[Letux-kernel] Pandaboard ES no longer booting properly

[H. Nikolaus Schaller]

> Am 01.03.2020 um 12:34 schrieb H. Nikolaus Schaller <hns at goldelico.com>:
> 
> When trying to use a Bullseye rootfs to test SGX DDK1.17
> on omap4 I was not able to boot the Pandaboard. It
> repeatedly does a reboot after ca. 12 seconds.
> 
> It turned out that it happens always after the line
> 
> [   12.609222] hub 2-1:1.0: 5 ports detected
> 
> Since I am not sure if the board is broken I tried older
> kernels than letux-5.6-rc3.
> 
> It turns out that 4.19.105 and 4.19.66 fail at exactly the
> same location while 4.19.8 does boot:
> 
> * 5.6-rc3			reboots after [   15.251495] hub 1-1:1.0: 5 ports detected
> * 4.19.106			reboots after [   12.609222] hub 2-1:1.0: 5 ports detected
> * 4.19.66			reboots after [   12.498352] hub 1-1:1.0: 5 ports detected
> * 4.19.8			continues after [ 12.361938] hub 2-1:1.0: 5 ports detected
> 
> So the board is ok but Letux kernel is broken somehow.
> 
> Next steps will be to "bisect" through kernel binaries
> between 4.19.8 and 4.19.66 to better locate the issue.
> 
> Then - maybe - a bisect may help. Or diff shows something
> with relevance.

After trying to find the issue between 4.19.8 and 4.19.66
I could not reproduce it. 4.19.66 did work and then 4.19.106
and finally 5.6-rc3 as well.

Strange.

One thing that may be part of an explanation are the next log lines after the hub message:

[   12.793212] hub 2-1:1.0: 5 ports detected
[   12.806610] omapdss_dss 58000000.dss: 58000000.dss supply vdda_video not found, using dummy regulator
[   12.825592] omapdss_dss 58000000.dss: 58000000.dss supply vdda_video not found, using dummy regulator

If there was a hickup/latchup with the vdda_video regulator the PMIC may have done the reboot.

[   13.134246] usb 2-1.1: new high-speed USB device number 3 using ehci-omap

Or there was a hickup in the Ethernet controller.

[  OK  ] Listening on Load/Save RF …itch Status /dev/rfkill Watch.
[   13.204681] usb 2-1.1: New USB device found, idVendor=0424, idProduct=ec00, bcdDevice= 1.00
[   13.228820] usb 2-1.1: New USB device strings: Mfr=0, Product=0, SerialNumber=0
         Starting Load/Save RF Kill Switch Statu[   13.253448] omapdss_dss 58000000.dss: 58000000.dss supply vdda_video not found, using dummy regulator
s...
[  OK  ] Reached target Bluetooth.
[  OK  ] Started Load/Save RF Kill Switch Status.
[   13.516784] cfg80211: Loading compiled-in X.509 certificates for regulatory database
[   13.608642] cfg80211: Loaded X.509 cert 'sforshee: 00b28ddf47aef9cea7'

But since I had powered down the PandaBoard between every reboot attempt it is a little
unlikely that there was some persistently bad voltage state.

Anyways, I can now try what I had planned: check the HDMI port and try to get SGX working.

BR,
Nikolaus