linux-kernel - Re: [PATCH] watchdog: pnx4008: readout watchddog state

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Thu, 11 Apr 2019 09:37:17 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Alexandre Belloni <alexandre.belloni@...tlin.com>
Cc:     Vladimir Zapolskiy <vz@...ia.com>,
        Wim Van Sebroeck <wim@...ux-watchdog.org>,
        Sylvain Lemieux <slemieux.tyco@...il.com>,
        Gregory Clement <gregory.clement@...tlin.com>,
        linux-arm-kernel@...ts.infradead.org,
        linux-watchdog@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] watchdog: pnx4008: readout watchddog state

On Thu, Apr 11, 2019 at 06:13:39PM +0200, Alexandre Belloni wrote:
> On 11/04/2019 09:03:05-0700, Guenter Roeck wrote:
> > Hi Alexandre,
> > 
> > On Thu, Apr 11, 2019 at 05:36:46PM +0200, Alexandre Belloni wrote:
> > > Readout the enabled state so it is possible to get the pre-userspace
> > > handler working. Also, avoid disabling the watchdog when nowayout is set to
> > > ensure the watchdog continues working and triggers if there is an issue
> > > later in the boot or if userspace fails to start.
> > > 
> > > Signed-off-by: Alexandre Belloni <alexandre.belloni@...tlin.com>
> > > ---
> > >  drivers/watchdog/pnx4008_wdt.c | 6 +++++-
> > >  1 file changed, 5 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/watchdog/pnx4008_wdt.c b/drivers/watchdog/pnx4008_wdt.c
> > > index 8e261799c84e..9e3714e3814c 100644
> > > --- a/drivers/watchdog/pnx4008_wdt.c
> > > +++ b/drivers/watchdog/pnx4008_wdt.c
> > > @@ -209,7 +209,11 @@ static int pnx4008_wdt_probe(struct platform_device *pdev)
> > >  	watchdog_set_nowayout(&pnx4008_wdd, nowayout);
> > >  	watchdog_set_restart_priority(&pnx4008_wdd, 128);
> > >  
> > > -	pnx4008_wdt_stop(&pnx4008_wdd);	/* disable for now */
> > > +	if (readl(WDTIM_CTRL(wdt_base)) & COUNT_ENAB)
> > > +		set_bit(WDOG_HW_RUNNING, &pnx4008_wdd.status);
> > > +
> > > +	if (!nowayout)
> > > +		pnx4008_wdt_stop(&pnx4008_wdd);	/* disable for now */
> > 
> > I don't see the point of stopping the watchdog here, even though that is the
> > old behavior. If it is possible to detect that the watchdog is running, we
> > can let the core handle pings until the watchdog device is opened. The code
> > above does that, but only if nowayout is set. That seems inconsistent and
> > unnecessary. Is there a downside to doing that unconditionally ?
> > 
> 
> I guess the downside is the change of behaviour. I don't think many will
> be affected but they will have to ensure the core pings the watchdog
> until the device is open.  WATCHDOG_HANDLE_BOOT_ENABLED is y by default
> so that should be fine for most.
> 
Your code already does change behavior if nowayout is set.

> Honestly, I would just never disable the watchdog, that is the whole
> point of this patch. If you are fine with that, I will just remove the
> line.
> 
Please.

Thanks,
Guenter

> 
> > Nitpick: s/watchddog/watchdog/ in subject line.
> > 
> 
> I'll fix it.
> 
> 
> -- 
> Alexandre Belloni, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com