Thanks for reporting. And weird, since this kernel is quite old and worked well since a year or more on all RPi 1 and Zero models, including my own RPi Zero W.
Thanks for your reply @Michalng!
I tried it. With the new image my Pi does not boot at all. After about 1sec. LED starts constantly blinking (sth like 0.5s on /0.5 off). Do not see anything on UART.
Btw - I’m confused with this situation as well. This Pi worked for a year or so with very old RPi image as a bridge for Xiaomi Thermometers to SUPLA system. Then I moved to another solution and put Pi into the drawer for a long time.
After I failed with Bookworm image I also tried DietPi Bullseye with 50% success. I was able to fininsh an installation and first boot and do an initial config. But the same situation happend during dowloading packages for PiHole and Unbound (with nginx).
I tested Raspbian Lite (bookworm) also - pihole worked for a few hours (aprox. 1AM - 8AM) than the same thing happend.
I tried 3 different SD cards (Sandisk, Sandisk Ultra, Kingstone).
I used bullseye image once again. System booted up correcly. I configured basic staff like STATIC IP, regions etc.
Update with dietpi-update was succesfull.
WiFi Internet connecion is stable ~72/72 Mbps.
No messages on levels 0,1,2,3.
Just some warnings:
[ 1.686764] usb_phy_generic phy: supply vcc not found, using dummy regulator
[ 1.687028] usb_phy_generic phy: dummy supplies not allowed for exclusive requests
[ 1.711087] armv6-pmu arm-pmu: hw perfevents: no irqs for PMU, sampling events not supported
[ 2.620845] Core Release: 2.80a
[ 2.622623] Setting default values for core params
[ 2.624277] Finished setting default values for core params
[ 2.825793] Using Buffer DMA mode
[ 2.827483] Periodic Transfer Interrupt Enhancement - disabled
[ 2.829123] Multiprocessor Interrupt Enhancement - disabled
[ 2.830815] OTG VER PARAM: 0, OTG VER FLAG: 0
[ 2.832568] Dedicated Tx FIFOs mode
[ 2.834928] WARN::dwc_otg_hcd_init:1074: FIQ DMA bounce buffers: virt = dad14000 dma = 0x9ad14000 len=9024
[ 2.839886] FIQ FSM acceleration enabled for :
Non-periodic Split Transactions
Periodic Split Transactions
High-Speed Isochronous Endpoints
Interrupt/Control Split Transaction hack enabled
[ 2.847905] WARN::hcd_init_fiq:457: FIQ on core 0
[ 2.851068] WARN::hcd_init_fiq:458: FIQ ASM at c065d3d8 length 36
[ 2.854164] WARN::hcd_init_fiq:497: MPHI regs_base at df810000
[ 2.862041] Init: Port Power? op_state=1
[ 2.863527] Init: Power Port (0)
[ 3.096657] mmc0: host does not support reading read-only switch, assuming write-enable
[ 3.777639] Kernel memory protection not selected by kernel config.
System crashed again during an installation of additional packages
I can’t believe, it’s a SD Card issue, but ofc it is not a problem to buy a new one.
Is there anything else I can do or test? I do not like to give up to quick :-]
Update 12/20/2024 2:46PM
Another try, another fail, but this time dmesg shows:
Strange. Maybe I messed up something as at that time I tried a few dietpi images.
I downloaded the image you shared a few posts above once again.
Below you can check the results.
Unfortunately I do not have UART cable with me so only logs are photos I took (as I have my pi connected to the TV now).
First run (only wifi enabled and sdid and wifi password added to the config files):
After system crashed with “Oops: 17 [#1] ARM” I turned it off an on again.
Then I went through all the dietpi-config menu, changed regional settings etc.
When I tried to search for the software, the system crashed again with the same “reason”:
@Joulinar@MichaIng - thanks!
I tried the latest RPi OS Lite image. It behaves same.
I also burned extremely old, RPi OS Lite Buster image (2021-05-07-raspios-buster-armhf-lite.zip - the latest Buster I found). Just to eliminate the hardware problem. And yeah - it seems everything working fine with Buster.
Does it make sense to test more older PRi OS images to locate a version where my Pi 0 starts crashing? @MichaIng I will try with “over_voltage=2” and let you know by editing this post.
No need to try older images. Different kernel versions can imply different peak power/voltage loads, which might cause issues/instabilities like that. A typical similar issue appears often when people attach a 2.5" SSD, powered via RPi USB only. This can work for a while, but the RPi USB port is not meant to distribute the needed power reliably, and often it does not work anymore, after a certain kernel upgrade, changing whichever little power draw timing or peak.
In your case, same could be with the PSU. “USB PD” with only 2A is weird in its own . If it is made for charging in the first place, it might just not provide stable voltage too reliably, which did still work with older kernels, but not with the newer anymore. Or it is the chip itself, an outliner which requires a little higher voltage to run stable with recent kernel, in which case overvoltage=2 might do.
Okay, not exactly great to require higher voltage, but it really is a small increase that should not be recognisable by CPU temperature and does not reduce hardware life time measurably or any such. Good to keep this in mind as solution attempt for future cases, when people face individual crashes with RPis.
. If it is made for charging in the first place, it might just not provide stable voltage too reliably, which did still work with older kernels, but not with the newer anymore. Or it is the chip itself, an outliner which requires a little higher voltage to run stable with recent kernel, in which case