kernel 5.14.1
|
Quote:
What are the other differences? |
only by the fact that it has not yet got into the main kernel and some wifi drivers
|
update kernel 5.15.5
slarm64-current-aarch64-server-raspberry_pi_4-5.15.5-build-20211128.img.zst slarm64-current-aarch64-server-raspberry_pi_4-5.15.5-build-20211128.img.zst.sha256 slarm64-current-aarch64-xfce-raspberry_pi_4-5.15.5-build-20211128.img.zst slarm64-current-aarch64-xfce-raspberry_pi_4-5.15.5-build-20211128.img.zst.sha256 |
Thanks for the links. Downloading as I speak.
I was feeling disloyal running Raspberry Pi OS (=Debian Buster). I avoided the initial spade work of setting up the Slackware distro, but excused myself on the basis of being an invalid, and RazPi setup was handy, I have to admit. Now RazPi OS (formerly raspbian) have dropped Debian 10 (Buster) and dropped all updates. They have released bullseye or debian 11, and apparently there's no end of problems, see post #4 particularly. Downloading the 64bit Slackware version as I speak. The one thing I'd mention is that Alien Bob's [x86_64] chromium-ungoogled Version 94.0.4606.81 (Official Build, ungoogled-chromium) (64-bit) pukes on the link; If I do a 'file/save as,' I get the error Code:
support.9oo91e.qjz9zk is blocked |
|
Installed that RazPi 4 20211128-xfce image, and it's not booting.
Apparently it does the initial stuff fine, remounts mounting on /mnt read only, boots, moves the / partition from /mnt to /, but doesn't mount it read write, so the thing runs init, but every write to / fails. That inevitably ends sadly, a reset in this case. Then we loop again, … and again… I tried it on a real usb disk (/=sda2), and an sdcard (/=mmcblk0p2) - both the same. Raspberry Pi OS save bootup bother by not having uEnv.txt. The cmdline.txt & /etc/fstab both have "root=PARTUUID=<number>. When you install an image, it's the partuuid of the disk the image was made on, I suppose. You could save bother for end users that way. I personally don't like partuuids, but I detest farting about. One other thing. What's the root password? |
Quote:
I documented "persistent storage device naming" in a SARPi mini-project, for the benefit of those who need/want to use it: https://sarpi.penthux.net/index.php?p=pers-devnam |
Quote:
|
@Exega: Thanks for the reply. I did as you suggested, but it didn't help. I gather from the page you linked that the canonical names are /dev/sdX, etc and the uuids are links to them.
@sndwvs: Thanks for the reply. It's good to know about root's password, but I have yet to get as far as logging in. Mine is the standard Razpi 4B, w/ 4G of ram. I tried to get a handle on this problem with mobile phone photos, but one handed phone photos are full of camera shake.I'm missing the clear output from / being moved from /mnt to /. After adding some overscan, I can see that it doesn't like my filesystem. I get "could be anything" error (code page, dirty bit set, superblock or other error) and then announces that root is a read only file system. root is ext4. It runs /sbin/init, gets past updating gtk modules, to the next thing (QT?) and then switches to runlevel 6, and resets. Got any ideas? It's not a filesystem fault. Usb disk & sdcard together? I checked the usb disk anyhow, and it's fine. |
Well, the way this fell out was that after examining the init scripts, it's pretty clear that / is mounted before /sbin/init is run. As /sbin/init is not included in the initrd, the mount of the / partition must be done read-only. I get a 'could be anything' type of error message, mentioning code page, superblock,… or other error. So booting fails. The kernel may work for others. It fails here.
I'm on the bog standard Pi 4B w/4G of ram and one firmware update for usb booting. I realise I'm talking to devs who have a developed instinct for issues, experience of possible problems and energy to solve them. I have none of those, poor health and dodgy balance. So I have wiped the promising looking slarm64 and grabbed the Slackware Arm 32bit. I'm fine with 32 a bit OS because 32bit is not significantly slower in my use case. |
discussion here
|
Hi sndwvs,
I have a PoE hat on top of one of my Pi4s. After I recently upgraded my slarm64 installation, (just after the reboot) the fan of the PoE hat seems to have stopped spinning. I went from kernel 5.12.something to 5.15.5. The fan was working fine before that. I don't see any fan or pwm related messages in dmesg. Do you have any idea what I should check? |
Hi wowbaggerHU,
let's see the dmesg output, also the lsmod output. and Code:
grep poe /boot/config.txt |
Quote:
dmesg: Code:
[Fri Dec 10 22:27:46 2021] Booting Linux on physical CPU 0x0000000000 [0x410fd083] Code:
Module Size Used by Code:
root@ntp:~/d# fgrep -i poe /boot/config |
looked at the changes, there are related commits.
try a new kernel 5.15.7 |
Quote:
Is it OK that the modules package is unusually large? (~600 MB while it's 16 MB for the Pi3) |
Quote:
no i updated. |
Hi sndwvs,
Unfortunately the fan is still standing still after the upgrade. I did not see any "pwm" or "poe" strings in the output of dmesg. |
try to manually load the modules and see the output dmesg:
Code:
modprobe gpio-fan gpio-pwm pwm-bcm2835 pwm-raspberrypi-poe raspberrypi-hwmon rpi_poe_power |
Quote:
Code:
root@ntp:~# modprobe gpio-fan; modprobe gpio-pwm; modprobe pwm-bcm2835; modprobe pwm-raspberrypi-poe; modprobe raspberrypi-hwmon; modprobe rpi_poe_power |
here they propose to update the firmware and try to change the pins of the gpio.
|
try the previous kernel 5.14.21
|
Quote:
Sorry but I had little time to work on this. I tested your older kernel, but the fan still doesn't turn. As others describe it: on boot, it moves slightly, but it doesn't even complete on full turn, and stands still from then on. I did not yet have time to explore update the firmware and to change the GPIO pins. (I also have a gps module plus an RTC connected to the board.) |
Quote:
Check out the code at the bottom of the installation tutorial for Slackware AArch64. Basically it sets the fan at max speed and decreases it to a speed that's suitable for the Slackware Installer. You could try doing something similar on the RPi. Just to clarify - the problem was resolved by writing the speed multiple times to the fan driver's interface, not running it at a lower speed. I always needed to start the fan at max rpm and decrement it. It's a complete hack but is suitable for the Installer. |
Quote:
Perhaps there is not enough power. |
Quote:
No I don't think that would be the case. The system is stable, (I have high uptimes), I power it over PoE, plus it has been running just fine with the same HW setup for months now without any problems. I would believe the system would spontaneously reboot or just run in an unstable way if the power was low. |
Quote:
Hi Stuart, Thanks for your suggestion, but unfortunately on the Rpi4, with the PoE hat, I only have the following files under /sys/devices/platform/pwm-fan: Code:
root@ntp:~# find /sys/devices/platform/pwm-fan |
Hi wowbaggerHU,
it was just described that at startup it twitches and stopped working on the old kernel. |
Quote:
Then we may have misunderstood each other somewhere. Originally I was running something like kernel 5.10.something, or 5.12.something, and the fan was working. Then I first upgraded to 5.15.7 (at boot time the fan moved for a fraction of a second and then stopped), and you suggested using 5.14.21, where the same thing happened. Sorry if I wasn't clear enough. Best Regards, János |
|
Quote:
The fan seems to work OK with this "old" kernel version. I believe that I may have been originally using something like 5.12.x, and that also worked OK, so whatever broke it, it may have happened after 5.12, but before (or up to) 5.14. |
Hi wowbaggerHU,
let's analyze the dmesg of this kernel and lsmod and compare with 5.15.y. |
Quote:
Code:
[17:09] janos@janos-work{43}:[~/_dissect_table/pwm-poe]{2}$ diff -U 1 dmesg-5.10.83 dmesg-5.15.5 Code:
[17:09] janos@janos-work{43}:[~/_dissect_table/pwm-poe]{1}$ diff -U 1 lsmod-5.10.83 lsmod-5.15.5 Code:
[17:13] janos@janos-work{43}:[~/_dissect_table/pwm-poe]{0}$ diff -U 1 modules-5.10.83 modules-5.15.5 |
updated config check 5.15.9
|
Quote:
Here is dmesg: Code:
[Thu Dec 16 21:47:03 2021] Booting Linux on physical CPU 0x0000000000 [0x410fd083] |
no, now everything is there
Code:
[Thu Dec 16 21:47:05 2021] raspberrypi-firmware soc:firmware: Request 0x00038049 returned status 0x80000001 only those that are here |
Quote:
Unfortunately I get the message '"blobs/broadcom/boot" did not exist on "arm"' when visiting your images-build-kit repo. |
Hi wowbaggerHU,
then try this image slarm64-current-aarch64-server-raspberry_pi_4-5.15.9-build-20211217.img.zst |
Quote:
I extracted most of the /boot files from your image, and placed them in my /boot partition. The fan seems to work with these new files! Thanks for your help! |
All times are GMT -5. The time now is 08:04 PM. |