Commit 4d389dcea8543161c3bb1e26f1c2ffb7e0822ff1

Authored by Alan Cox
Committed by Wim Van Sebroeck
1 parent ccb8f430ac

[WATCHDOG] clean-up watchdog documentation

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>

Showing 4 changed files with 60 additions and 321 deletions Side-by-side Diff

Documentation/watchdog/pcwd-watchdog.txt
  1 +Last reviewed: 10/05/2007
  2 +
1 3 Berkshire Products PC Watchdog Card
2 4 Support for ISA Cards Revision A and C
3 5 Documentation and Driver by Ken Hollis <kenji@bitgate.com>
... ... @@ -14,8 +16,8 @@
14 16  
15 17 The Watchdog Driver will automatically find your watchdog card, and will
16 18 attach a running driver for use with that card. After the watchdog
17   - drivers have initialized, you can then talk to the card using the PC
18   - Watchdog program, available from http://ftp.bitgate.com/pcwd/.
  19 + drivers have initialized, you can then talk to the card using a PC
  20 + Watchdog program.
19 21  
20 22 I suggest putting a "watchdog -d" before the beginning of an fsck, and
21 23 a "watchdog -e -t 1" immediately after the end of an fsck. (Remember
... ... @@ -61,7 +63,4 @@
61 63  
62 64 -- Ken Hollis
63 65 (kenji@bitgate.com)
64   -
65   -(This documentation may be out of date. Check
66   - http://ftp.bitgate.com/pcwd/ for the absolute latest additions.)
Documentation/watchdog/watchdog-api.txt
  1 +Last reviewed: 10/05/2007
  2 +
  3 +
1 4 The Linux Watchdog driver API.
2 5  
3 6 Copyright 2002 Christer Weingel <wingel@nano-system.com>
... ... @@ -22,7 +25,7 @@
22 25 notifications cease to occur, and the hardware watchdog will reset the
23 26 system (causing a reboot) after the timeout occurs.
24 27  
25   -The Linux watchdog API is a rather AD hoc construction and different
  28 +The Linux watchdog API is a rather ad-hoc construction and different
26 29 drivers implement different, and sometimes incompatible, parts of it.
27 30 This file is an attempt to document the existing usage and allow
28 31 future driver writers to use it as a reference.
29 32  
... ... @@ -46,14 +49,16 @@
46 49 shutdown on close", CONFIG_WATCHDOG_NOWAYOUT. If it is set to Y when
47 50 compiling the kernel, there is no way of disabling the watchdog once
48 51 it has been started. So, if the watchdog daemon crashes, the system
49   -will reboot after the timeout has passed.
  52 +will reboot after the timeout has passed. Watchdog devices also usually
  53 +support the nowayout module parameter so that this option can be controlled
  54 +at runtime.
50 55  
51   -Some other drivers will not disable the watchdog, unless a specific
52   -magic character 'V' has been sent /dev/watchdog just before closing
53   -the file. If the userspace daemon closes the file without sending
54   -this special character, the driver will assume that the daemon (and
55   -userspace in general) died, and will stop pinging the watchdog without
56   -disabling it first. This will then cause a reboot.
  56 +Drivers will not disable the watchdog, unless a specific magic character 'V'
  57 +has been sent /dev/watchdog just before closing the file. If the userspace
  58 +daemon closes the file without sending this special character, the driver
  59 +will assume that the daemon (and userspace in general) died, and will stop
  60 +pinging the watchdog without disabling it first. This will then cause a
  61 +reboot if the watchdog is not re-opened in sufficient time.
57 62  
58 63 The ioctl API:
59 64  
... ... @@ -226,219 +231,4 @@
226 231 WDIOS_TEMPPANIC Kernel panic on temperature trip
227 232  
228 233 [FIXME -- better explanations]
229   -
230   -Implementations in the current drivers in the kernel tree:
231   -
232   -Here I have tried to summarize what the different drivers support and
233   -where they do strange things compared to the other drivers.
234   -
235   -acquirewdt.c -- Acquire Single Board Computer
236   -
237   - This driver has a hardcoded timeout of 1 minute
238   -
239   - Supports CONFIG_WATCHDOG_NOWAYOUT
240   -
241   - GETSUPPORT returns KEEPALIVEPING. GETSTATUS will return 1 if
242   - the device is open, 0 if not. [FIXME -- isn't this rather
243   - silly? To be able to use the ioctl, the device must be open
244   - and so GETSTATUS will always return 1].
245   -
246   -advantechwdt.c -- Advantech Single Board Computer
247   -
248   - Timeout that defaults to 60 seconds, supports SETTIMEOUT.
249   -
250   - Supports CONFIG_WATCHDOG_NOWAYOUT
251   -
252   - GETSUPPORT returns WDIOF_KEEPALIVEPING and WDIOF_SETTIMEOUT.
253   - The GETSTATUS call returns if the device is open or not.
254   - [FIXME -- silliness again?]
255   -
256   -booke_wdt.c -- PowerPC BookE Watchdog Timer
257   -
258   - Timeout default varies according to frequency, supports
259   - SETTIMEOUT
260   -
261   - Watchdog cannot be turned off, CONFIG_WATCHDOG_NOWAYOUT
262   - does not make sense
263   -
264   - GETSUPPORT returns the watchdog_info struct, and
265   - GETSTATUS returns the supported options. GETBOOTSTATUS
266   - returns a 1 if the last reset was caused by the
267   - watchdog and a 0 otherwise. This watchdog cannot be
268   - disabled once it has been started. The wdt_period kernel
269   - parameter selects which bit of the time base changing
270   - from 0->1 will trigger the watchdog exception. Changing
271   - the timeout from the ioctl calls will change the
272   - wdt_period as defined above. Finally if you would like to
273   - replace the default Watchdog Handler you can implement the
274   - WatchdogHandler() function in your own code.
275   -
276   -eurotechwdt.c -- Eurotech CPU-1220/1410
277   -
278   - The timeout can be set using the SETTIMEOUT ioctl and defaults
279   - to 60 seconds.
280   -
281   - Also has a module parameter "ev", event type which controls
282   - what should happen on a timeout, the string "int" or anything
283   - else that causes a reboot. [FIXME -- better description]
284   -
285   - Supports CONFIG_WATCHDOG_NOWAYOUT
286   -
287   - GETSUPPORT returns CARDRESET and WDIOF_SETTIMEOUT but
288   - GETSTATUS is not supported and GETBOOTSTATUS just returns 0.
289   -
290   -i810-tco.c -- Intel 810 chipset
291   -
292   - Also has support for a lot of other i8x0 stuff, but the
293   - watchdog is one of the things.
294   -
295   - The timeout is set using the module parameter "i810_margin",
296   - which is in steps of 0.6 seconds where 2<i810_margin<64. The
297   - driver supports the SETTIMEOUT ioctl.
298   -
299   - Supports CONFIG_WATCHDOG_NOWAYOUT.
300   -
301   - GETSUPPORT returns WDIOF_SETTIMEOUT. The GETSTATUS call
302   - returns some kind of timer value which ist not compatible with
303   - the other drivers. GETBOOT status returns some kind of
304   - hardware specific boot status. [FIXME -- describe this]
305   -
306   -ib700wdt.c -- IB700 Single Board Computer
307   -
308   - Default timeout of 30 seconds and the timeout is settable
309   - using the SETTIMEOUT ioctl. Note that only a few timeout
310   - values are supported.
311   -
312   - Supports CONFIG_WATCHDOG_NOWAYOUT
313   -
314   - GETSUPPORT returns WDIOF_KEEPALIVEPING and WDIOF_SETTIMEOUT.
315   - The GETSTATUS call returns if the device is open or not.
316   - [FIXME -- silliness again?]
317   -
318   -machzwd.c -- MachZ ZF-Logic
319   -
320   - Hardcoded timeout of 10 seconds
321   -
322   - Has a module parameter "action" that controls what happens
323   - when the timeout runs out which can be 0 = RESET (default),
324   - 1 = SMI, 2 = NMI, 3 = SCI.
325   -
326   - Supports CONFIG_WATCHDOG_NOWAYOUT and the magic character
327   - 'V' close handling.
328   -
329   - GETSUPPORT returns WDIOF_KEEPALIVEPING, and the GETSTATUS call
330   - returns if the device is open or not. [FIXME -- silliness
331   - again?]
332   -
333   -mixcomwd.c -- MixCom Watchdog
334   -
335   - [FIXME -- I'm unable to tell what the timeout is]
336   -
337   - Supports CONFIG_WATCHDOG_NOWAYOUT
338   -
339   - GETSUPPORT returns WDIOF_KEEPALIVEPING, GETSTATUS returns if
340   - the device is opened or not [FIXME -- I'm not really sure how
341   - this works, there seems to be some magic connected to
342   - CONFIG_WATCHDOG_NOWAYOUT]
343   -
344   -pcwd.c -- Berkshire PC Watchdog
345   -
346   - Hardcoded timeout of 1.5 seconds
347   -
348   - Supports CONFIG_WATCHDOG_NOWAYOUT
349   -
350   - GETSUPPORT returns WDIOF_OVERHEAT|WDIOF_CARDRESET and both
351   - GETSTATUS and GETBOOTSTATUS return something useful.
352   -
353   - The SETOPTIONS call can be used to enable and disable the card
354   - and to ask the driver to call panic if the system overheats.
355   -
356   -sbc60xxwdt.c -- 60xx Single Board Computer
357   -
358   - Hardcoded timeout of 10 seconds
359   -
360   - Does not support CONFIG_WATCHDOG_NOWAYOUT, but has the magic
361   - character 'V' close handling.
362   -
363   - No bits set in GETSUPPORT
364   -
365   -scx200.c -- National SCx200 CPUs
366   -
367   - Not in the kernel yet.
368   -
369   - The timeout is set using a module parameter "margin" which
370   - defaults to 60 seconds. The timeout can also be set using
371   - SETTIMEOUT and read using GETTIMEOUT.
372   -
373   - Supports a module parameter "nowayout" that is initialized
374   - with the value of CONFIG_WATCHDOG_NOWAYOUT. Also supports the
375   - magic character 'V' handling.
376   -
377   -shwdt.c -- SuperH 3/4 processors
378   -
379   - [FIXME -- I'm unable to tell what the timeout is]
380   -
381   - Supports CONFIG_WATCHDOG_NOWAYOUT
382   -
383   - GETSUPPORT returns WDIOF_KEEPALIVEPING, and the GETSTATUS call
384   - returns if the device is open or not. [FIXME -- silliness
385   - again?]
386   -
387   -softdog.c -- Software watchdog
388   -
389   - The timeout is set with the module parameter "soft_margin"
390   - which defaults to 60 seconds, the timeout is also settable
391   - using the SETTIMEOUT ioctl.
392   -
393   - Supports CONFIG_WATCHDOG_NOWAYOUT
394   -
395   - WDIOF_SETTIMEOUT bit set in GETSUPPORT
396   -
397   -w83877f_wdt.c -- W83877F Computer
398   -
399   - Hardcoded timeout of 30 seconds
400   -
401   - Does not support CONFIG_WATCHDOG_NOWAYOUT, but has the magic
402   - character 'V' close handling.
403   -
404   - No bits set in GETSUPPORT
405   -
406   -w83627hf_wdt.c -- w83627hf watchdog
407   -
408   - Timeout that defaults to 60 seconds, supports SETTIMEOUT.
409   -
410   - Supports CONFIG_WATCHDOG_NOWAYOUT
411   -
412   - GETSUPPORT returns WDIOF_KEEPALIVEPING and WDIOF_SETTIMEOUT.
413   - The GETSTATUS call returns if the device is open or not.
414   -
415   -wdt.c -- ICS WDT500/501 ISA and
416   -wdt_pci.c -- ICS WDT500/501 PCI
417   -
418   - Default timeout of 60 seconds. The timeout is also settable
419   - using the SETTIMEOUT ioctl.
420   -
421   - Supports CONFIG_WATCHDOG_NOWAYOUT
422   -
423   - GETSUPPORT returns with bits set depending on the actual
424   - card. The WDT501 supports a lot of external monitoring, the
425   - WDT500 much less.
426   -
427   -wdt285.c -- Footbridge watchdog
428   -
429   - The timeout is set with the module parameter "soft_margin"
430   - which defaults to 60 seconds. The timeout is also settable
431   - using the SETTIMEOUT ioctl.
432   -
433   - Does not support CONFIG_WATCHDOG_NOWAYOUT
434   -
435   - WDIOF_SETTIMEOUT bit set in GETSUPPORT
436   -
437   -wdt977.c -- Netwinder W83977AF chip
438   -
439   - Hardcoded timeout of 3 minutes
440   -
441   - Supports CONFIG_WATCHDOG_NOWAYOUT
442   -
443   - Does not support any ioctls at all.
Documentation/watchdog/watchdog.txt
1   - Watchdog Timer Interfaces For The Linux Operating System
2   -
3   - Alan Cox <alan@lxorguk.ukuu.org.uk>
4   -
5   - Custom Linux Driver And Program Development
6   -
7   -
8   -The following watchdog drivers are currently implemented:
9   -
10   - ICS WDT501-P
11   - ICS WDT501-P (no fan tachometer)
12   - ICS WDT500-P
13   - Software Only
14   - SA1100 Internal Watchdog
15   - Berkshire Products PC Watchdog Revision A & C (by Ken Hollis)
16   -
17   -
18   -All six interfaces provide /dev/watchdog, which when open must be written
19   -to within a timeout or the machine will reboot. Each write delays the reboot
20   -time another timeout. In the case of the software watchdog the ability to
21   -reboot will depend on the state of the machines and interrupts. The hardware
22   -boards physically pull the machine down off their own onboard timers and
23   -will reboot from almost anything.
24   -
25   -A second temperature monitoring interface is available on the WDT501P cards
26   -and some Berkshire cards. This provides /dev/temperature. This is the machine
27   -internal temperature in degrees Fahrenheit. Each read returns a single byte
28   -giving the temperature.
29   -
30   -The third interface logs kernel messages on additional alert events.
31   -
32   -Both software and hardware watchdog drivers are available in the standard
33   -kernel. If you are using the software watchdog, you probably also want
34   -to use "panic=60" as a boot argument as well.
35   -
36   -The wdt card cannot be safely probed for. Instead you need to pass
37   -wdt=ioaddr,irq as a boot parameter - eg "wdt=0x240,11".
38   -
39   -The SA1100 watchdog module can be configured with the "sa1100_margin"
40   -commandline argument which specifies timeout value in seconds.
41   -
42   -The i810 TCO watchdog modules can be configured with the "i810_margin"
43   -commandline argument which specifies the counter initial value. The counter
44   -is decremented every 0.6 seconds and default to 50 (30 seconds). Values can
45   -range between 3 and 63.
46   -
47   -The i810 TCO watchdog driver also implements the WDIOC_GETSTATUS and
48   -WDIOC_GETBOOTSTATUS ioctl()s. WDIOC_GETSTATUS returns the actual counter value
49   -and WDIOC_GETBOOTSTATUS returns the value of TCO2 Status Register (see Intel's
50   -documentation for the 82801AA and 82801AB datasheet).
51   -
52   -Features
53   ---------
54   - WDT501P WDT500P Software Berkshire i810 TCO SA1100WD
55   -Reboot Timer X X X X X X
56   -External Reboot X X o o o X
57   -I/O Port Monitor o o o X o o
58   -Temperature X o o X o o
59   -Fan Speed X o o o o o
60   -Power Under X o o o o o
61   -Power Over X o o o o o
62   -Overheat X o o o o o
63   -
64   -The external event interfaces on the WDT boards are not currently supported.
65   -Minor numbers are however allocated for it.
66   -
67   -
68   -Example Watchdog Driver: see Documentation/watchdog/src/watchdog-simple.c
69   -
70   -
71   -Contact Information
72   -
73   -People keep asking about the WDT watchdog timer hardware: The phone contacts
74   -for Industrial Computer Source are:
75   -
76   -Industrial Computer Source
77   -http://www.indcompsrc.com
78   -ICS Advent, San Diego
79   -6260 Sequence Dr.
80   -San Diego, CA 92121-4371
81   -Phone (858) 677-0877
82   -FAX: (858) 677-0895
83   ->
84   -ICS Advent Europe, UK
85   -Oving Road
86   -Chichester,
87   -West Sussex,
88   -PO19 4ET, UK
89   -Phone: 00.44.1243.533900
90   -
91   -
92   -and please mention Linux when enquiring.
93   -
94   -For full information about the PCWD cards see the pcwd-watchdog.txt document.
Documentation/watchdog/wdt.txt
  1 +Last Reviewed: 10/05/2007
  2 +
  3 + WDT Watchdog Timer Interfaces For The Linux Operating System
  4 + Alan Cox <alan@lxorguk.ukuu.org.uk>
  5 +
  6 + ICS WDT501-P
  7 + ICS WDT501-P (no fan tachometer)
  8 + ICS WDT500-P
  9 +
  10 +All the interfaces provide /dev/watchdog, which when open must be written
  11 +to within a timeout or the machine will reboot. Each write delays the reboot
  12 +time another timeout. In the case of the software watchdog the ability to
  13 +reboot will depend on the state of the machines and interrupts. The hardware
  14 +boards physically pull the machine down off their own onboard timers and
  15 +will reboot from almost anything.
  16 +
  17 +A second temperature monitoring interface is available on the WDT501P cards
  18 +This provides /dev/temperature. This is the machine internal temperature in
  19 +degrees Fahrenheit. Each read returns a single byte giving the temperature.
  20 +
  21 +The third interface logs kernel messages on additional alert events.
  22 +
  23 +The wdt card cannot be safely probed for. Instead you need to pass
  24 +wdt=ioaddr,irq as a boot parameter - eg "wdt=0x240,11".
  25 +
  26 +Features
  27 +--------
  28 + WDT501P WDT500P
  29 +Reboot Timer X X
  30 +External Reboot X X
  31 +I/O Port Monitor o o
  32 +Temperature X o
  33 +Fan Speed X o
  34 +Power Under X o
  35 +Power Over X o
  36 +Overheat X o
  37 +
  38 +The external event interfaces on the WDT boards are not currently supported.
  39 +Minor numbers are however allocated for it.
  40 +
  41 +
  42 +Example Watchdog Driver: see Documentation/watchdog/src/watchdog-simple.c