bodhi Wrote: ------------------------------------------------------- > Correct, it was wrong. It should be: > > # cat /etc/fw_env.config > > # MTD device name Device offset Env. size Flash > sector size Number of sectors > # > # Thecus N2350 > /dev/mtd2 0x00000 0x10000 0x10000 > Ok, i fixed it, and now fw_printenv shows all the envs correctly. I am sby lordzahl - uBoot
Hi, I get mjung@vega:~$ cat /proc/cmdline console=ttyS0,115200 root=LABEL=rootfs rootdelay=10 mtdparts=armada-nand:-(ubifs);spi1.0:0x00400000(uboot),0x00010000@0x00100000(uboot_env),0x00010000@0x00300000(uboot_scr) earlyprintk=serial systemd.unified_cgroup_hierarchy=0 mjung@vega:~$ cat /etc/fw_env.config # Configuration file for fw_(printenv/saveenv) utility. # Up to two entries are vby lordzahl - uBoot
Hi bodhi, Thanks for you advice. Today i connected by serial and indeed my envs were not set correctly. Actually my log from /proc/cmdline above includes "mtdparts=pxa3xx_nand-0" which i never set at all. Maybe my fw_setenv wrote to the wrong addresses? I also installed a few days back your new uboot (finally) and its envs were not saved at all. I did that now from the stock os aby lordzahl - uBoot
Hi Bodhi, Hi @All, I recently updated my Thecus N2350 to Debian Bullseye as well as the newest kernel 5.14.9. Working fine as always, thanks! I do now have some problems with Docker and cgroups v2 and wanted to change the init system (init=/bin/systemd) or even set a custom systemd config (systemd.unified_cgroup_hierarchy=0) as kernel parameters. I did set my set_bootargs parameter in ubooby lordzahl - uBoot
Great! I briefly looked over it, i hope some people will use it. I have no time to test it. Also my box is in "production" use now for quite some time. Running great!by lordzahl - Debian
bodhi Wrote: ------------------------------------------------------- > Manuel, > > dmesg | grep -i machine | cut -c40-120 > Thecus N2350 > > uname -a > Linux version 5.2.9-mvebu-tld-1 (root@tldDebianVM) > (gcc version 7.4.0 (Ubuntu/Linaro > 7.4.0-1ubuntu1~18.04.1)) #1 SMP PREEMPT Mon Aug 19 > 12:47:39 PDT 2019 > > cat /etc/debian_version > Dby lordzahl - Debian
Hi bodhi, Quotebodhi However, let's compare the system info on our boxes. How did you get that information? mjung@vega:~$ dpkg -l | grep udev ii libgudev-1.0-0:armhf 232-2 armhf GObject-based wrapper library for libudev ii libinput-bin 1.12.6-2 armhf input device managementby lordzahl - Debian
Hi bodhi, Thanks for the Update. I upgraded to your Kernel 5.2.9 and also updated the envs. I also upgraded to Debian Buster and got the MAC address problem, which has been described in another thread. I just hardcoded the address now. Otherwise everything works fine. Thanks again! Cheers, Manuelby lordzahl - Debian
Hi bodhi, Netconsole sounds great. I probably won't flash it soon, but it may come in handy. Hm, maybe some part of the memory is written, which needs to be zero when initializing the PHY? So it should be properly initialized, but it isn't. And removing power kind of initializes it? Well, but this theory does not really help with anything.. Cheers, Manuelby lordzahl - Debian
Hi, Great that you have been able to figure it out. A hardware bug would have been bad news. Cheers, Manuelby lordzahl - Debian
Hi, I need to attach to serial to see the uboot output, right? Or is it available in any log file? I can try that maybe later today. So after this you don't have ethernet? I had no problem with my network connection at all. Cheersby lordzahl - Debian
Hi bodhi! First: Thanks for the newest kernel - it is already running on my box. I only need to reboot it for kernel updates - great. > Are you booting with SATA rootfs? could you > refresh my memory if we have done this setup? I don't boot with SATA rootfs. I keep the SATA trays purely for data drives. My boot device is an old/small SSD in USB enclosure with a good chipset.by lordzahl - Debian
bodhi Wrote: ------------------------------------------------------- > Kernel linux-4.18.5-mvebu-tld-1 package has been > uploaded. Please see 1st post for download link. Thanks, bodhi! I did the update on my Thecus N2350 and it is working fine. MTDs are also ok - i used your provided command. The box has been running absolutely stable in the last months. Cheers, Manuelby lordzahl - Debian
Hi bodhi, I am not quite sure. The setup process is a little bit weird. For the initial configuration the box can not be reached through the browser at that ip. One has to use their mobile phone app or desktop (windows) program to setup the router. I think that it is not communicating at that point over tcp but maybe over udp. But i am not familiar with these basic levels of networking. I canby lordzahl - Debian
Hi, Yeah, good that you write that down here. I think i did it already like this, but i never reported it. Sometimes one misses little stuff, which might be very important to others. Cheers, Manuelby lordzahl - Debian
Hi bodhi, I removed the rescue_fw partition from my disk, so the system/debian image is the first partition on it, and it booted without problems using your envs/custom uboot. Great! :) Cheers, Manuelby lordzahl - Debian
Hi, Here we go: $ sudo nanddump --noecc --omitoob -f thecus_n2350.mtd0 /dev/mtd0 Block size 131072, page size 2048, OOB size 64 Dumping data starting at 0x00000000 and ending at 0x20000000... $ sudo dd if=/dev/mtd1 of=thecus_n2350.mtd1 bs=4096k conv=sync 1+0 records in 1+0 records out 4194304 bytes (4.2 MB, 4.0 MiB) copied, 7.02599 s, 597 kB/s $ sudo dd if=/dev/mtd2 of=thecus_nby lordzahl - Debian
Hi bodhi, Thanks for clarifying that. So i use the debian definition mtdparts=armada-nand:-(ubifs);spi1.0:0x00400000(uboot),0x00010000@0x00100000(uboot_env),0x00010000@0x00300000(uboot_scr) which leads: $ cat /proc/mtd dev: size erasesize name mtd0: 20000000 00020000 "ubifs" mtd1: 00400000 00001000 "uboot" mtd2: 00010000 00001000 "uboot_env&quby lordzahl - Debian
Hi bodhi, Thanks for your detailed description. You show the result of "cat /proc/mtd". Can you give me your line for the mtdparts definition as well? Your script only includes e.g. mtdparts=armada-nand:-(ubifs);spi1.0:0x00400000(uboot),0x00010000@0x00100000(uboot_env),0x00010000@0x00300000(uboot_scr) which leads to: $ cat /proc/mtd dev: size erasesize name mtd0: 2by lordzahl - Debian
Hi bodhi, Yes, i should be able to test this, maybe this evening or in the following days. I need to fix my mtds first. You posted a correction to their definition a couple of posts back. I will check back with you, if they are allright, before flashing, just to be sure. Cheers, Manuelby lordzahl - Debian
Hi bodhi, Thanks for the new kernel! I need to catch up to your updated mtd definitions soon.. Ok, i didn't try the fan yet. I only use 2.5" HDDs, so not much heat. Bute only full speed is not good. Sorry i don't have a lot of time at the moment. I will be very busy with work this month. Cheers, Manuelby lordzahl - Debian
Hi bodhi, Sorry, i forgot to look up my toolchain yesterday. You don't need the PATH variable, if you have your toolchain already available. The thecus guys just load their toolchain with this. The only thing you need to do, is to set CROSS_COMPILE[_BH] with an arm cross compiling toolchain. I can give detailed instructions this evening. Why did the Thecus guys use a softfp toolchaiby lordzahl - Debian
Hi bodhi, No there were no other scripts in other tar balls. There is only the "Build_Thecus.sh" script, which defines the CROSS_COMPILE(_BH) commands and then calls ./build.pl, which seems to be also a custom Thecus/Marvell build script. I have just set CROSS_COMPILE(_BH) to some appropriate cross compiler, which was installable with my package system (on ARCH linux). Should bby lordzahl - Debian
Hi, I have looked at this myself, but i think for arm boxes there is no real open source alternative to OMV. I tried OMV, but i have to say i don't really see the value. So maybe ask yourself, why do you need it and if you can not reach your goal differently? I setup my server tasks simply by hand and have a nextcloud instance running for remote file access, listing to music etc. What elsby lordzahl - Debian
Hi bodhi, Sorry for not replying for a few days. We had/still have a holiday here in germany. The hard coded bootcmd really sucks. It's a really stupid idea from the thecus guys. I will try to catch up with your posts when i am home tomorrow. I am happy to see that you had time to get your box running! Cheers, Manuelby lordzahl - Debian
Hi bodhi, Thanks for the explanations. I can also see the advantages of your more general approach - especially if one answers a lot of questions to a lot of people on a forum ;). Maybe i will switch to that method later on, but than i will simply change my ssd back, so the rootfs partition with /boot is the first partition on the drive again. Let me know, if there is anything i can do/testby lordzahl - Debian
Hi bodhi, Sure, i know my 1. configuration had to fail with your boot commands. Maybe i don't get it, but why is the partition table of my ssd needed? Or what are you planing to do? You were talking about updating u-boot. But what would be the gain? Actually i am quite happy about the boot process like it is: I don't have to change anything from the original system: I keep theby lordzahl - Debian
And now the second log, which is with my old USB stick, which only has a single partition with the rootfs: BootROM - 1.73 Booting from SPI flash General initialization - Version: 1.0.0 AVS selection from EFUSE disabled (Skip reading EFUSE values) Overriding default AVS value to: 0x23 Detected Device ID 6820 High speed PHY - Version: 2.0 Init Customer board board SerDes lanes topby lordzahl - Debian
Hi, > The correct specification for ethernet is > > > aliases { > ethernet0 = ð0; > }; > Thanks - corrected. And now the serial bootlog(s). There was a " ' " missing in a line. This is the correct one: setenv set_bootargs 'setenv bootargs "console=ttyS0,115200 root=LABEL=rootfs rootdelay=10 $mtdparts earlyprintk=serial&quby lordzahl - Debian
I have an other small but important update. My MAC address was changing with every boot. This is of course annoying for a small server box in an dhcp environment.. Some googling brought this (similar) snippet from a clearfog config up: aliases { /* So that mvebu u-boot can update the MAC addresses */ ethernet1 = ð0; }; This fixes the problem and the MAC address saved in theby lordzahl - Debian