[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160111214257.GE6613@spo001.leaseweb.nl>
Date: Mon, 11 Jan 2016 22:42:57 +0100
From: Wim Van Sebroeck <wim@...ana.be>
To: Guenter Roeck <linux@...ck-us.net>
Cc: Dan Carpenter <dan.carpenter@...cle.com>,
linux-watchdog@...r.kernel.org, linux-kernel@...r.kernel.org,
kernel-janitors@...r.kernel.org
Subject: Re: [patch] watchdog: pnx833x_wdt: fix typo in MODULE_PARM_DESC
Hi All,
> On 01/08/2016 03:02 AM, Dan Carpenter wrote:
> >The module_param() is "pnx833x_wdt_timeout" and MODULE_PARM_DESC()
> >should match.
> >
> >Signed-off-by: Dan Carpenter <dan.carpenter@...cle.com>
> >
> >diff --git a/drivers/watchdog/pnx833x_wdt.c
> >b/drivers/watchdog/pnx833x_wdt.c
> >index 882fdcb..69da25a 100644
> >--- a/drivers/watchdog/pnx833x_wdt.c
> >+++ b/drivers/watchdog/pnx833x_wdt.c
> >@@ -52,8 +52,9 @@ static int pnx833x_wdt_alive;
> > /* Set default timeout in MHZ.*/
> > static int pnx833x_wdt_timeout = PNX_WATCHDOG_TIMEOUT;
> > module_param(pnx833x_wdt_timeout, int, 0);
> >-MODULE_PARM_DESC(timeout, "Watchdog timeout in Mhz. (68Mhz clock),
> >default="
> >- __MODULE_STRING(PNX_TIMEOUT_VALUE) "(30 seconds).");
> >+MODULE_PARM_DESC(pnx833x_wdt_timeout,
> >+ "Watchdog timeout in Mhz. (68Mhz clock), default="
> >+ __MODULE_STRING(PNX_TIMEOUT_VALUE) "(30 seconds).");
>
> The intent here was to have a module parameter named 'timeout'.
> Of course that was never the case. Wonder if anyone ever noticed.
> The correct fix would be to use something like
> module_param_named(timeout, pnx833x_wdt_timeout, int, 0);
> but of course that would change the ABI. On the other side,
> 'timeout' is the documented module parameter in
> Documentation/watchdog/watchdog-parameters.txt, so one could argue
> that using module_param_named() would be a bug fix and not an ABI change.
>
> Not really sure what the best approach is here. Wim, any comments ?
> Provide module_param_named() in _addition_ to the existing module_param(),
> maybe, if that is possible ?
Good question. If timeout would have been a timeout in seconds then I would
have said to go for module_param_named(). But since it is a value in Mhz, I
would rather go for Dan's patch. It will break the parameters for everyone
that has set in manually, but I personnaly think that people are using the
default value or change it via the watchdog daemon.
But if we change this, we should also make sure that
Documentation/watchdog/watchdog-parameters.txt gets corrected.
Kind regards,
Wim.
Powered by blists - more mailing lists