Re: [PATCH] WATCHDOG: don't auto-grab eurotechwdt.

Previous thread: [PATCH] sky2: Fix suspend/hibernation/shutdown regression with WOL enabled by Rafael J. Wysocki on Saturday, August 9, 2008 - 2:32 pm. (6 messages)

Next thread: [PATCH] [PATCH] x86: Restore proper vector locking during cpu hotplug by Eric W. Biederman on Saturday, August 9, 2008 - 3:09 pm. (2 messages)
From: Rene Herman
Date: Saturday, August 9, 2008 - 2:42 pm

[ Andrew: not submitted ]

Hi Wim.

I'm going over a list of drivers that break the boot of randconfig 
kernels by keeping resources busy:

http://lkml.org/lkml/2008/8/1/96

and eurotechwdt and plain wdt are in that list. Below is an eurotechwdt 
that makes passing in io and irq values mandatory as to keep that from 
happening.

I saw pcwd was already isafied and followed that -- due to the comment 
in eurwdt_release() though, I'm not sure about a .shutdown() method.

Do you want this in the first place? The randconfig testing does find 
actual bugs so it is useful.

I'll do wdt and the other applicable ones in there as well if yes.

Rene.
From: Alan Cox
Date: Saturday, August 9, 2008 - 2:35 pm

On Sat, 09 Aug 2008 23:42:50 +0200

NAK this one. Same reason - if its a module the auto-grab should be
enabled. 
--

From: Rene Herman
Date: Sunday, August 10, 2008 - 2:44 pm

Wim, do you want them with the grab when modular, not grab when builtin 
distinction?

Rene.
--

From: Wim Van Sebroeck
Date: Tuesday, August 12, 2008 - 11:31 am

The Berkshire PC-Watchdog card is indeed allready isafied. This is (just
like the mixcom and the ICS WDT50x watchdog cards) an ISA card that can
be inserted into an ISA slot. (The other two will be isafied also).
The shutdown notifier does the same thing then the reboot_notifier.

Most of the other drivers (except the 2 PCI cards and 1 USB card)
are however directly attached to the motherboard of the system. These
are thus in my opinion more platform devices then isa device drivers.

The problem with a lott of these devices (however) is that they can not
be probed. You either have the hardware or you don't. Some can't even be
stopped once started, some are allready active on boot.
My feeling: If you have the hardware, you will compile and load the driver
of these "platform" watchdog devices, if not you won't.

What we do best, however, to avoid problems at boot with random kernels
looks to be another issue to me.

Kind regards,
Wim.

--

From: Adrian Bunk
Date: Tuesday, August 12, 2008 - 5:58 am

WTF?

You are expecting randconfig kernels to boot?

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

--

From: Rene Herman
Date: Tuesday, August 12, 2008 - 6:12 am

90% does it seems. Ingo has been using it as part of automated testing 
for some time now.

Rene.
--

From: Adrian Bunk
Date: Tuesday, August 12, 2008 - 6:18 am

Please define "boot".

E.g. for your harddisk driver alone the probability of not having it 
included in a randconfig kernel is definitely > 10%.

Same goes for the network driver.


cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

--

From: Rene Herman
Date: Tuesday, August 12, 2008 - 6:30 am

I can't since I'm not doing the testing, Ingo is. When you ask him, do 

I suppose he has local allrandom.config files.

Rene.
--

From: Adrian Bunk
Date: Tuesday, August 12, 2008 - 6:34 am

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

--

Previous thread: [PATCH] sky2: Fix suspend/hibernation/shutdown regression with WOL enabled by Rafael J. Wysocki on Saturday, August 9, 2008 - 2:32 pm. (6 messages)

Next thread: [PATCH] [PATCH] x86: Restore proper vector locking during cpu hotplug by Eric W. Biederman on Saturday, August 9, 2008 - 3:09 pm. (2 messages)