going

I have a question. If it works well, why switch to another interface (Ethernet)? This device does not have a battery-powered clock and it requires the Internet to get an accurate time stamp from the time server and set the SOC's internal clock in real time. You have disabled WI-FI and connected Ethernet. Do the settings in the connection management system look correct now? Is the network cable working? Does the banana receive the DHCP settings from the router? Did the router have Internet access at the time of banana connection? You didn't write anything about it. Just for reference. There has been a strong magnetic storm on our planet for the last two days. My Internet went down several times for one hour or more.

Boot the system in the case when it is stable, i.e. with wi-fi. Look at the system logs and post the one when the Ethernet connection failed.

Maybe yes, maybe not if there are smart devices on both ends. I have all the cables full. The device itself decides how many lines to use. dwmac-sun8i 1c30000.ethernet end0: Link is Up - 100Mbps/Full - flow control rx/tx There is an inscription on the switch: 10/100 Fast Ethernet switch

All my boards are connected by cable to an Ethernet switch. And it is connected to the router accordingly The switch is old and it is 10mbit slow. I did not perform a health check at a speed of 1 Gbit.

Almost all of these little rafts won't work without the internet. The reason is the lack of a watch with a battery. The OS finds itself in the past and thinks that the file system is corrupted because the timestamps of the files are from the future.

set command: > ls /usr/sbin | grep fsck > sudo fsck --help > man fsck dmesg | tail ..... [29283.479644] sdf: sdf1 [29283.482102] sd 4:0:0:0: [sdf] Attached SCSI removable disk [29283.882104] EXT4-fs (sdf1): mounted filesystem with writeback data mode. Opts: errors=remount-ro. Quota mode: none. fs is mounted!! > df -h .... /dev/sdf1 7,2G 1004M 6,1G 14% /run/media/leo/armbi_root > sudo umount /run/media/leo/armbi_root > sudo fsck.ext4 /dev/sdf1 e2fsck 1.46.4 (18-Aug-2021) armbi_root: clean, 28006/470112 files, 295400/1907712 blocks it is good The first three lines are for reference dmesg | tail - Get information where the SD card was mounted The disk with the SD card file system must be unmounted. If the fsck team finds bad blocks, it will ask questions. They must be answered in the affirmative. Sometimes it takes a long time. > sudo fsck.ext4 -fcy /dev/sdf1 e2fsck 1.46.4 (18-Aug-2021) Checking for bad blocks (read-only test): done armbi_root: Updating bad block inode. Pass 1: Checking inodes, blocks, and sizes Pass 2: Checking directory structure Pass 3: Checking directory connectivity Pass 4: Checking reference counts Pass 5: Checking group summary information armbi_root: ***** FILE SYSTEM WAS MODIFIED ***** armbi_root: 28006/470112 files (0.1% non-contiguous), 295400/1907712 blocks fsck.ext4 -fcy - it is forced to check and fix (It took about 7 minutes).

Remove the SD card and use the SD-USB adapter. Use another computer with a Linux OS. Check the file system on the sections of the SD card. This is a common Linux issue. There are a lot of instructions on the Internet.

People read the documentation. When they don't find something in it or.... they look at the code. You and I are among the latter. Unfortunately, it's hard for me to read the source code today. I became inattentive due to the deterioration of my eyesight. If you want to know my opinion, we can communicate in a personal conversation on the site.

@Tu Hu @AaronNGray I tested this image on my board: Armbian_23.10_Bananapim3_bookworm_edge_6.4.16_minimal.img Please confirm that the OS is loaded and working properly. If you have problems downloading, please the output in the UART console and post it here.

@Gunjan Gupta I apologize. The machine translator wrote changing the meaning. It should be tested by someone who has these boards. First of all, it's me. Yes. I'm not suggesting anything. I'm just thinking. I will not make the decisions to do this or that. My thoughts on the code: function post_config_uboot_target__extra_configs_for_bananapipro() { display_alert "$BOARD" "set dram clock" "info" run_host_command_logged scripts/config --set-val CONFIG_DRAM_CLK "384" } I didn't see the difference in the names bananapipro and bananapim3. I'm sorry, the glasses are on my nose. But that doesn't change the point. The user will only see the message: bananapipro set dram clock What if he added his own custom patch? Does the build system allow the user to add patches today? Maybe give the user the opportunity to set the value of this variable? For example: UBOOT_CONFIG_DRAM_CLK="${UBOOT_CONFIG_DRAM_CLK:-384}" function post_config_uboot_target__extra_configs_for_bananapipro() { display_alert "$BOARD" "set dram clock to $UBOOT_CONFIG_DRAM_CLK" "info" run_host_command_logged scripts/config --set-val CONFIG_DRAM_CLK "$UBOOT_CONFIG_DRAM_CLK" } Please don't listen to my grumbling, but do as you see fit.

____ ____ _ __ __ _____ | __ )| _ \(_) | \/ |___ / | _ \| |_) | | | |\/| | |_ \ | |_) | __/| | | | | |___) | |____/|_| |_| |_| |_|____/ Welcome to Armbian 23.10 Bookworm with Linux 6.4.16 No end-user support: built from trunk & unsupported (bookworm) userspace! System load: 11% Up time: 0 min Memory usage: 5% of 1.96G IP: 192.168.100.101 CPU temp: 45�°C Usage of /: 14% of 7.2G The system has become more responsive. For my board, changes are good.

This is a very big problem of the build system. Manufacturers solder different chips in different batches, but the name of the board remains the same. The most active members of the community receive the boards first and the probability that they have these devices from the same manufacturer's batch is very high. As a result, a patch appears that suits everyone. But the next batch of the manufacturer makes us strain answering questions on the forum. To remove a patch is to deprive the support of those users for whom it is good. We should be able to change some settings on the fly without making changes to the algorithm and code of the build system itself. Check it out: armbian/build> grep -nr CONFIG_DRAM_CLK ./* First, you need to check how the changes will work for different users, and then we will think about how to make universal changes.

Armbian_23.10_Bananapim3_bookworm_edge_6.4.16_minimal.img P.S. Only for the test Changes for u-boot: CONFIG_DRAM_CLK <= 480 /* DDR3-1066G @480MHz, timings: 8-8-8-18 */ .cas = 8, .tpr0 = 0x34b28890, .tpr1 = 0xa0c0, .tpr2 = 0x23200, .emr2 = 0x8,

Unfortunately, the build system does not guarantee to reproduce the build, but this information helped me find out that the uboot RAM settings were changed after the time this image was created. @Gunjan Gupta Today, the CONFIG_DRAM_CLK parameter is set to 384 for this BPI-M3 board. This happens both in the patch for this board and encoded in the build system. But the author of the topic uses an image where this parameter in uboot is left by default and is equal to 480. I assumed that the uboot developers had the same fee. I ran the build with these changes in uboot -> CONFIG_DRAM_CLK=480.

sudo dpkg-reconfigure locales

I tried to "kill" the flash drive and the OS on it by pulling the plug out of the socket. It's on the image: Armbian_community_24.2.0-trunk.449_Bananapim3_bookworm_current_6.6.13 It didn't work out. There is a slight freeze for 1 minute, and then everything continues to work. The reboot & poweroff commands are working fine.

The repository is in a shaky state. A configuration file is required. Tomorrow I'll try to build something and publish a working configuration file. By the way, you can rebuild any system source package in this branch.

That's the problem! While I was reading and comparing logs, you seem to have found it. Can't you pull the plug out of the socket, just turn off the power? It is necessary to turn off the poweroff command!

Works

Unfortunately, this happens. In order to understand the difference, let's try to compare the output of this working image with what you provided last time. To do this, please publish the dmesg output file from this working image without timestamps: dmesg --notime >lod_5.83_notime.txt P.S. Please add the DTB file from this image

Try to set the encoding en_US.UTF-8 both in the system on the BPI-M3 and in the terminal window in which you receive the log.

I checked my image. In the minicom console (BPI-M3 -UART -> FTD232 -> USB - Host Linux), in the SSH session, and on the monitor screen connected via HDMI, I see the same greeting pictures: ____ ____ _ __ __ _____ | __ )| _ \(_) | \/ |___ / | _ \| |_) | | | |\/| | |_ \ | |_) | __/| | | | | |___) | |____/|_| |_| |_| |_|____/ Welcome to Armbian 23.10 Bookworm with Linux 6.4.10 No end-user support: built from trunk & unsupported (bookworm) userspace! System load: 1% Up time: 4 min Memory usage: 6% of 1.96G IP: 192.168.100.101 CPU temp: 46�°C Usage of /: 6% of 15G leo@bananapim3:~$ dmesg | grep -i hdmi [ 0.000000] Kernel command line: root=UUID=41970142-87ec-4b91-8231-1b55ee68d69f rootwait rootfstype=ext4 splash=verbose console=ttyS0,115200 console=1 [ 0.084376] platform 1ee0000.hdmi: Fixed dependency cycle(s) with /soc/lcd-controller@1c0d000/ports/port@1/endpoint@1 [ 0.086788] platform connector: Fixed dependency cycle(s) with /soc/hdmi@1ee0000/ports/port@1/endpoint [ 2.327031] sun8i-dw-hdmi 1ee0000.hdmi: supply hvcc not found, using dummy regulator [ 2.327713] sun8i-dw-hdmi 1ee0000.hdmi: Detected HDMI TX controller v1.32a with HDCP (sun8i_dw_hdmi_phy) [ 2.328344] sun8i-dw-hdmi 1ee0000.hdmi: EVENT=plugin [ 2.328730] sun8i-dw-hdmi 1ee0000.hdmi: registered DesignWare HDMI I2C bus driver [ 2.329389] sun4i-drm display-engine: bound 1ee0000.hdmi (ops 0xc0baecbc) [ 2.330594] sun8i-dw-hdmi 1ee0000.hdmi: read_hpd result: 1 [ 43.228319] sun8i-dw-hdmi 1ee0000.hdmi: EVENT=plugin [ 44.121951] sun8i-dw-hdmi 1ee0000.hdmi: EVENT=plugin Observation: My monitor is very smart. He thinks he has correctly identified the entrance from the three possible ones. But he was wrong. I had to press the button with my hand to switch the mode and see the command prompt for entering the password.

Is the terminal connected via HDMI?

P.S. @Tu Hu There is no watch with a battery on this device. You must have an Ethernet connection. During startup, the system will request the exact time from an external server and if it does not receive it, then it will freeze (i.e. it will wait for a response) or it will be in the past with respect to the timestamps of the file system and will not be able to work.

On my board: After the reboot: All these output texts are received in the minicom console and a physical connection to the board via a USB-UART adapter. SSH connection: > ssh leo@192.168.100.101 leo@192.168.100.101's password: ____ ____ _ __ __ _____ | __ )| _ \(_) | \/ |___ / | _ \| |_) | | | |\/| | |_ \ | |_) | __/| | | | | |___) | |____/|_| |_| |_| |_|____/ Welcome to Armbian 23.10 Bookworm with Linux 6.4.10 No end-user support: built from trunk & unsupported (bookworm) userspace! System load: 1% Up time: 8 min Memory usage: 5% of 1.96G IP: 192.168.100.101 CPU temp: 43°C Usage of /: 6% of 15G [ Menu-driven system configuration (beta): sudo apt update && sudo apt install armbian-config ] Last login: Sat Jan 27 12:07:12 2024 leo@bananapim3:~$ sudo apt update [sudo] пароль для leo: Пол:1 http://security.debian.org bookworm-security InRelease [48,0 kB] Пол:2 http://deb.debian.org/debian bookworm InRelease [151 kB] ...... I have not tested HDMI and other connected devices. I communicate with the device only via UART or SSH. You will have to install the graphical environment yourself.