Just a remark, I don’t have any HDD attached to the device, only the flash drive which I use to boot the kernel. Can that be a problem?by vzhilov - Debian
I mean, since my /dev/mtdblock1 partition is of a problem I still can boot if I dont stop boot process. As I understand it boots from some other default partition: Enable HD1 Enable HD2 et: I port I Interface I PHY address 1 egiga0 I RGMII In—Band 1 egigal RGMII In—Band 1 egiga2 I SGMII 0x00 egiga0, egigal, egiga2 Hit any key to stop autoboot: 0by vzhilov - Debian
I have tried some so far, but don’t get any results. I’m thinking to try with the rescue kernel that always boots in case of failover. Which one is that? How do I get that, like below? dd if=/dev/mtd /mnt/USB/uImage ?by vzhilov - Debian
Device UUID: 34375840-037F—B8C1-2285-24A3009F87FF Service UUID: FFEO RX UUID: FFE1 TX UUID: FFE1 Connecting Successfully connected to BLE device. Enable HD2 Net: port I Interface 1 PHY address I egiga0 I RGMII 1 In—Band I egigal RGMII 1 In—Band I egiga2 I SGMII 1 0x00 egiga0, egigal, egiga2 Hit any key to stop autoboot: 0 Marvell» pby vzhilov - Debian
Connecting Successfully connected to BLE device. Enable HD2 Net: 1 port I Interface I PHY address 1 egiga0 RGMII In—Band 1 egigal RGMII In—Band 1 egiga2 I SGMII I 0x00 egiga0, egigal, egiga2 Hit any key to stop autoboot: 0 Marvell» printenv CASset=max MALLOC_len=5 MPmode=SMP autoload=no baudrate=115200 boot_order=hd_scr usb_by vzhilov - Debian
If I would get a single message for a clue I will be happy Device UUID: 34375840-037F—B8C1-2285-24A3009F87FF Service UUID: FFEO RX UUID: FFE1 TX UUID: FFE1 Connecting Successfully connected to BLE device. et: 1 port Interface 1 PHY address 1 1 1 egiga0 RGMII 1 In—Band 1 egigal RGMII 1 In—Band 1 egiga2 SGMII 1 0x00 egby vzhilov - Debian
Device UUID: 34375840-037F—B8C1-2285-24A3009F87FF Service UUID: FFEO RX UUID: FFE1 TX UUID: FFE1 Connecting Successfully connected to BLE device. Enable HD2 et: 1 port I Interface PHY address 1 1 egiga0 I RGMII In—Band 1 egigal RGMII In—Band 1 egiga2 I SGMII I 0x00 egiga0, egigal, egiga2 Hit any key to stop autoboot: 0 Marvell» uby vzhilov - Debian
After I flashed OS5, I flashed my old u-boot, my initrd and tried old or new kernel. But I can only boot from recovery partition of 2015. The custom images do not boot.by vzhilov - Debian
Quote bodhi Wrote: ------------------------------------------------------- > That's probably not enough. You need to use my > kernel config and patches to make sure that it > will work. Oh, yes, you are right. Is that something new? Quote > And also use earlyprintk to see more info during > boot. > > Current: > > bootargs=root=/dev/ram console=tby vzhilov - Debian
bodhi Wrote: ------------------------------------------------------- > @dtd, > > Here is the revised version of the DTS and its > DTB. I've changed the file name to be consistent > with what we have currently. I also change the > machine name in the DTS (do a diff with your > version to see the changes). > > @all, > > Please try this armada-385by vzhilov - Debian
Thank you very much, bodhi! Great solution. Just another question. I'm putting my Ex2 Ultra device onto work now. WIth syncing raid to second HDD and in the same time `rsync` data from old disk storage I get CPU temerature up to 70-75 celcius. I can manually set fan to 75% of speed with my `mcu_ctl fan_set_75` but that doesn't seem to help a lot, maybe it gets 1-2 degrees lower. Is tby vzhilov - Debian
Maybe someone can help me on LEDs. WIth the above settings the power led stops flashing when rootfs boots. And Frist HDD led blinking red showing disk activity. This is good enough. When I run `poweroff` command on rootfs all proccesses get stopped and I get "System halted" in the console. But LEDs does not reflect this anyhow, all of them stay blue. The only difference that red lighby vzhilov - Debian
Ok, I have handled the LEDs and fan speed with scripts I got from fox-exe. Compile mcu_ctl from the attached file: apt install build-essential cc mcu_test.c -o mcu_ctl cp mcu_ctl usr/sbin/ Then put below to your rc.local file: #!/bin/sh -e # # rc.local # # This script is executed at the end of each multiuser runlevel. # Make sure that the script will "exit 0" on succeby vzhilov - Debian
And just another question. When I turn NAS on, the fan keeps running loudly. When I boot it does not go away. Where it is controlled?by vzhilov - Debian
CyberPK, Quote1) done with the attached patch and "1 -" DTSs. Tested working; 2) done with the attached "2 -" DTSs and rc.local. Tested Working;\ is that gpio_trigger.patch? I thought it only works with DTS file #2 (the one I have). And then I see bodhi later says - use point 2, it is better. Quote2nd is a better approach than 1st . Because while 1st is easier, butby vzhilov - Debian
bodhi, QuoteIf you could, edit your post to list the u-boot envs changes (permanently or on the fly during booting). Please make it a more detailed procedure so I can add to the Wiki. yes, I will do that. CyperPK, I think I use your .DTS file #2, the one that handles blue and red led lights with rc.local. But I get: ~ # echo gpio > /sys/class/leds/wdmcex2u\:blue\:hdd1by vzhilov - Debian
bodhi, I love your rootfs. Thank you for the great work! It also works perfectly with my setup! I guess I have completed my work and can write a repot. I use bodhi kernel. It is in /dev/mtdblock1. I changed it slightly so it fits 5Mb lilmit. I took some media, TV, radio modules out. This is my kernel config and my final uImage based on bodhi kernel. Then my Initrd is Busybox + updatedby vzhilov - Debian
dtd, I compiled the kernel without marvell_cesa module (option M) just to compare software encrytption vs builtin harware encryption. First I run benchmark without this module, then I add this and run benchmark again. The difference is actually 100%. Here are the results: / # cryptsetup benchmark # Tests are approximate using memory only (no storage I0). PBKDF2—shal 215578 itby vzhilov - Debian
bodhi, Quotelinux-image-5.5.3-mvebu-tld-1_1.0_armhf.deb linux-headers-5.5.3-mvebu-tld-1_1.0_armhf.deb zImage-5.5.3-mvebu-tld-1 config-5.5.3-mvebu-tld-1 linux-dtb-5.5.3-mvebu-tld-1.tar linux-5.5.3-mvebu-tld-1.patch Which one of these is the rootfs? I checked all of them, I see config and DTS files, patch for U-Boot, Kernel, Kernel modules in libs but I can't find the actuall rootby vzhilov - Debian
QuoteCould you tell me the average Read/Write speed when LUKS is used (e.g. with scp, ftp). Before I settled on encryption algorythm I have tftp. Once I settled I was thinking to use samba to transfer all files from my old storage. For now I'm a bit stuck on making rootfs. When I do: debbootstrap --arm=armhf --include=net-tools buster rootfs chroot rootfs passwd I get &quby vzhilov - Debian
CyberPK, QuoteIn order to offload disk encryption on the CESA unit, you will need to specify to cryptsetup the following cipher: aes-cbc-essiv:sha256. Therefore the command to create your encrypted disk should looked as follow: sudo cryptsetup -v -y -c aes-cbc-essiv:sha256 luksFormat /dev/sda1 You can verify running cryptsetup benchmark I had to re-build the kernel as it was missingby vzhilov - Debian
QuoteCant you share your initrd? Hello! Yes, sure. Couldn't upload 5Mb, but here is the link This Initrd is a kind of a layer that add cryptsetup before the main rootfs will be loaded. It is still on LVM though. It takes partition from /dev/md0, unencrypts its and boots from there. Can be ajusted for unencrypted partitions too. For example we could put bodhi rootfs on HDD and boot thby vzhilov - Debian
I have now made Initrd less than 5Mb to fit it to NAND. It is Busybox + Cryptsetup binaries and libraries it depends on. I'm now really thinking on advice of CyberPK to get rid of LVM and Ext4 QuoteAbout using BTRFS: It should be BTRFS over LUKS directly, managing both volumes and raid instead of EXT4 over LUKS over RAID. Also take a moment to understand what is better between LUKS oby vzhilov - Debian
Ok, I have compiled cryptsetup and Glibc for it and got over 15 Mb, So I will have to put Initrd to USB and boot from USB. So I would like to ask a question. Is it possible to boot kernel from nand and Initrd from USB? At the end of Initrd I'm going to use switch_root to Debian on HDD. So after booting Debian I will be able to remove my USB drive from NAS and still have rootfs running.by vzhilov - Debian
Thank you very much, @dtd, I understood it now. I flashed the u-boot and I can now boot fine. Great. So now I have Kernel v5.5.3 in NAND and Busybox 1.31 in NAND too. WIth a simple init script lke: #!/bin/busybox ash # init-script for WD Mycloud Mirror gen2 & Ex2 ultra # Boots from partition labeled "rootfs" /bin/busybox mkdir -p /dev /mnt/root /mnt/key /mnt/config /mntby vzhilov - Debian
Before flashing the custom u-boot I would like to understand all the steps. If I do: usb start fatload usb 0:1 0xa00000 uImage fatload usb 0:1 0x2000020 uInitrd bootm 0xa00000 0x2000020 I can boot with Kernel 5.5.3 (4.9Mb) and my uInitrd. So those are the locatations which were choosed by @CyberPK and @dtd and approved by @bodhi. But if I reboot and put just: bootm 0xa0000by vzhilov - Debian
Ok, we have updated .dts file and kernel.config and with all that we can build uImage lesser then 5Mb. So it fits to NAND and can be booted from there. In this case it would be really good to have uInitrd in NAND too, avoiding booting from USB drive. Do I have to flash u-boot anyway? Just to test it, if I do: setenv bootcmd 'nand read.e 0x02000020 0x0500000 0x0500000; nand readby vzhilov - Debian
Ah, Ok. I got it. It should be in quotes. 257 default-state = "keep";by vzhilov - Debian
QuoteHave a look at the new DTS from me or from CyberPK, the DTS from Fox seems too old. I jave tried to make your new .dts file, but I get an error. As I understood here is the new .dts file: // SPDX-License-Identifier: (GPL-2.0+ OR MIT) /* * Device Tree file for Western Digital My Cloud Expert Series EX2 Ultra * (BVBZ/Ranger Peak) * * Copyright (C) 2020 * * Based on theby vzhilov - Debian