Hi Support Team,
We need for production test the use of kexec to avoid a reboot to load an other devicetree.
We are running an imx93 CPU on Yocto Scartgap linux 6.6.23.
We try to restart a linux kernel by:
kexec -l /mnt/Image --dtb="/mnt/msc-osm-sf-imx93-03N0A00I-module-panel-tester-dsi.dtb" --serial=ttyLP0 --append="console=ttyLP0,115200 earlycon root=/dev/mmcblk1p2 rootwait ro" -c -d
kexec -d -e --serial=ttyLP0 --command-line="console=ttyLP0,115200 earlycon root=/dev/mmcblk1p2 rootwait ro dtb-module=msc-osm-sf-imx93-03N0A00I-module-panel-tester-dsi.dtb dtb-overlay=undefined loglevel=7" -c
After the reboot trigger (kexec -e), the process goes to
[ 64.659344] ci_hdrc ci_hdrc.1: remove, state 4
[ 64.663906] usb usb1: USB disconnect, device number 1
[ 64.672819] ci_hdrc ci_hdrc.1: USB bus 1 deregistered
[ 64.710580] kvm: exiting hardware virtualization
[ 64.715254] kexec_core: Starting new kernel
[ 64.765413] psci: CPU1 killed (polled 0 ms)
[ 64.770711] Bye!
And stucked.
I could reproduce this behaviour on the EVK.
Do someone have any Idea?
