Greg KH
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
kobjsize() has been abusing page->index as a method for sorting out
compound order, which blows up both for page cache pages, and SLOB's
reuse of the index in struct slob_page.Presently we are not able to accurately size arbitrary pointers that
don't come from kmalloc(), so the best we can do is sort out the
compound order from the head page if it's a compound page, or default
to 0-order if it's impossible to ksize() the object.Obviously this leaves quite a bit to be desired in terms of object
sizing accuracy, but the behaviour is unchanged over the existing
implementation, while fixing the page->index oopses originally reported
here:http://marc.info/?l=linux-mm&m=121127773325245&w=2
Accuracy could also be improved by having SLUB and SLOB both set PG_slab
on ksizeable pages, rather than just handling the __GFP_COMP cases
irregardless of the PG_slab setting, as made possibly with Pekka's
patches:http://marc.info/?l=linux-kernel&m=121139439900534&w=2
http://marc.info/?l=linux-kernel&m=121139440000537&w=2
http://marc.info/?l=linux-kernel&m=121139440000540&w=2This is primarily a bugfix for nommu systems for 2.6.26, with the aim
being to gradually kill off kobjsize() and its particular brand of
object abuse entirely.Reviewed-by: Pekka Enberg
Signed-off-by: Paul Mundt
Acked-by: David Howells
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
As some m68k machines have plenty of libc5 binaries in active use, enable
CONFIG_COMPAT_BRK by default.Signed-off-by: Geert Uytterhoeven
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix a regression introduced by
commit 4cc6028d4040f95cdb590a87db478b42b8be0508
Author: Jiri Kosina
Date: Wed Feb 6 22:39:44 2008 +0100brk: check the lower bound properly
The check in sys_brk() on minimum value the brk might have must take
CONFIG_COMPAT_BRK setting into account. When this option is turned on
(i.e. we support ancient legacy binaries, e.g. libc5-linked stuff), the
lower bound on brk value is mm->end_code, otherwise the brk start is
allowed to be arbitrarily shifted.Signed-off-by: Jiri Kosina
Tested-by: Geert Uytterhoeven
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
As commit 6089093e588ee3f6aed99d08b1cf5ea37c52cf97 ("ip2: fix crashes on
load/unload") fixed the ip2 crashes on load/unload by making ip2/ip2main
one module (ip2), Kconfig shouldn't mention a now non-existing module.Signed-off-by: Roland.Kletzing
Acked-by: Alan Cox
Cc: Jiri Slaby
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Alexey Dobriyan
Cc: Matt Domsch
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The page decrypt calls in ecryptfs_write() are both pointless and buggy.
Pointless because ecryptfs_get_locked_page() has already brought the page
up to date, and buggy because prior mmap writes will just be blown away by
the decrypt call.This patch also removes the declaration of a now-nonexistent function
ecryptfs_write_zeros().Thanks to Eric Sandeen and David Kleikamp for helping to track this
down.Eric said:
fsx w/ mmap dies quickly ( < 100 ops) without this, and survives
nicely (to millions of ops+) with it in place.Signed-off-by: Michael Halcrow
Cc: Eric Sandeen
Cc: Dave Kleikamp
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This version is a bit of a whopper. This version brings a few new checks,
improvements to a number of checks mostly through modifications to the
way types are parsed, several fixes to quote/comment handling, as well as
the usual slew of fixes for false positives.Of note:
- return is not a function and is now reported,
- preprocessor directive detection is loosened to match C99 standard,
- we now intuit new type modifiers, and
- comment handling is much improvedAndy Whitcroft (18):
Version: 0.19
fix up a couple of missing newlines in reports
colon to parenthesis spacing varies on asm
values: #include is a preprocessor statement
quotes: fix single character quotes at line end
add typedef exception for the non-pointer "function types"
kerneldoc parameters must be on one line, relax line length
types: word boundary is not always required
improved #define bracketing reports
uninitialized_var is an annotation not a function name
possible types: add possible modifier handling
possible types: fastcall is a type modifier
types: unsigned is not a modifier on all types
static/external initialisation to zero should allow modifiers
checkpatch: fix recognition of preprocessor directives -- part 2
comments: fix inter-hunk comment tracking
return is not a function
do not report include/asm/foo.h use in include/linux/foo.h
return is not a function -- tighten test[jengelh@computergmbh.de: fix recognition of preprocessor directives]
Signed-off-by: Andy Whitcroft
Cc: Jan Engelhardt
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Some drivers are using too generic "serial" name for driver_name, this
might cause issues, like this:Freescale QUICC Engine UART device driver
proc_dir_entry 'serial' already registered
Call Trace:
[cf82de50] [c0007f7c] show_stack+0x4c/0x1ac (unreliable)
[cf82de90] [c00b03fc] proc_register+0xfc/0x1ac
[cf82dec0] [c00b05c8] create_proc_entry+0x60/0xac
[cf82dee0] [c00b23dc] proc_tty_register_driver+0x60/0x98
[cf82def0] [c016dbd8] tty_register_driver+0x1b4/0x228
[cf82df20] [c0184d70] uart_register_driver+0x144/0x194
[cf82df40] [c030a378] ucc_uart_init+0x2c/0x94
[cf82df50] [c02f21a0] kernel_init+0x98/0x27c
[cf82dff0] [c000fa74] kernel_thread+0x44/0x60^^ The board is using ucc_uart.c and 8250.c, both registered as
"serial".This patch fixes two drivers that are using "serial" for driver_name and
not "ttyS" for dev_name. Drivers that are using "ttyS" for dev_name, will
conflict anyway, so we don't bother with these.Signed-off-by: Anton Vorontsov
Acked-by: Alan Cox
Acked-By: Timur Tabi
Acked-by: Maciej W. Rozycki
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
=============================================
[ INFO: possible recursive locking detected ]
2.6.26-rc4 #30
---------------------------------------------
heap-overflow/2250 is trying to acquire lock:
(&mm->page_table_lock){--..}, at: [] .copy_hugetlb_page_range+0x108/0x280but task is already holding lock:
(&mm->page_table_lock){--..}, at: [] .copy_hugetlb_page_range+0xfc/0x280other info that might help us debug this:
3 locks held by heap-overflow/2250:
#0: (&mm->mmap_sem){----}, at: [] .dup_mm+0x134/0x410
#1: (&mm->mmap_sem/1){--..}, at: [] .dup_mm+0x144/0x410
#2: (&mm->page_table_lock){--..}, at: [] .copy_hugetlb_page_range+0xfc/0x280stack backtrace:
Call Trace:
[c00000003b2774e0] [c000000000010ce4] .show_stack+0x74/0x1f0 (unreliable)
[c00000003b2775a0] [c0000000003f10e0] .dump_stack+0x20/0x34
[c00000003b277620] [c0000000000889bc] .__lock_acquire+0xaac/0x1080
[c00000003b277740] [c000000000089000] .lock_acquire+0x70/0xb0
[c00000003b2777d0] [c0000000003ee15c] ._spin_lock+0x4c/0x80
[c00000003b277870] [c0000000000cf2e8] .copy_hugetlb_page_range+0x108/0x280
[c00000003b277950] [c0000000000bcaa8] .copy_page_range+0x558/0x790
[c00000003b277ac0] [c000000000050fe0] .dup_mm+0x2d0/0x410
[c00000003b277ba0] [c000000000051d24] .copy_process+0xb94/0x1020
[c00000003b277ca0] [c000000000052244] .do_fork+0x94/0x310
[c00000003b277db0] [c000000000011240] .sys_clone+0x60/0x80
[c00000003b277e30] [c0000000000078c4] .ppc_clone+0x8/0xcFix is the same way that mm/memory.c copy_page_range does the
lockdep annotation.Acked-by: KOSAKI Motohiro
Acked-by: Adam Litke
Acked-by: Nishanth Aravamudan
Signed-off-by: Nick Piggin
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The temporary structure for calculated CVT mode is not initialized. Few
fields have only bits or-ed or and-ed so they may be left in incorrect
(random) state.Testing of the tridentfb seems like a good exercise for the fbdev layer.
Signed-off-by: Krzysztof Helt
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix the following compile error:
CC fs/binfmt_flat.o
In file included from
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:36:
/home/bunk/linux/kernel-2.6/git/linux-2.6/include/linux/flat.h:14:22: error: asm/flat.h: No such file or directory
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c: In function 'create_flat_tables':
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:124: error: implicit declaration of function 'flat_stack_align'
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:125: error: implicit declaration of function 'flat_argvp_envp_on_stack'
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c: In function 'calc_reloc':
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:347: error: implicit declaration of function 'flat_reloc_valid'
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c: In function 'load_flat_file':
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:479: error: implicit declaration of function 'flat_old_ram_flag'
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:755: error: implicit declaration of function 'flat_set_persistent'
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:757: error: implicit declaration of function 'flat_get_relocate_addr'
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:765: error: implicit declaration of function 'flat_get_addr_from_rp'
/home/bunk/linux/kernel-2.6/git/linux-2.6/fs/binfmt_flat.c:781: error: implicit declaration of function 'flat_put_addr_at_rp'Reported-by: Adrian Bunk
Signed-off-by: Adrian Bunk
Tested-by: David Howells
Acked-by: David Howells
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Make the doc consistent with current cpusets implementation.
Signed-off-by: Miao Xie
Acked-by: Paul Jackson
Cc: David Rientjes
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
If a block is computed (rather than read) then a check/repair operation
may be lead to believe that the data on disk is correct, when infact it
isn't. So only compute blocks for failed devices.This issue has been around since at least 2.6.12, but has become harder to
hit in recent kernels since most reads bypass the cache.echo repair > /sys/block/mdN/md/sync_action will set the parity blocks to the
correct state.Cc:
Signed-off-by: Dan Williams
Signed-off-by: Neil Brown
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
If an array was created with --assume-clean we will oops when trying to
set ->resync_max.Fix this by initializing ->recovery_wait in mddev_find.
Cc:
Signed-off-by: Dan Williams
Signed-off-by: Neil Brown
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
During the initial array synchronization process there is a window between
when a prexor operation is scheduled to a specific stripe and when it
completes for a sync_request to be scheduled to the same stripe. When
this happens the prexor completes and the stripe is unconditionally marked
"insync", effectively canceling the sync_request for the stripe. Prior to
2.6.23 this was not a problem because the prexor operation was done under
sh->lock. The effect in older kernels being that the prexor would still
erroneously mark the stripe "insync", but sync_request would be held off
and re-mark the stripe as "!in_sync".Change the write completion logic to not mark the stripe "in_sync" if a
prexor was performed. The effect of the change is to sometimes not set
STRIPE_INSYNC. The worst this can do is cause the resync to stall waiting
for STRIPE_INSYNC to be set. If this were happening, then STRIPE_SYNCING
would be set and handle_issuing_new_read_requests would cause all
available blocks to eventually be read, at which point prexor would never
be used on that stripe any more and STRIPE_INSYNC would eventually be set.echo repair > /sys/block/mdN/md/sync_action will correct arrays that may
have lost this race.Cc:
Signed-off-by: Dan Williams
Signed-off-by: Neil Brown
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This addresses other oopsing paths in "spidev" by changing how it manages
refcounting. It decouples the lifecycle of the per-device data from the
class device (not just the spi device):- Use class_{create,destroy} not class_{register,unregister}.
- Use device_{create,destroy} not device_{register,unregister}.
- Free the per-device data only when TWO conditions are true:
* Driver is unbound from underlying SPI device, and
* Device is no longer open (new)Also, spi_{get,set}_drvdata not dev_{get,set}_drvdata for simpler code.
Signed-off-by: David Brownell
Sebastian Siewior
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
