Update 2023-04-17: Much of the below has changed; in no particular order:
- Most of the work to rescue these P1020 boards has been through the use of boot wrappers to protect from non-functional vendor U-Boot(s).
- A u-boot port for these boards has ultimately not materialized. It was more challenging than it looked!
- The glut of P1020 boards reaching EOL has slowed down.
Hi there,
TL;DR: I'm making this post to bring attention to Freescale's P10XX series, particularly to needed cleanup, OEM EOLs and porting potential.
It's come to my attention that there are a number of mpc85xx devices out there needing some attention. I want to focus that effort and keep it organized, particularly when submitting pull requests to the OpenWrt Github or patch requests to the mailing lists, as review work is very heavy for maintainers.
The mpc85xx target presents an interesting opportunity at this point in time: there are a large number of enterprise devices based on the P1020RDB / P1020WLAN which are nearing end-of-life in the very near future (or the very recent past), which means availability will be very high.
Some have outstanding pull requests, or potentially re-submissible PRs that just need cleanup:
- EdgeCore ECW7210 (thread here)
-
Aerohive HiveAP 370
- Work started by @dmascord; I have helped them clean up the port, and it is essentially ready to merge.
Some devices will need pull requests soon:
-
DONE: Extreme Networks AP3825i
- My port is working; I still need to
- [x] finish fixing LEDs (done by blocktrron)
- [x] clean up device tree
- [x] clean up and squash commits, and add instructions
- [x] test with more devices
- @Gadorach has done plenty of this, but I'd like to make sure to test against my devices. I'm sure, as with the Enterasys AP3710i, some of the devices will have passwords, and I want to verify that factory-reset instructions to replace the password work.
- [x] specify how large the kernel must grow for instructions to stop working
- growth of the kernel has caused problems repeatedly in the past
- My port is working; I still need to
Some devices are ready to be ported:
-
- This device uses a P1010E. It appears to require a u-boot patch in order to autoboot OpenWrt.
- It's noted by @Slimey that many Adtran Bluesocket devices in this series use P1010 / P1020.
Some devices need fixing:
- [x] DONE The Aerohive HiveAP 330 can no longer decompress its own kernel; I have an idea process to patch the forced bootcmd in u-boot.
- [x] Start an issue thread for this fixup.
- Done; see my repo's issue thread here.
- [x] Implement the fix.
- Done; see this commit here.
- [x] Test the fix.
- [x] Start an issue thread for this fixup.
Finally: On any P1 / P2 QorIQ device with a JTAG interface, it is possible to use a BDI2000 to recover the device, making a U-boot port feasible to write.
Many vendors' u-boots are a pain to deal with, but can be manually circumvented in a way that can be used to overwrite the bootloader. I have purchased a BDI2000 and gotten access to the firmware to drive these boards; I plan to use it to write a U-boot replacement for the HP MSM466.
Other useful things all of these P1 / P2 devices seem to have:
- an external console port
- a watchdog timer (see page 16) to allow us to force a reset and inspect RAM, particularly to see how U-boot is manipulating the device tree
Note that this 'QorIQ' is not @stintel's ppc64 QorIQ -- those processors are a later generation, e.g. T1042.