linux-kernel - Re: [PATCH v3] watchdog: Add watchdog timer support for the WinSystems EBC-C384

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 25 Jan 2016 17:58:06 -0800
From:	Guenter Roeck <linux@...ck-us.net>
To:	William Breathitt Gray <vilhelm.gray@...il.com>,
	One Thousand Gnomes <gnomes@...rguk.ukuu.org.uk>
Cc:	wim@...ana.be, linux-watchdog@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3] watchdog: Add watchdog timer support for the
 WinSystems EBC-C384

On 01/25/2016 03:36 PM, William Breathitt Gray wrote:
> On 01/25/2016 03:42 PM, Guenter Roeck wrote:
>> On 01/25/2016 11:28 AM, One Thousand Gnomes wrote:
>>> If ask for 299 seconds surely I should get 300 not 240 ?
>>> (Whether to round off or round up is an interesting question for the
>>> middle range - does it go off early or late - I'd have said late but...)
>>>
>>
>> Matter of endless discussion. Some argue that the value should be rounded
>> up, some argue that it should be rounded down, some argue that it should
>> be rounded to the closest match. Each camp has its own valid arguments.
>> I usually leave it up to the driver's author to decide, with a slight
>> preference to never select a value larger than requested.
>
> I implemented it to round down simply because it was the simplest
> solution (i.e. integer truncation). Although I see merit in an
> implementation that rounds to the closest valid value, I'll keep the
> current implementation for now due to its simplicity; if enough users of
> the driver prefer a different implementation, then I'll add it in a
> later patch.
>
>>> Is there no ACPI entry for it ?
>>>
>> Same here. As long as the board is identified, I tend to leave it up
>> to the driver author to decide _how_ to identify it.
>>
>> Only question for me would be if the watchdog timer is implemented
>> in a Super-IO chip, and if so, if it would be possible to use the chip
>> identification instead of a DMI (or ACPI) entry to instantiate the driver.
>
> I do not believe there is an ACPI entry for it. Interestingly, the
> watchdog timer BIOS configuration option for this motherboard is listed
> under the Super I/O menu; perhaps this watchdog timer is implemented in
> the Super I/O chip.
>
> The manual for this motherboard does not provide much information about
> the Super I/O chip (no model number, etc.), and neither sensors-detect
> nor superiotool was able to detect it. I've sent an email to the
> motherboard company (WinSystems) requesting further information about
> the Super I/O chip and whether the watchdog timer is built-in to the
> Super I/O chip.
>

I had a close look at the board pictures in the manual. Looks like there
is a NCT7802Y hardware monitoring chip on the board. This suggests that the
Super-IO chip does probably not implement hardware monitoring. Not that
it helps much to know that, but it reduces the possibilities.

There is also a pretty large Lattice FPGA on the board, which makes it at
least somewhat likely that the functionality is implemented in that FPGA.

Guenter