Hi,
I'm trying to get the reset functionality work on a Zynqmp Ultra96+ rev2 board.
The goal is to have (part of) the board reset on a watchdog timeout.
I'm able to trigger a watchdog reset on both APU and RPU, but the board inevitably ends up in a perpetual reset state:
xsdb% targets
1 PS TAP
2 PMU
3 PL
5* PSU
6 RPU (Reset)
7 Cortex-R5 #0 (RPU Reset)
8 Cortex-R5 #1 (RPU Reset)
9 APU (L2 Cache Reset)
10 Cortex-A53 #0 (APU Reset)
11 Cortex-A53 #1 (APU Reset)
12 Cortex-A53 #2 (APU Reset)
13 Cortex-A53 #3 (APU Reset)
The PMU did its job resetting everything, but doesn't bring the system back to life.
Some facts:
- Booting from SD card or from JTAG produces the same behavior.
- When doing a system reset, PMU goes all the way in XPfw_ResetSystem until writing in CRL_APB_RESET_CTRL, but never seems to complete this last step.
A PS-only reset seems to hang before (just after PmResetSlaveStates).
- When done before Linux runs, a 'rst -system' puts me in this state:
xsdb% targets
1 PS TAP
2 PMU
3 PL
5* PSU
6 RPU (Reset)
7 Cortex-R5 #0 (RPU Reset)
8 Cortex-R5 #1 (RPU Reset)
9 APU
10 Cortex-A53 #0 (Reset Catch, EL3(S)/A64)
11 Cortex-A53 #1 (Reset)
12 Cortex-A53 #2 (Reset)
13 Cortex-A53 #3 (Reset)
and I can get out of reset with a 'con' (FSBL will run again fine).
- If I do a subsystem reset from Linux with 'echo "PM_SYSTEM_SHUTDOWN 1 0" > /sys/kernel/debug/zynqmp-firmware/pm', then the FSBL runs again, but fails to initialize the SD card it booted from (fails in XSdPs_CardIfCond sending CMD8 and getting no response).
- PMU is compiled with EM, RECOVERY, ESCALATION, IDLE_PERIPHERALS, NODE_IDLING, and ATF is compiled with WARM_RESTART and located in OCM. FSBL runs first, from A53#0, and also in OCM.
I've defined the 4 isolated subsystems in Vivado, and I'm using the 2020.2 version.
Any help is greatly welcome, as the reset is a must-have for us.
