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  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Sat, 30 Oct 2021 19:59:42 +0300
From:   Sam Protsenko <semen.protsenko@...aro.org>
To:     Guenter Roeck <linux@...ck-us.net>
Cc:     Wim Van Sebroeck <wim@...ux-watchdog.org>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski@...onical.com>,
        linux-watchdog@...r.kernel.org,
        devicetree <devicetree@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        linux-arm Mailing List <linux-arm-kernel@...ts.infradead.org>,
        Linux Samsung SOC <linux-samsung-soc@...r.kernel.org>
Subject: Re: [PATCH 7/7] watchdog: s3c2410: Let kernel kick watchdog

On Sat, 30 Oct 2021 at 18:14, Guenter Roeck <linux@...ck-us.net> wrote:
>
> On 10/30/21 7:29 AM, Sam Protsenko wrote:
> > On Fri, 29 Oct 2021 at 03:30, Guenter Roeck <linux@...ck-us.net> wrote:
> >>
> >> On 10/28/21 11:35 AM, Sam Protsenko wrote:
> >>> When "tmr_atboot" module param is set, the watchdog is started in
> >>> driver's probe. In that case, also set WDOG_HW_RUNNING bit to let
> >>> watchdog core driver know it's running. This way wathcdog core can kick
> >>> the watchdog for us (if CONFIG_WATCHDOG_HANDLE_BOOT_ENABLED option is
> >>> enabled), until user space takes control.
> >>>
> >>> Signed-off-by: Sam Protsenko <semen.protsenko@...aro.org>
> >>> ---
> >>>    drivers/watchdog/s3c2410_wdt.c | 26 +++++++++++++++-----------
> >>>    1 file changed, 15 insertions(+), 11 deletions(-)
> >>>
> >>> diff --git a/drivers/watchdog/s3c2410_wdt.c b/drivers/watchdog/s3c2410_wdt.c
> >>> index ca082b1226e3..9af014ff1468 100644
> >>> --- a/drivers/watchdog/s3c2410_wdt.c
> >>> +++ b/drivers/watchdog/s3c2410_wdt.c
> >>> @@ -732,6 +732,21 @@ static int s3c2410wdt_probe(struct platform_device *pdev)
> >>>        wdt->wdt_device.bootstatus = s3c2410wdt_get_bootstatus(wdt);
> >>>        wdt->wdt_device.parent = dev;
> >>>
> >>> +     /*
> >>> +      * If "tmr_atboot" param is non-zero, start the watchdog right now. Also
> >>> +      * set WDOG_HW_RUNNING bit, so that watchdog core can kick the watchdog.
> >>> +      *
> >>> +      * If we're not enabling the watchdog, then ensure it is disabled if it
> >>> +      * has been left running from the bootloader or other source.
> >>> +      */
> >>> +     if (tmr_atboot && started == 0) {
> >>> +             dev_info(dev, "starting watchdog timer\n");
> >>> +             s3c2410wdt_start(&wdt->wdt_device);
> >>> +             set_bit(WDOG_HW_RUNNING, &wdt->wdt_device.status);
> >>> +     } else if (!tmr_atboot) {
> >>> +             s3c2410wdt_stop(&wdt->wdt_device);
> >>> +     }
> >>> +
> >>
> >> This doesn't cover the case where the watchdog is already enabled by the BIOS.
> >> In that case, WDOG_HW_RUNNING won't be set, and the watchdog will time out
> >> if the userspace handler is not loaded fast enough. The code should consistently
> >> set WDOG_HW_RUNNING if the watchdog is running.
> >>
> >
> > As I understand, in the case when bootloader started the watchdog, the
> > driver just stops it. You can see it in the code you replied to.
> >
> >      } else if (!tmr_atboot) {
> >              s3c2410wdt_stop(&wdt->wdt_device);
> >
> > In other words, having "tmr_atboot" module param makes it irrelevant
> > whether bootloader enabled WDT or no.
> >
>
> Sure, but I am concerned about "if (tmr_atboot && started)", which doesn't
> set WDOG_HW_RUNNING with your current code, and I was looking for something
> like
>
>         if (tmr_atboot) {
>                 if (!started) {
>                         dev_info(dev, "starting watchdog timer\n");
>                         s3c2410wdt_start(&wdt->wdt_device);
>                 }
>                 set_bit(WDOG_HW_RUNNING, &wdt->wdt_device.status);
>         } else {
>                 s3c2410wdt_stop(&wdt->wdt_device);
>         }
>

Wow, I really overlooked that case. Nice catch! Not having '} else {'
section is vicious...

Though if started != 0, it means s3c2410wdt_set_heartbeat() failed to
set wdd->timeout, and without that the watchdog core won't be able to
calculate correctly ping interval in watchdog_next_keepalive(), and
WDOG_HW_RUNNING bit won't do much good, right? So I'll probably just
call s3c2410wdt_stop() in that case, to be on the safe side.

Also this 'started' variable name is misleading, I'll convert it to
"bool timeout_ok" while at it.

> Guenter

Powered by blists - more mailing lists