User avatar
thagrol
Posts: 2941
Joined: Fri Jan 13, 2012 4:41 pm
Location: Darkest Somerset, UK
Contact: Website

Re: 3A+ usbboot/rpiboot

Sat Nov 17, 2018 5:34 pm

hippy wrote:
Sat Nov 17, 2018 1:43 pm
The A+ ( and I believe A but haven't confirmed ) will also enter device boot mode if it has no SD Card fitted so a PC or other Pi can push bootcode.bin to it.
I can confirm that both the A and A+ do this, though not usably so with the default gpio expander setup from rpiboot.

The A+ fully boots but due to the files sent is in USB host mode so can't be accessed.

The A is detected and gets the initial files sent to it but doesn't boot, though if it did, I'd expect the same USB behaviour as the A+.

Unless the data sent by rpiboot has been updated for the 2837 CPU I'd expect things to fail on the 3A+ as it'l be getting the wrong kernel.
Attempts to contact me outside of these forums will be ignored unless signed in triplicate, sent in, sent back, queried, lost, found, subjected to public enquiry, lost again, and finally buried in soft peat for three months and recycled as firelighters

ejolson
Posts: 5157
Joined: Tue Mar 18, 2014 11:47 am

Re: 3A+ usbboot/rpiboot

Sat Nov 17, 2018 7:57 pm

thagrol wrote:
Sat Nov 17, 2018 5:34 pm
Unless the data sent by rpiboot has been updated for the 2837 CPU I'd expect things to fail on the 3A+ as it'l be getting the wrong kernel.
I think people have been using rpiboot with the newer compute modules for some time, so this is not the problem.

User avatar
thagrol
Posts: 2941
Joined: Fri Jan 13, 2012 4:41 pm
Location: Darkest Somerset, UK
Contact: Website

Re: 3A+ usbboot/rpiboot

Sun Nov 18, 2018 12:25 am

ejolson wrote:
Sat Nov 17, 2018 7:57 pm
thagrol wrote:
Sat Nov 17, 2018 5:34 pm
Unless the data sent by rpiboot has been updated for the 2837 CPU I'd expect things to fail on the 3A+ as it'l be getting the wrong kernel.
I think people have been using rpiboot with the newer compute modules for some time, so this is not the problem.
True, but I suspect mostly for the mass storage device mode. RPF would, of course, have updated that or the CM3 would be unusable. Was the GPIO expander/extender part updated too?

It's a moot point anyway. As gsh stated above it's a a bootloader issues based off how the OTP bits and the OTG_ID pin are set in the factory/board design. Not something rpiboot has any control over.
Attempts to contact me outside of these forums will be ignored unless signed in triplicate, sent in, sent back, queried, lost, found, subjected to public enquiry, lost again, and finally buried in soft peat for three months and recycled as firelighters

gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1551
Joined: Sat Sep 10, 2011 11:43 am

Re: 3A+ usbboot/rpiboot

Sun Nov 18, 2018 2:31 pm

Yes, that's right...

The whole problem is as described above... RE OTP bits that have been set and why this causes a problem. It only effects booting from USB as a device (i.e. using rpiboot). Although if you want to make the 3A+ into an ethernet RNDIS device (so it appears on your computer's network automatically) you can still do this as before by using the dwc3 driver, you just need to boot from an SD card and force it into device mode:

Add this to config.txt

dtoverlay=dwc2,dr_mode=peripheral

And append 'modules-load=dwc2,g_ether' to cmdline.txt (must be all on a single line)

Or other things, Andrew Mulholland has written quite a bit about doing it here:
https://gist.github.com/gbaman/50b6cca61dd1c3f88f41
--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

Return to “General discussion”