summaryrefslogtreecommitdiff
path: root/Documentation/watchdog
AgeCommit message (Collapse)AuthorFilesLines
2013-07-11watchdog: delete mpcore_wdt driverViresh Kumar1-8/+0
Interrupt request doesn't use the right API: The TWD watchdog uses a per-cpu interrupt (usually interrupt #30), and the GIC configuration should flag it as such. With this setup, request_irq() should fail, and the right API is request_percpu_irq(), together with enable_percpu_irq()/disable_percpu_irq(). Nothing ensures the userspace ioctl() will end-up kicking the watchdog on the right CPU. There are no users of this driver since a long time and it makes more sense to get rid of it as nobody is looking to fix it. In case somebody wakes up after this has been removed and needs it, please revert this driver and pick these updates (These were never pushed to mainline): http://comments.gmane.org/gmane.linux.ports.arm.kernel/245998 Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Acked-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2013-03-01watchdog: core: dt: add support for the timeout-sec dt propertyFabio Porcedda1-1/+13
Add support for watchdog drivers to initialize/set the timeout field of the watchdog_device structure. The timeout field is initialised either with the module timeout parameter value (if valid) or with the timeout-sec dt property (if valid). If both are invalid the initial value is unchanged. Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com> Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-08-29watchdog: fix watchdog-test.c build warningRandy Dunlap1-1/+1
Fix compiler warning by making the function static: Documentation/watchdog/src/watchdog-test.c:34:6: warning: no previous prototype for 'term' Signed-off-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-05-30Watchdog: DA9052/53 PMIC watchdog supportAshish Jangam1-0/+5
This driver adds support for the watchdog functionality provided by the Dialog Semiconductor DA9052 PMIC chip. Tested on samsung smdkv6410 and i.mx53 QS boards. Signed-off-by: Anthony Olech <Anthony.Olech@diasemi.com> Signed-off-by: Ashish Jangam <ashish.jangam@kpitcummins.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-05-30watchdog: Add support for dynamically allocated watchdog_device structsHans de Goede1-1/+27
If a driver's watchdog_device struct is part of a dynamically allocated struct (which it often will be), merely locking the module is not enough, even with a drivers module locked, the driver can be unbound from the device, examples: 1) The root user can unbind it through sysfd 2) The i2c bus master driver being unloaded for an i2c watchdog I will gladly admit that these are corner cases, but we still need to handle them correctly. The fix for this consists of 2 parts: 1) Add ref / unref operations, so that the driver can refcount the struct holding the watchdog_device struct and delay freeing it until any open filehandles referring to it are closed 2) Most driver operations will do IO on the device and the driver should not do any IO on the device after it has been unbound. Rather then letting each driver deal with this internally, it is better to ensure at the watchdog core level that no operations (other then unref) will get called after the driver has called watchdog_unregister_device(). This actually is the bulk of this patch. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-05-30watchdog: Add Locking supportHans de Goede1-0/+2
This patch fixes some potential multithreading issues, despite only allowing one process to open the /dev/watchdog device, we can still get called multiple times at the same time, since a program could be using thread, or could share the fd after a fork. This causes 2 potential problems: 1) watchdog_start / open do an unlocked test_n_set / test_n_clear, if these 2 race, the watchdog could be stopped while the active bit indicates it is running or visa versa. 2) Most watchdog_dev drivers probably assume that only one watchdog-op will get called at a time, this is not necessary true atm. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-05-30watchdog: create all the proper device filesAlan Cox1-0/+5
Create the watchdog class and it's associated devices. Signed-off-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-05-30watchdog: Add multiple device supportAlan Cox1-1/+9
We keep the old /dev/watchdog interface file for the first watchdog via miscdev. This is basically a cut and paste of the relevant interface code from the rtc driver layer tweaked for watchdog. Revised to fix problems noted by Hans de Goede Signed-off-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-05-23Documentation/watchdog: Fix the file descriptor leak when no cmdline arg givenDevendra Naga1-0/+11
we start a infinite loop when user gives ./watchdog-test, and when user ctrl + c's the program, we just exit immeadiately with out closing the filedescriptor of the watchdog device. a signal handler is used to do the job of closing the filedescriptor and exiting the program. Signed-off-by: Devendra Naga <devendra.aaru@gmail.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-05-23Documentation/watchdog: close the fd when cmdline arg givenDevendra Naga1-3/+6
in the watchdog test code, the ioctl is performed on the watchdog device and just doing exit(0) so we leak a filedescripor. Signed-off-by: Devendra Naga <devendra.aaru@gmail.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-05-23Documentation/watchdog: Fix a small typoDevendra Naga1-1/+1
Signed-off-by: Devendra Naga <devendra.aaru@gmail.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-03-27watchdog: Add support for WDIOC_GETTIMELEFT IOCTL in watchdog coreViresh Kumar2-1/+7
This patch adds support for WDIOC_GETTIMELEFT IOCTL in watchdog core. So, there is another function pointer added to struct watchdog_ops, which can be passed by drivers to support this IOCTL. Related documentation is updated too. Signed-off-by: Viresh Kumar <viresh.kumar@st.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-03-27watchdog: watchdog_dev: Let the driver update the timeout field on ↵Hans de Goede1-3/+4
set_timeout success When a set_timeout operation succeeds this does not necessarily mean that the exact timeout requested has been achieved, because the watchdog does not necessarily have a 1 second resolution. So rather then have the core set the timeout member of the watchdog_device struct to the exact requested value, instead the driver should set it to the actually achieved timeout value. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-03-27watchdog: documentation: remove index-fileWolfram Sang1-19/+0
The 00-index file in the watchdog directory is, like many others, outdated (conversion-howto is missing) and doesn't contain worthwhile additional information. As it seems to be a maintenance burden without much gain, simply remove it. Signed-off-by: Wolfram Sang <w.sang@pengutronix.de> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-02-21Documentation: Fix typo in watchdog-kernel-api.txtMasanari Iida1-1/+1
Correct spelling "retruns" to "returns" in Documentations/watchdog/watchdog-kernel-api.txt Signed-off-by: Masanari Iida <standby24x7@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2012-01-06watchdog: documentation: describe nowayout in coversion-guideWolfram Sang1-0/+19
nowayout is also handled by the watchdog core. Describe how this needs to be addressed in the conversion guide. Signed-off-by: Wolfram Sang <w.sang@pengutronix.de> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-01-06watchdog: documentation: update index fileWim Van Sebroeck1-0/+2
Add the convert_drivers_to_kernel_api.txt to the 00-INDEX file. Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2012-01-06watchdog: add nowayout helpers to Watchdog Timer Driver Kernel APIWim Van Sebroeck1-1/+9
Add two nowayout helpers for the Watchdog Timer Driver Kernel API. And apply this to the already converted drivers. Note: s3c2410_wdt lost the nowayout feature during the conversion. Reviewed-by: Wolfram Sang <w.sang@pengutronix.de> Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2011-11-06Documentation: watchdog: add guide how to convert drivers to new frameworkWolfram Sang1-0/+195
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de> Cc: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2011-07-28watchdog: WatchDog Timer Driver Core - Add minimum and max timeoutWim Van Sebroeck1-0/+4
Add min_timeout (minimum timeout) and max_timeout values so that the framework can check if the new timeout value is between the minimum and maximum timeout values. If both values are 0, then the framework will leave the check for the watchdog device driver itself. Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-07-28watchdog: WatchDog Timer Driver Core - Add ioctl callWim Van Sebroeck1-0/+5
Add support for extra ioctl calls by adding a ioctl watchdog operation. This operation will be called before we do our own handling of ioctl commands. This way we can override the internal ioctl command handling and we can also add extra ioctl commands. The ioctl watchdog operation should return the appropriate error codes or -ENOIOCTLCMD if the ioctl command should be handled through the internal ioctl handling of the framework. Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-07-28watchdog: WatchDog Timer Driver Core - Add nowayout featureWim Van Sebroeck1-5/+8
Add support for the nowayout feature to the WatchDog Timer Driver Core framework. This feature prevents the watchdog timer from being stopped. Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-07-28watchdog: WatchDog Timer Driver Core - Add Magic Close featureWim Van Sebroeck1-0/+7
Add support for the Magic Close feature to the WatchDog Timer Driver Core framework. Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-07-28watchdog: WatchDog Timer Driver Core - Add WDIOC_SETTIMEOUT and ↵Wim Van Sebroeck1-0/+10
WDIOC_GETTIMEOUT ioctl This part add's the WDIOC_SETTIMEOUT and WDIOC_GETTIMEOUT ioctl functionality to the WatchDog Timer Driver Core framework. Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-07-28watchdog: WatchDog Timer Driver Core - Add WDIOC_SETOPTIONS ioctlWim Van Sebroeck1-2/+7
This part add's the WDIOC_SETOPTIONS ioctl functionality to the WatchDog Timer Driver Core framework. Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-07-28watchdog: WatchDog Timer Driver Core - Add WDIOC_KEEPALIVE ioctlWim Van Sebroeck1-0/+3
This part add's the WDIOC_KEEPALIVE ioctl functionality to the WatchDog Timer Driver Core framework. Please note that the WDIOF_KEEPALIVEPING bit has to be set in the watchdog_info options field. Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-07-28watchdog: WatchDog Timer Driver Core - Add basic ioctl functionalityWim Van Sebroeck1-0/+6
This part add's the basic ioctl functionality to the WatchDog Timer Driver Core framework. The supported ioctl call's are: WDIOC_GETSUPPORT WDIOC_GETSTATUS WDIOC_GETBOOTSTATUS Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-07-28watchdog: WatchDog Timer Driver Core - Add basic frameworkWim Van Sebroeck2-0/+121
The WatchDog Timer Driver Core is a framework that contains the common code for all watchdog-driver's. It also introduces a watchdog device structure and the operations that go with it. This is the introduction of this framework. This part supports the minimal watchdog userspace API (or with other words: the functionality to use /dev/watchdog's open, release and write functionality as defined in the simplest watchdog API). Extra functionality will follow in the next set of patches. Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Wolfram Sang <w.sang@pengutronix.de>
2011-03-31Fix common misspellingsLucas De Marchi1-1/+1
Fixes generated by 'codespell' and manually reviewed. Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
2010-07-01watchdog: docs: add an entry for imx2_wdtFabio Estevam1-0/+5
Add an entry for imx2_wdt in watchdog-parameters.txt Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> Acked-by: Wolfram Sang <w.sang@pengutronix.de> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2010-05-25watchdog: update/improve/consolidate watchdog driverRandy Dunlap3-4/+406
Move the limited watchdog driver help from kernel-parameters.txt to Documentation/watchdog/watchdog-parameters.txt and add info to it for all watchdog drivers except the ones that have driver-specific files already. Correct minor comments and MODULE_PARM_DESC() text in 2 places. Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2010-04-06[WATCHDOG] doc: Fix use of WDIOC_SETOPTIONS ioctl.James Hogan2-5/+8
In the watchdog-test program and watchdog-api.txt, pass the values to the WDIOC_SETOPTIONS ioctl as a pointer to an integer containing the values intead of directly in the third ioctl argument. The actual watchdog drivers in drivers/watchdog don't read the options directly from the argument but use get_user and copy_from_user. Signed-off-by: James Hogan <james.hogan@imgtec.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2010-04-04[WATCHDOG] doc: watchdog simple example: don't fail on fsync()Giel van Schijndel1-3/+0
Don't terminate the watchdog daemon when fsync() fails because no watchdog driver actually implements the fsync() syscall. Signed-off-by: Giel van Schijndel <me@mortis.eu> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2009-09-23Documentation/: fix warnings from -Wmissing-prototypes in HOSTCFLAGSLadinu Chandrasinghe1-1/+1
Fix up -Wmissing-prototypes in compileable userspace code, mainly under Documentation/. Signed-off-by: Ladinu Chandrasinghe <ladinu.pub@gmail.com> Signed-off-by: Trevor Keith <tsrk@tsrk.net> Cc: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-06-23[WATCHDOG] hpwdt: Add NMI priority optionTom Mingarelli1-4/+15
Add a priority option so that the user can choose if we do the NMI first or last. Signed-off-by: Thomas Mingarelli <thomas.mingarelli@hp.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2009-06-18[WATCHDOG] hpwdt: Add NMI sourcingThomas Mingarelli1-0/+84
Add NMI sourcing functionality (Can only be active if nmi_watchdog is inactive). Signed-off-by: Thomas Mingarelli <thomas.mingarelli@hp.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2008-10-30.gitignore updatesAlexey Dobriyan1-0/+2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-08-13docsrc: build Documentation/ sourcesRandy Dunlap1-0/+8
Currently source files in the Documentation/ sub-dir can easily bit-rot since they are not generally buildable, either because they are hidden in text files or because there are no Makefile rules for them. This needs to be fixed so that the source files remain usable and good examples of code instead of bad examples. Add the ability to build source files that are in the Documentation/ dir. Add to Kconfig as "BUILD_DOCSRC" config symbol. Use "CONFIG_BUILD_DOCSRC=1 make ..." to build objects from the Documentation/ sources. Or enable BUILD_DOCSRC in the *config system. However, this symbol depends on HEADERS_CHECK since the header files need to be installed (for userspace builds). Built (using cross-tools) for x86-64, i386, alpha, ia64, sparc32, sparc64, powerpc, sh, m68k, & mips. Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Reviewed-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-01-19[WATCHDOG] clarify watchdog operation in documentationAndrew Dyer1-17/+21
It was not clear what the difference is/was between the nowayout feature and the Magic Close feature. Signed-off-by: "Andrew Dyer" <amdyer@gmail.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2007-10-24[WATCHDOG] Documentation/watchdog/src/watchdog-simple.c: improve this codeWANG Cong1-4/+14
Make some improvements for Documentation/watchdog/src/watchdog-simple.c. Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2007-08-17[WATCHDOG] Add a 00-INDEX file to Documentation/watchdog/Jesper Juhl1-0/+10
Add a 00-INDEX file to Documentation/watchdog/ Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2007-05-24[WATCHDOG] clean-up watchdog documentationAlan Cox4-321/+60
Random sampling of some URLs in the Documentation tree to see how many were stale found that one watchdog driver was now a porn site. In fact if the watchdogs document directory was any older it would be written in latin Clean it up somewhat and add Last reviewed headers, something all the Documentation could do with IMHO. Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2006-11-30Fix typos in /Documentation : 'T''Matt LaPlante1-1/+1
This patch fixes typos in various Documentation txts. The patch addresses some +words starting with the letter 'T'. Signed-off-by: Matt LaPlante <kernel1@cyberdogtech.com> Acked-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Adrian Bunk <bunk@stusta.de>
2006-10-12[WATCHDOG] includes for sample watchdog program.Jean Delvare1-0/+2
Add missing includes to sample watchdog program. Signed-off-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
2006-10-04Fix "can not" in Documentation and KconfigMatt LaPlante1-2/+2
Randy brought it to my attention that in proper english "can not" should always be written "cannot". I donot see any reason to argue, even if I mightnot understand why this rule exists. This patch fixes "can not" in several Documentation files as well as three Kconfigs. Signed-off-by: Matt LaPlante <kernel1@cyberdogtech.com> Acked-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Adrian Bunk <bunk@stusta.de>
2006-10-04Fix typos in Documentation/: 'H'-'M'Matt LaPlante1-1/+1
This patch fixes typos in various Documentation txts. The patch addresses some words starting with the letters 'H'-'M'. Signed-off-by: Matt LaPlante <kernel1@cyberdogtech.com> Acked-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Adrian Bunk <bunk@stusta.de>
2006-10-04Fix typos in Documentation/: 'F'-'G'Matt LaPlante1-1/+1
This patch fixes typos in various Documentation txts. The patch addresses some words starting with the letters 'F'-'G'. Signed-off-by: Matt LaPlante <kernel1@cyberdogtech.com> Signed-off-by: Adrian Bunk <bunk@stusta.de>
2006-06-28[WATCHDOG] Documentation/watchdog updateRandy Dunlap5-110/+88
Documentation/watchdog/: Expose example and tool source files in the Documentation/ directory in their own files instead of being buried (almost hidden) in readme/txt files. This will make them more visible/usable to users who may need to use them, to developers who may need to test with them, and to janitors who would update them if they were more visible. Also, if any of these possibly should not be in the kernel tree at all, it will be clearer that they are here and we can discuss if they should be removed. Signed-off-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Signed-off-by: Andrew Morton <akpm@osdl.org>
2006-06-20[WATCHDOG] add WDIOC_GETTIMELEFT ioctlWim Van Sebroeck1-0/+9
Some watchdog drivers have the ability to report the remaining time before the system will reboot. With the WDIOC_GETTIMELEFT ioctl you can now read the time left before the watchdog would reboot your system. The following drivers support this new IOCTL: i8xx_tco.c, pcwd_pci.c and pcwd_usb.c . Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Signed-off-by: Andrew Morton <akpm@osdl.org>
2006-06-20[WATCHDOG] Pre-Timeout flagsCorey Minyard1-1/+29
Some watchdog timers support the concept of a "pretimeout" which occurs some time before the real timeout. The pretimeout can be delivered via an interrupt or NMI and can be used to panic the system when it occurs (so you get useful information instead of a blind reboot). Signed-off-by: Corey Minyard <minyard@acm.org> Signed-off-by: Wim Van Sebroeck <wim@iguana.be> Signed-off-by: Andrew Morton <akpm@osdl.org>