| 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
| ||
|
Message-ID: <96a752d7-2818-70a5-ef69-0322f6b9bc42@roeck-us.net>
Date: Tue, 9 Mar 2021 12:49:25 -0800
From: Guenter Roeck <linux@...ck-us.net>
To: Jerry Hoemann <jerry.hoemann@....com>
Cc: Flavio Suligoi <f.suligoi@...m.it>,
"linux-watchdog@...r.kernel.org" <linux-watchdog@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Wim Van Sebroeck <wim@...ux-watchdog.org>,
Mika Westerberg <mika.westerberg@...ux.intel.com>
Subject: Re: [PATCH v1 0/2] Watchdog Core Global Parameters
On 3/9/21 10:42 AM, Jerry Hoemann wrote:
> On Tue, Mar 09, 2021 at 07:22:28AM -0800, Guenter Roeck wrote:
>> On 3/9/21 2:26 AM, Flavio Suligoi wrote:
>>> Hi Guenter,
>>>
>>>>> Instead of adding this kind of module parameters independently to each
>>>>> driver, the best solution is declaring each feature only once, in the
>>>>> watchdog core.
>>>>>
>>>>
>>>> I agree to and like the idea, but I don't see the point of letting drivers opt in
>>>> or opt out. This adds a lot of complexity for little if any gain.
>>>
>>> Do you mean that all the support for this "global parameters" should be done
>>> in the watchdog-core only, without write any code in each single
>>> "hardware" driver?
>>
>> Correct. It should not be up to the driver author to decide if they
>> want to opt out from global parameters or not. It should be up to
>> users, and users can opt out by not providing the parameters.
>
>
> Guenter,
>
> What about parameters like "pretimeout" that only some WD drivers have
> hw to support?
>
Those drivers are supposed to set WDIOF_PRETIMEOUT. If they don't, any associated
module parameter would be ignored. That is quite similar to any other non-existing
module parameter.
> Might be nice to centralize these parameters as well, but leaving it up
> to users to decide might not make sense.
>
Decide what, exactly ? Users can still provide a pretimeout module
parameter even if pretimeout is not supported for a given watchdog.
That is the case today, and it won't change.
Given that, I must admit that I don't really understand your concern.
> Or do you see the recent work to allow for a software pretimeout
> mechanism covering this?
>
That is completely orthogonal.
Guenter
> thanks
>
> Jerry
>
>>
>> Guenter
>>
>>>>
>>>> Guenter
>>>
>>> Regards,
>>>
>>> Flavio
>>>
>>>>
>>>>> Additionally, I added a implementation example of this "global"
>>>>> parameters using the module "wdat_wdt"
>>>>>
>>>>> In details:
>>>>>
>>>>> ===============================
>>>>> Watchdog Core Global Parameters
>>>>> ===============================
>>>>>
>>>>> Information for watchdog kernel modules developers.
>>>>>
>>>>> Introduction
>>>>> ============
>>>>>
>>>>> Different watchdog modules frequently require the same type of
>>>>> parameters (for example: *timeout*, *nowayout* feature,
>>>>> *start_enabled* to start the watchdog on module insertion, etc.).
>>>>> Instead of adding this kind of module parameters independently to each
>>>>> driver, the best solution is declaring each feature only once, in the
>>>>> watchdog core.
>>>>>
>>>>> In this way, each driver can read these "global" parameters and then,
>>>>> if needed, can implement them, according to the particular hw watchdog
>>>>> characteristic.
>>>>>
>>>>> Using this approach, it is possible reduce some duplicate code in the
>>>>> *new* watchdog drivers and simplify the code maintenance. Moreover,
>>>>> the code will be clearer, since the same kind of parameters are often
>>>>> called with different names (see Documentation/watchdog/watchdog-
>>>> parameters.rst).
>>>>> Obviously, for compatibility reasons, we cannot remove the already
>>>>> existing parameters from the code of the various watchdog modules, but
>>>>> we can use this "global" approach for the new watchdog drivers.
>>>>>
>>>>>
>>>>> Global parameters declaration
>>>>> ==============================
>>>>>
>>>>> The global parameters data structure is declared in
>>>>> include/linux/watchdog.h, as::
>>>>>
>>>>> struct watchdog_global_parameters_struct {
>>>>> int timeout;
>>>>> int ioport;
>>>>> int irq;
>>>>> unsigned long features;
>>>>> /* Bit numbers for features flags */
>>>>> #define WDOG_GLOBAL_PARAM_VERBOSE 0
>>>>> #define WDOG_GLOBAL_PARAM_TEST_MODE 1
>>>>> #define WDOG_GLOBAL_PARAM_START_ENABLED 2
>>>>> #define WDOG_GLOBAL_PARAM_NOWAYOUT 3
>>>>> };
>>>>>
>>>>> The variable "feature" is a bitwise flags container, to store boolean
>>>>> features, such as:
>>>>>
>>>>> * nowayout
>>>>> * start_enable
>>>>> * etc...
>>>>>
>>>>> Other variables can be added, to store some numerical values and other
>>>>> data required.
>>>>>
>>>>> The global parameters are declared (as usual for the module
>>>>> parameters) in the first part of drivers/watchdog/watchdog_core.c file.
>>>>> The above global data structure is then managed by the function *void
>>>>> global_parameters_init()*, in the same file.
>>>>>
>>>>> Global parameters use
>>>>> =====================
>>>>>
>>>>> Each watchdog driver, to check if one of the global parameters is
>>>>> enabled, can use the corresponding in-line function declared in
>>>>> include/linux/watchdog.h.
>>>>> At the moment the following functions are ready to use:
>>>>>
>>>>> * watchdog_global_param_verbose_enabled()
>>>>> * watchdog_global_param_test_mode_enabled()
>>>>> * watchdog_global_param_start_enabled()
>>>>> * watchdog_global_param_nowayout_enabled()
>>>>>
>>>>>
>>>>>
>>>>> Flavio Suligoi (2):
>>>>> watchdog: add global watchdog kernel module parameters structure
>>>>> watchdog: wdat: add start_enable global parameter
>>>>>
>>>>> Documentation/watchdog/index.rst | 1 +
>>>>> .../watchdog-core-global-parameters.rst | 74 +++++++++++++++++++
>>>>> drivers/watchdog/watchdog_core.c | 74 +++++++++++++++++++
>>>>> drivers/watchdog/wdat_wdt.c | 2 +
>>>>> include/linux/watchdog.h | 42 +++++++++++
>>>>> 5 files changed, 193 insertions(+)
>>>>> create mode 100644
>>>>> Documentation/watchdog/watchdog-core-global-parameters.rst
>>>>>
>>>
>
Powered by blists - more mailing lists