pironman 5 case fans always on in the last butterfly + wrong filesystem reported by the case display

[ZeroNoRyouki]

Batocera build version

41 2024/12/17 02:17 (butterfly)

Your architecture

Raspberry Pi 5 B 1.0

Your Graphic Processor Unit(s) (GPU)

integrated VideoCore VII

Issue description

in V40, the pironman 5 case RGB fans turned on only if the Pi temperature required it, after installing V41 they are always active from the get-go

The "free drive space" reported by the case LCD display is the one of, I think, the root filesystem (about 256 MB free) which is not that usefull: it should report the free space of /userdata (I was told by the pironman support guys to report this to you guys)

Detailed reproduction steps

n/a

Details of any attempts to fix this yourself

I've reinstalled the pironman 5 services in V41 (see below)

Details of any modifications you have made to Batocera.

I've installed (first in V40 and then again in V41 butterfly) the pironman 5 services as detailed here: https://docs.sunfounder.com/projects/pironman5/en/latest/set_up/set_up_batocera.html

Logs and data

No response

[ZeroNoRyouki]

I found a bit of time today to debug the fan problem, starting from this error from the pironman5 service:

24/12/30 13:40:57.991 [DEBUG] FanControl init
24/12/30 13:40:57.992 [DEBUG] Init GPIO Fan with pin: 6
/usr/lib/python3.11/site-packages/gpiozero/devices.py:295: PinFactoryFallback: Falling back from lgpio: 'can not open gpi
warnings.warn(
GPIO Fan init error: Cannot determine SOC peripheral base address
24/12/30 13:40:58.035 [WARNING] GPIO Fan init failed, disable gpio_fan control

I've traced it back to a problem in gpiozero : gpiozero/gpiozero#1166

The proposed workaround of symlinking /dev/gpiochip4 to /dev/gpiochip0 seems to do its job: the RGB fans are now off. The PWM fan is now recognised by the kernel correctly too (something that was not happening since I've installed V41 but I didn't notice before):

cat /sys/class/thermal/cooling_device0/cur_state
1

cat /sys/devices/platform/cooling_fan/hwmon/*/fan1_input
768

Both were 0 before creating the symlink. I'll keep an eye the fans and how they respond to changes in the temperature of the Pi

[ZeroNoRyouki]

The fans works fine, they respond correctly to temperature changes.

Fixed the oled problem too:

[root@BATOCERA /usr/lib/python3.11/site-packages/pm_auto]# diff pm_auto.py.ORG pm_auto.py
174c174
<         disk_info = get_disk_info()
---
>         disk_info = get_disk_info('/userdata')

[jlm70]

Hi @ZeroNoRyouki!
Thanks for your suggestions: for v41 I've the very same problems.
I'm not a super expert of linux, but I tried following your suggestions.

In the first case I did modify the pm_auto.py file... inserting '/userdata' at line 174.
But, after a reboot, my NVME 1TB SSD still reads as a 256MB one.
I looked back at the pm_auto.py file and... is back as the original one (while I did save my changes, closed vi editor, loaded it again and the new string was there... so... is the RPi restoring the original file at start? I entered as root so...)
Is the system in a sort of read only mode?

For the second hint, instead, I tried to issue the command:
ln -s /dev/gpiochip0 /dev/gpiochip4

But my two "cats" still reply with 0. And the fans are always on.

Any hints? Super thanks

[ZeroNoRyouki]

Hi there.
Batocera is restoring the file on boot. What I did, as a quick and dirty fix, was storing a copy of the modified pm_auto.py file in a directory under /userdata and running a couple of scripts at startup.

cat /boot/postshare.sh <-- this is executed after Batocera has mounted /userdata, it just execute the second script

/userdata/system/fixes/pironman5/pironman5_fix.sh

cat /userdata/system/fixes/pironman5/pironman5_fix.sh <-- this one creates the symlink and copy over the modified .py file

#!/bin/bash
# pironman5 temp fix

ln -s /dev/gpiochip0 /dev/gpiochip4

cp --force /userdata/system/fixes/pironman5/pm_auto.py /usr/lib/python3.11/site-packages/pm_auto/pm_auto.py

ciao
Z