[SOLVED] OP boot problem

[CVL]

Hello

thanks for releasing OP 0.10 which works nice on simple RPi 2. But as soon as I put on the UPS Pico Hat it won't boot anymore. After the colourful first boot screen, it tells

sh: can't  access tty; job control turned off

and drops out to an emergency prompt: /#

dmesg shows nothing suspicious, it stops after initializing USB.

Did you change anything with regard to the "normal" raspian boot sequence ?

The Pico needs to reserve GPIO 22 et 27 - are they used by any other process on OP 0.10 ?

OP 0.8 worked nicely with the Pico but : http://forum.openmarine.net/showthread.php?tid=384

Thanks in advance
Christian

[Sailoog]

Nothing special on normal raspbian boot.
GPIO 22 and 27 are free of use.

Maybe you need enable the serial interface? is disabled by default on OP 0.10.0. Go to menu - Raspberry Pi configuration - interfaces
have you installed the software your hat needs?

[CVL]

Nothing special on normal raspbian boot.
GPIO 22 and 27 are  free of use.

Maybe you need enable the serial interface? is disabled by default on OP 0.10.0. Go to menu - Raspberry Pi configuration - interfaces
have you installed the software your hat needs?

thanks for the hint on the serial interface - I had already tried that - but the real culprit was the serial console in /boot/cmdline.txt
After removing it, OP starts normally.

Cheers
Christian

[CVL]

Nothing special on normal raspbian boot.
GPIO 22 and 27 are  free of use.

Maybe you need enable the serial interface? is disabled by default on OP 0.10.0. Go to menu - Raspberry Pi configuration - interfaces
have you installed the software your hat needs?

thanks for the hint on the serial interface - I had already tried that - but the real culprit was the serial console in /boot/cmdline.txt
After removing it, OP starts normally.

Cheers
Christian

Sadly, I am coming back on this issue, as I don't get OP to start anymore with the Pico hat since I updated to 0.11. It always boots into safe mode. I tried all combinations on the serial line to no avail.

To be sure, I installed a clean Rasbian lite on another SD-Card, and there it is working all right. So there must be some difference between clean Rasbian and OP. Any ideas to where to look?
Modules are the same, except bluetooth modules loaded by OP, could it be useful to blacklist them?

I am stuck, as there are no entrys in the log-files, and dmesg on the emergency prompt just shows the messages of the safe mode.

 Going back to the original 0.10 image won't give better results, no clue why.

Any help is much appreciated
Christian

[e-sailing]

Hello

thanks for releasing OP 0.10 which works nice on simple RPi 2. But as soon as I put on the UPS Pico Hat it won't boot anymore. After the colourful first boot screen, it tells

sh: can't  access tty; job control turned off

and drops out to an emergency prompt: /#

dmesg shows nothing suspicious, it stops after initializing USB.

Did you change anything with regard to the "normal" raspian boot sequence ?

The Pico needs to reserve GPIO 22 et 27 - are they used by any other process on OP 0.10 ?

OP 0.8 worked nicely with the Pico but : http://forum.openmarine.net/showthread.php?tid=384

Thanks in advance
Christian

The difference between 0.8 and 0.10 on gpio is the use of spi bus for mcp3008

mcp3008  Pin  
Din            19
Dout         21
CL            23
CE0          24

You can try if disable SPI helps. Menu->Preferences->Raspberry Pi Configuration->Interfaces

[CVL]

The difference between 0.8 and 0.10 on gpio is the use of spi bus for mcp3008

mcp3008  Pin  
Din            19
Dout         21
CL            23
CE0          24

You can try if disable SPI helps. Menu->Preferences->Raspberry Pi Configuration->Interfaces

I tried to disable all interfaces except I2C and serial in order to eliminate this possibility of interference. Badly, it doesn't have any impact.

I wonder whether it is possible to have output of this early stage of booting? On linux bootloaders (grub) one can load line by line the kernel, initial ramdisk etc. et see where there might be a bug. Is there any such possibility on RPi ?

[CVL]

additional question : I saw that the size of kernel images of stock Rasbian and OP are not the same : OP uses a custom kernel?
Which changes where added with regard to a stock kernel ?

[CVL]

Trying to get more logs, I stumbled upon information about NOOBS partitioning problems. With NOOBS, the boot partition is #6 and partition #1 a recovery partition. I reckoned that the "safe mode" was actually the recovery partition booting. I tried to put the normal boot files (kernel.img, kernel7.img, cmdline.txt, config.txt) into the first partition and it booted all-right. So my problem has nothing to do with the GPIO nor custom kernel, but with the partition layout of NOOBS. This explains why OP 0.8 was booting fine with it's simple partition layout.

So the Pi with the Pico attached boots always the first partition, no matter what cmdline.txt says, or the partition.json in the SETTINGS partition. I asked on the Pîco forum if this is expected behaviour, I will report back.

[e-sailing]

Trying to get more logs, I stumbled upon information about NOOBS partitioning problems. With NOOBS, the boot partition is #6 and partition #1 a recovery partition. I reckoned that the "safe mode" was actually the recovery partition booting. I tried to put the normal boot files (kernel.img, kernel7.img, cmdline.txt, config.txt) into the first partition and it booted all-right. So my problem has nothing to do with the GPIO nor custom kernel, but with the partition layout of NOOBS. This explains why OP 0.8 was booting fine with it's simple partition layout.

So the Pi with the Pico attached boots always the first partition, no matter what cmdline.txt says, or the partition.json in the SETTINGS partition. I asked on the Pîco forum if this is expected behaviour, I will report back.

Ok. So the problem is that you changed the cmdline.txt to:

Code:

dwc_otg.lpm_enable=0 console=tty1 root=/dev/mmcblk0p2 rootfstype=ext4 elevator=deadline rootwait

As the manual shows in green letters. (noobs doesn't have a mmcblk0p2 partition so it can't start)

It says you only should delete the red part. So it should look like this:

Code:

dwc_otg.lpm_enable=0 console=tty1 root=/dev/mmcblk0p7 rootfstype=ext4 elevator=deadline fsck.repair=yes rootwait

[CVL]

So the Pi with the Pico attached boots always the first partition, no matter what cmdline.txt says, or the partition.json in the SETTINGS partition. I asked on the Pîco forum if this is expected behaviour, I will report back.

Ok. So the problem is that you changed the cmdline.txt to:

Code:

dwc_otg.lpm_enable=0 console=tty1 root=/dev/mmcblk0p2 rootfstype=ext4 elevator=deadline rootwait

As the manual shows in green letters. (noobs doesn't have a mmcblk0p2 partition so it can't start)

It says you only should delete the red part. So it should look like this:

Code:

dwc_otg.lpm_enable=0 console=tty1 root=/dev/mmcblk0p7 rootfstype=ext4 elevator=deadline fsck.repair=yes rootwait

Well, I didn't change the partition number in cmdline.txt   It was and is /dev/mmcblk0p7 which is the root partition of OP. The problem IMHO opinion is rather the boot partition /dev/mmcblk0p6 which is not addressed ed by the bootloader.

Could you please give me the link to the documentation, I couldn't find a hint to the partition numbers of NOOBS at the online manuat at gitbook.


Cheers