The hack itself seems to be copied from Linux rti_wdt.c, but the WDT reset
principle is different in U-Boot. While Linux relies on correct frequencies
and timers and doesn't check the actual WDT counter value U-Boot driver
seems to be more robust: it does compare RTIDWDCNTR vs RTIDWDPRLD.
Now the root cause of the original motivation to manipulate the clock rate
is said to be understood and fixed in Linux commit cae58516534e
("watchdog: rti_wdt: Set min_hw_heartbeat_ms to accommodate a safety margin")
which simultaneously removed the hack itself.
While is fix part of the mentioned patch is neither applicable nor requried
for the U-Boot driver just drop the hack setting WDT clock rate to 90% of
the real rate. This has a nice effect that the WDT timeout is now as
requested and not 10% shorter.
Signed-off-by: Alexander Sverdlin <alexander.sverdlin@siemens.com>
Reviewed-by: Jan Kiszka <jan.kiszka@siemens.com>
Reviewed-by: Stefan Roese <sr@denx.de>
If the RTI watchdog has been enabled in SPL, enabling it in U-Boot proper
fails because it can only be enabled once in HW and never stopped. This
however leads to a situation that wdt_cyclic() watchdog trigger is not
being started any longer and the WDT fires at some point.
Allow for WDT re-start by not bailing out if the [previously] configured
period matches the one to be configured.
Enabling in [A53] SPL has been tested on AM62x-based HW (where [A53] SPL is
responsible for loading R5 DM firmware and not this driver).
Signed-off-by: Alexander Sverdlin <alexander.sverdlin@siemens.com>
Reviewed-by: Jan Kiszka <jan.kiszka@siemens.com>
Reviewed-by: Stefan Roese <sr@denx.de>
As part of bringing the master branch back in to next, we need to allow
for all of these changes to exist here.
Reported-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Tom Rini <trini@konsulko.com>
When bringing in the series 'arm: dts: am62-beagleplay: Fix Beagleplay
Ethernet"' I failed to notice that b4 noticed it was based on next and
so took that as the base commit and merged that part of next to master.
This reverts commit c8ffd1356d42223cbb8c86280a083cc3c93e6426, reversing
changes made to 2ee6f3a5f7550de3599faef9704e166e5dcace35.
Reported-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Tom Rini <trini@konsulko.com>
Remove <common.h> from this driver directory and when needed
add missing include files directly.
Acked-by: Angelo Dureghello <angelo@kernel-space.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
When running against RC_OSC_32k, the watchdog may suffer from running
faster than expected, expiring earlier. The Linux kernel adds a 10%
margin to the timeout calculation by slowing down the read clock rate
accordingly. Do the same here, also to have comparable preset values
for both drivers.
Along this, fix the name of the local var holding to frequency - in Hz,
not kHz.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Stefan Roese <sr@denx.de>
To avoid the need of extra boot scripting on AM65x for loading a
watchdog firmware, add the required rproc init and loading logic for the
first R5F core to the watchdog start handler. In case the R5F cluster is
in lock-step mode, also initialize the second core. The firmware itself
is embedded into U-Boot binary to ease access to it and ensure it is
properly hashed in case of secure boot.
One possible firmware source is https://github.com/siemens/k3-rti-wdt.
The board is responsible for providing the firmware as additional
loadable via the U-Boot fit image. The driver will pick up its location
from /fit-images/k3-rti-wdt-firmware then.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
This flag has the word 'REMOVE' in it which means it conflicts with
the DM_REMOVE flags. Rename it to DM_FLAG_LEAVE_PD_ON which seems to
indicate its purpose well enough.
Signed-off-by: Simon Glass <sjg@chromium.org>
This construct is quite long-winded. In earlier days it made some sense
since auto-allocation was a strange concept. But with driver model now
used pretty universally, we can shorten this to 'auto'. This reduces
verbosity and makes it easier to read.
Coincidentally it also ensures that every declaration is on one line,
thus making dtoc's job easier.
Signed-off-by: Simon Glass <sjg@chromium.org>
This is based on the Linux kernel driver for the RTI watchdog.
To actually reset the system on an AM65x, it requires firmware running
on the R5 that accepts the NMI and issues the actual system reset via
TISCI. Kind of an iTCO, except that this watchdog hardware has support
for no-way-out, and only for that.
On the J721E, reset works without extra firmware help when routing the
RTI interrupt via the ESM.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
