Thank you very much for the answer.

When doing :
cat /proc/config.gz|gunzip -c|grep -i CONFIG_WATCHDOG I have:

CONFIG_WATCHDOG=y
CONFIG_WATCHDOG_CORE=y
# CONFIG_WATCHDOG_NOWAYOUT is not set
*# CONFIG_WATCHDOG_HANDLE_BOOT_ENABLED is not set*
# CONFIG_WATCHDOG_SYSFS is not set
# CONFIG_WATCHDOG_PRETIMEOUT_GOV is not set

So I suppose I have this part correct since
CONFIG_WATCHDOG_HANDLE_BOOT_ENABLED is not set.

Yet my EBG watchdog doesn't work :(
Could there be something else interfering with the boot-enabled iTCO
watchdog?
I see that Lynx Point iTCO support was added in v0.6 yet I'm at 0.7. I
suppose it *SHOULD* work.

Le ven. 29 oct. 2021 à 01:12, Jan Kiszka <[email protected]> a écrit :

> On 29.10.21 02:08, François Dussault wrote:
> > Hi,
> >
> > I wonder what could be the reason for the iTCO watchdog not firing when
> > not confirming (i.e. not calling bg_setenv -c) for the duration of 10
> > secs (value set in BGENV.DAT of both EFI partitions).
> > We're using a standard Haswell CPU from Intel.
> >
> > When booting linux I can see:
> > iTCO_wdt: Intel TCO Watchdog Timer Driver v1.11
> > iTCO_wdt: Found a Lynx Point TCO device (Version=2, TCOBASE=0x1860)
> > iTCO_wdt: initialized. heartbeat=30sec (nowayout=0)
> >
> > We ship a yocto release with efibootguard along with swupdate.
> > Nevertheless, articifically provoking a partition switch by incrementing
> > the revision and setting the watchdog timer to 10 sec and ustate on
> > "INSTALLED" and rebooting (which becomes "TESTING" after reboot), and
> > then waiting, I wait for over 5 mins and nothing reboots.
> > If I reboot manually the ustate becomes "FAILED" and revision is 0, but
> > the reboot didn't happen from the watchdog (only manually from me).
> >
> > I wonder if there could be something that pings the watchdog in
> > linux(something pinging it and thus preventing the efibg watchdog from
> > firing up).
> >
> > The EFI console says (just before booting linux):
> > Detected Intel TCO watchdog
> > Starting C:BOOT0:bzImage with watchdog set to 10 seconds
> >
> > Yet nothing really happens past those 10 secs.
> >
> > This is problematic. We have machines in the medical field that don't
> > update properly and since the watchdog never fires nobody saw it until
> > the issue was specifically reported during a demo we did!
> >
>
> If Linux boots, it may also take over the watchdog from EFI Boot Guard
> and drive it until userland picks up. Check
> CONFIG_WATCHDOG_HANDLE_BOOT_ENABLED (can also be tuned via kernel
> command line).
>
> Jan
>
> --
> Siemens AG, T RDA IOT
> Corporate Competence Center Embedded Linux
>

-- 
You received this message because you are subscribed to the Google Groups "EFI 
Boot Guard" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/efibootguard-dev/CAPWAaBvcbY3k5zWHcv_d1wsfgtF-QfdWZypHt4Le_73XJi71Bw%40mail.gmail.com.

Reply via email to