uboot 2016.09-1 whats new

[summers]

Odd. I checked the pogo_e02 uboot installer pogo_e02.sh, it had the command in it:

$this->bbcode_second_pass_code('', 'ln -s fw_printenv fw_setenv')

from:

http://de3.mirror.archlinuxarm.org/os/armv5te/boot/pogo_e02/pogo_e02.sh

So I assumed it had set it up, and that on pogo e02 it was the same executable as fw_printenv, just with a different command name ...

But I note that the script downloads to /tmp - so probably removed on a reboot. Still from the script you can see how to download fw_printenv ...

Oh yes - note the $this->bbcode_second_pass_code('', 'setenv usb_pgood_delay 10000') is a uboot command - so thats the hassle, you have to execute a uboot command without access to uboot. This you can only do by rewriting the scripts in the uboot variables. So I'd scan the uboot varaibles for the $this->bbcode_second_pass_code('', 'usb start') command, then add the $this->bbcode_second_pass_code('', 'setenv') command to run just before ...

e.g. see the boot commands on my NSA325 which I unrolled here:

https://archlinuxarm.org/forum/viewtopic.php?f=58&t=9798&start=21

[hfrjkse]

I'm affraid that this is a bit dangerous for me to play with. If I'd make a mistake I may brick my pink devil Maybe moonman will issue some fix to his uBoot sometime. At least we know what is the problem.

[moonman]

There is no fix really. It is just just certain flash drives don't play nice with the uboot. I can bring it up to 2015.07 but I doubt it will fix anything. If you want a 2015.07 build let me know.

[summers]

@hfrjkse

yes i understand your concerns, changing variables in uboot is the equivalent of changing setting in a PC BIOS. moonman doing a later version of uboot, is like flashing a new bios. Hence if it goes wrong, its a hassle to recover.

In this case, if it goes wrong, its a case of cracking the pogo case, and pluging in a serial connector. Seems like most pogo have a UART connector on the board, so its just a case of getting the right lead (sounds like a CD ROM lead!), so only something like £15 for the USB to UART TTL + cable.

I can understand that you aren't comfortable.

There is another approach, suggested by moonman. As far as I can see everyone else thatr has had this problem, its only been with specific USB memory stick. e.g. some USB memory sticks run very close to the USB specs, and uboot also runs close to the specs - and so this causes some failures. It means another solution for you is to try a different USB stick - thats also stands a good chance of solving the problem.

[hfrjkse]

That's exactly what I did. I found some Lexar flashdrive in my drawer and it is working perfectly. No more booting problems, although I found that I can't spin down my 2TB Seagate using hdparm. This is strange as it used to work perfectly when it was plugged into B04 running some old Arch, 2012 I think... 1TB Toshiba does spin down though. Of course this is not related with uBoot. Eventually I used Seagate's tool to change spindown time using internal firmware, so everything is now up and running... at least I hope so Thanks for all your suggestions and help!

[nwestfal]

Would it be possible to get a default value added for ethaddr in CONFIG_EXTRA_ENV_SETTINGS in pogo_e02.h? That way those of us that have an E02 with a bad nand block at 0xc0000 and can't load ethaddr from environment would at least have a valid ethaddr and should therefore theoretically be able to use netconsole. I have a workaround for Arch, so it's not a huge deal, but would be nice to be able to use netconsole!

[Sagittarius]

I've updated remotely for a friend his goflexnet with latest uboot and kernel-dt.Everything works like a charm, thanks to the devs.

But because the right slot of the goflexnet has not been used for years (before the system was installed on the left sata slot), it's really oxydated and dmesg shows me ata errors

What is the equivalent of the old syntax we used before (sata_device=1:1) to be able to boot on the left sata slot now ?

I've managed it by setting : setenv devnum 1

[summers]

Pity in 2016.09-1 there was no interesting in updating the install to my suggestion:

https://archlinuxarm.org/forum/viewtopic.php?f=53&t=10499&p=52403#p52478

Still as I wasn't sure if it would be changed, I didn't install the new uboot automatically, but did by hand.