zjd Posted February 8 Posted February 8 (edited) I updated to Armbian 26.2.1 on OPi5, and it can't boot from MTD (rootfs at NVMe), and I have to use SD card to boot to NVMe as workaround. After looking into this issue, I found rkspi_loader.img is no longer available for current (6.18.8 kernel) u-boot pkg and armbian-config flashes u-boot-rockchip-spi.bin ( or similar sata version) to /dev/mtdblock0 and I think this is wrong image for MTD bootloader. If I use older vendor image, which u-boot pkg has rkspi_loader.img, system can boot to NVMe rootfs and I notice that /dev/mtdblock0 has several gpt partitions including u-boot one, and flashing u-boot-rockchip-spi.bin erases partitions in SPI flash. I noticed that several related changes were introduced since 25.11, such as https://github.com/armbian/build/commit/e5b845f9432abb0408287599bba5889a86e6d068 , but I couldn't do quick fix to bring back working rkspi_loader.img because I am not familiar with whole armbian build. One thing to note that if board was flashed with old rkspi_loader.img and not updated with 26.2 MTD image, it still boots to NVMe rootfs, but I have other issues such as video playback not working with 6.18 kernel likely due to mismatch u-boot and kernel version leading to wrong HW initialization. BTW, I checked 26.2 u-boot pkg for Rock 5B (similar rk3588 board), and rkspi_loader.img is still included. Edited February 8 by zjd 0 Quote
zjd Posted February 9 Author Posted February 9 After digging into more details, I think Armbian has moved to use new U-Boot generated FIT image u-boot-rockchip-spi.bin for SPI boot, however, it doesn't work to boot NVMe rootfs and sometimes breaks the board and I have to erase SPI to get SD card booting. Couldn't find more information how Armbian does MTD boot now. Note that older vendor SPI bootloader can boot kernel 6.18.8 but I ran into other issue like freezing system for a couple of minutes so that I'd say to avoid use old MTD image with 26.2 6.18 kernel. 0 Quote
zjd Posted February 13 Author Posted February 13 I connected UART today and did some debugging, and it looks like that this u-boot 2025.10 for Armbian 26.02 doesn't see NVMe drive, i.e. 'nvme scan' returns nothing. Is OPi5 NVMe support included in upstream mainline u-boot? 0 Quote
Werner Posted February 13 Posted February 13 The related PR states this was tested. @Efe Çetin can you confirm this still works as expected? 0 Quote
Efe Çetin Posted February 16 Posted February 16 On 2/13/2026 at 7:43 AM, Werner said: The related PR states this was tested. @Efe Çetin can you confirm this still works as expected Will check it when i back to PC, will be away from pc for around a week 2 Quote
xsetiadi Posted February 27 Posted February 27 (edited) how about armbian 26.2.1 but still using vendor kernel? did we still have this problem? Edited February 28 by xsetiadi 0 Quote
zjd Posted March 1 Author Posted March 1 If you have flashed working rkspi_loader.img to MTD before, i.e. from previous releases, you can keep using it with vendor kernel w/o updating MTD. The 26.2.1 image with vendor kernel uses new FIT u-boot loader as well, I doubt it will work with NVMe. BTW, I need 6.18 kernel and new u-boot and noticed that 26.2.1 image default to vendor kernel instead of 6.18 kernel now. 0 Quote
xsetiadi Posted March 1 Posted March 1 ok, i guess i just dd new image then. have you read github release a few minutes ago? orangepi5: bump u-boot to v2026.04-rc3 0 Quote
Igor Posted March 1 Posted March 1 On 2/27/2026 at 7:28 PM, xsetiadi said: but still using vendor kernel? Vendor kernel will stay for awhile (this year for sure) as mainline is not on this level yet. 10 minutes ago, xsetiadi said: orangepi5: bump u-boot to v2026.04-rc3 This part might not be part of the release - you can grab it from nightly release - but most likely it won't have any affect to this topic. 1 Quote
mircsicz Posted March 10 Posted March 10 (edited) Oh damn, wish I would have read this before I tried to install (26.2 with vendor kernel build on the 9th) from SD to Nvme... How can I reflash MTD if I only have a MacBook available? Note to self and the next one hitting that wall: Going to maskrom mode kinda fixed it, after that I started again from a mSD and from that running ArmBian I flashed the same img to the Nvme on the board. Another reboot and it was started from the SSD 😉 Edited March 10 by mircsicz Fixed it 0 Quote
xsetiadi Posted March 12 Posted March 12 so today im finally reinstall my opi5, and when choose mtd to flash i got two option, i guess the sata version is if im using sata ssd instead of nvme ssd? 0 Quote
Werner Posted March 13 Posted March 13 8 hours ago, xsetiadi said: i guess the sata version is if im using sata ssd instead of nvme ssd? correct 0 Quote
xsetiadi Posted March 13 Posted March 13 i find it is very convenient now, not only able to choose boot from nvme or sata ssd. but can also install the image from armbian-config. last time im installing armbian i need to dd the image manualy into nvme. very nice work for armbian team 0 Quote
Werner Posted March 13 Posted March 13 Well the only thing armbian-config does, is calling "armbian-install" which has been there for ages 1 Quote
edgar-vincent Posted April 16 Posted April 16 Hi everyone, I just got an Orange Pi 5 and am experiencing the same issue as the OP. I've since then installed the bootloader on the SD card and the rootfs on the NVME drive. I don't really understand what came out of the discussion here: is it possible to boot from the MTD? Thanks a lot! EV 0 Quote
Werner Posted April 16 Posted April 16 Booting from mtd/nvme should just work. Most common problems: - dirty mtd/emmc. vendor/unmatching boot loader can cause issues. - mixup between m.2 sata and actual nvme. Correct overlay/dt must be used - wrong rootfs uuid in armbianEnv.txt after running armbian-install (after done, do not reboot but mount freshly written nvme to a random folder and check if the uuid is correct) 0 Quote
edgar-vincent Posted April 17 Posted April 17 Thank you very much for your reply, Werner - and indeed for all the tremendous work you do! The drive is listed as an /dev/nvme0n1p1, so I think I did choose the right overlay. Rootfs UUID is correct. I tested SPI/MTD boot with Armbian 26.2.1, rootfs on NVMe. I erased and reflashed the SPI flash manually: ``` shell sudo flash_eraseall /dev/mtd0 sudo dd if=/usr/lib/linux-u-boot-vendor-orangepi5/u-boot-rockchip-spi.bin \ of=/dev/mtdblock0 bs=1M conv=fsync status=progress sync ``` I then verified that the running system was still using /boot from the SD card: ``` shell findmnt /boot ``` which returned: ``` shell /boot /dev/mmcblk1p1[/boot] ``` I mounted the NVMe rootfs and compared /boot on SD vs NVMe: ``` shell sudo mkdir -p /mnt/nvme sudo mount /dev/nvme0n1p1 /mnt/nvme ls -la /boot ls -la /mnt/nvme/boot ``` The NVMe /boot contents were older/stale compared with the SD /boot, so I synchronized them: ``` shell sudo rsync -aHAX --delete /boot/ /mnt/nvme/boot/ sync ``` I then powered the board off completely, removed the SD card, and tested booting from SPI + NVMe again but it still won't boot without the SD card. I watched your video on UART debugging, but unfortunately, my serial cable is too slow, so I'll get one of those you recommend and try again. Thanks again! 0 Quote
zjd Posted 1 hour ago Author Posted 1 hour ago I tried u-boot 2026.04 from rolling release, still has same problem: DDR 9fa84341ce typ 24/09/06-09:51:11,fwver: v1.18 ch0 ttot10 ch1 ttot10 ch2 ttot10 ch3 ttot10 ch0 ttot16 LPDDR4X, 2112MHz channel[0] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB ch1 ttot16 channel[1] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB ch2 ttot16 channel[2] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB ch3 ttot16 channel[3] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB Manufacturer ID:0x1 DQS rds:h6,h1 CH0 RX Vref:28.9%, TX Vref:20.8%,22.8% DQ rds:h5 h1 l0 l0 l1 h1 h1 h2, l0 l0 h1 h2 h1 h2 h1 h4 DQS rds:l0,h1 CH1 RX Vref:29.7%, TX Vref:21.8%,20.8% DQ rds:h4 l0 h3 h5 l1 h6 h1 h2, h1 h6 h3 h7 h1 h2 h2 l1 DQS rds:h1,l0 CH2 RX Vref:32.2%, TX Vref:20.8%,20.8% DQ rds:h3 l0 h5 h2 h1 h6 l0 h1, h1 h4 h3 h1 h1 h3 h5 h4 DQS rds:h1,h1 CH3 RX Vref:29.7%, TX Vref:23.8%,22.8% DQ rds:h1 h1 h1 h4 h4 l0 h1 h1, l0 h3 h7 l1 l0 h3 h1 h3 stride=0x2, ddr_config=0x4 hash ch_mask0-1 0x20 0x40, bank_mask0-3 0xa00 0x1400 0x2800 0x0, rank_mask0 0x401000 change to F1: 528MHz ch0 ttot10 ch1 ttot10 ch2 ttot10 ch3 ttot10 change to F2: 1068MHz ch0 ttot12 ch1 ttot12 ch2 ttot12 ch3 ttot12 change to F3: 1560MHz ch0 ttot14 ch1 ttot14 ch2 ttot14 ch3 ttot14 change to F0: 2112MHz ch0 ttot16 ch1 ttot16 ch2 ttot16 ch3 ttot16 out U-Boot SPL 2026.04_armbian-2026.04-S88dc-P9ef2-H95ba-Va0a8-B138f-R448a (Apr 25 2026 - 16:56:23 +0000) Trying to boot from SPI ## Checking hash(es) for config config-1 ... OK ## Checking hash(es) for Image atf-1 ... sha256+ OK ## Checking hash(es) for Image u-boot ... sha256+ OK ## Checking hash(es) for Image fdt-1 ... sha256+ OK ## Checking hash(es) for Image atf-2 ... sha256+ OK ## Checking hash(es) for Image atf-3 ... sha256+ OK INFO: Preloader serial: 2 NOTICE: BL31: v2.3():v2.3-868-g040d2de11:derrick.huang, fwver: v1.48 NOTICE: BL31: Built : 15:02:44, Dec 19 2024 INFO: spec: 0x13 INFO: code: 0x88 INFO: ext 32k is not valid INFO: ddr: stride-en 4CH INFO: GICv3 without legacy support detected. INFO: ARM GICv3 driver initialized in EL3 INFO: valid_cpu_msk=0xff bcore0_rst = 0x0, bcore1_rst = 0x0 INFO: l3 cache partition cfg-0 INFO: system boots from cpu-hwid-0 INFO: disable memory repair INFO: idle_st=0x21fff, pd_st=0x11fff9, repair_st=0xfff70001 INFO: dfs DDR fsp_params[0].freq_mhz= 2112MHz INFO: dfs DDR fsp_params[1].freq_mhz= 528MHz INFO: dfs DDR fsp_params[2].freq_mhz= 1068MHz INFO: dfs DDR fsp_params[3].freq_mhz= 1560MHz INFO: BL31: Initialising Exception Handling Framework INFO: BL31: Initializing runtime services WARNING: No OPTEE provided by BL2 boot loader, Booting device without OPTEE initialization. SMC`s destined for OPTEE K ERROR: Error initializing runtime service opteed_fast INFO: BL31: Preparing for EL3 exit to normal world INFO: Entry point address = 0x800000 INFO: SPSR = 0x3c9 U-Boot 2026.04_armbian-2026.04-S88dc-P9ef2-H95ba-Va0a8-B138f-R448a (Apr 25 2026 - 16:56:23 +0000) Model: Xunlong Orange Pi 5 SoC: RK3588S DRAM: 8 GiB Core: 726 devices, 33 uclasses, devicetree: separate MMC: mmc@fe2c0000: 0 Loading Environment from nowhere... OK In: serial@feb50000 Out: serial@feb50000 Err: serial@feb50000 Model: Xunlong Orange Pi 5 SoC: RK3588S Net: eth0: ethernet@fe1c0000 Hit any key to stop autoboot: 0 <comment: stopped for mannual command > => nvme scan <comment: No NVMe drive found> => boot Scanning for bootflows in all bootdevs Seq Method State Uclass Part Name Filename --- ----------- ------ -------- ---- ------------------------ ---------------- Scanning global bootmeth 'efi_mgr': Card did not respond to voltage select! : -110 Cannot persist EFI variables without system partition Scanning bootdev 'mmc@fe2c0000.bootdev': Card did not respond to voltage select! : -110 Card did not respond to voltage select! : -110 Card did not respond to voltage select! : -110 Card did not respond to voltage select! : -110 scanning bus for devices... USB EHCI 1.00 USB OHCI 1.0 USB EHCI 1.00 USB OHCI 1.0 USB XHCI 1.10 Bus usb@fc800000: 1 USB Device(s) found Bus usb@fc840000: 1 USB Device(s) found Bus usb@fc880000: 1 USB Device(s) found Bus usb@fc8c0000: 1 USB Device(s) found Bus usb@fcd00000: 1 USB Device(s) found Scanning bootdev 'ethernet@fe1c0000.bootdev': ethernet@fe1c0000 Waiting for PHY auto negotiation to complete....... done BOOTP broadcast 1 BOOTP broadcast 2 DHCP client bound to address 192.168.0.50 (257 ms) missing environment variable: pxeuuid Retrieving file: pxelinux.cfg/01-02-2a-1c-db-3c-a0 Using ethernet@fe1c0000 device TFTP from server 192.168.0.1; our IP address is 192.168.0.50 Filename 'pxelinux.cfg/01-02-2a-1c-db-3c-a0'. Load address: 0xe00000 Loading: T T <comment: went through predefined boot devices and stuck at waiting for TFTP server, which doesn't exist> When SD card plugged in, u-boot went through boot devices and found that it is bootable and load OS: DDR 9fa84341ce typ 24/09/06-09:51:11,fwver: v1.18 ch0 ttot10 ch1 ttot10 ch2 ttot10 ch3 ttot10 ch0 ttot16 LPDDR4X, 2112MHz channel[0] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB ch1 ttot16 channel[1] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB ch2 ttot16 channel[2] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB ch3 ttot16 channel[3] BW=16 Col=10 Bk=8 CS0 Row=16 CS1 Row=16 CS=2 Die BW=16 Size=2048MB Manufacturer ID:0x1 DQS rds:h4,h1 CH0 RX Vref:29.3%, TX Vref:20.8%,21.8% DQ rds:h7 h1 l0 l0 l0 h1 h2 h2, l0 l0 h1 h2 h1 h3 h1 h4 DQS rds:l0,h2 CH1 RX Vref:29.3%, TX Vref:21.8%,20.8% DQ rds:h3 l0 h2 h4 l2 h4 l0 h1, h1 h4 h2 h4 h1 h2 h1 l1 DQS rds:h1,l0 CH2 RX Vref:32.2%, TX Vref:20.8%,20.8% DQ rds:h3 l0 h5 h2 l0 h5 l0 h1, h1 h7 h3 h1 h1 h3 h5 h4 DQS rds:h1,h1 CH3 RX Vref:29.7%, TX Vref:22.8%,22.8% DQ rds:h1 l0 l0 h3 h4 l0 h1 h1, l0 h3 h7 l0 l0 h3 h1 h4 stride=0x2, ddr_config=0x4 hash ch_mask0-1 0x20 0x40, bank_mask0-3 0xa00 0x1400 0x2800 0x0, rank_mask0 0x401000 change to F1: 528MHz ch0 ttot10 ch1 ttot10 ch2 ttot10 ch3 ttot10 change to F2: 1068MHz ch0 ttot12 ch1 ttot12 ch2 ttot12 ch3 ttot12 change to F3: 1560MHz ch0 ttot14 ch1 ttot14 ch2 ttot14 ch3 ttot14 change to F0: 2112MHz ch0 ttot16 ch1 ttot16 ch2 ttot16 ch3 ttot16 out U-Boot SPL 2026.04_armbian-2026.04-S88dc-P9ef2-H95ba-Va0a8-B138f-R448a (Apr 25 2026 - 16:56:23 +0000) Trying to boot from SPI ## Checking hash(es) for config config-1 ... OK ## Checking hash(es) for Image atf-1 ... sha256+ OK ## Checking hash(es) for Image u-boot ... sha256+ OK ## Checking hash(es) for Image fdt-1 ... sha256+ OK ## Checking hash(es) for Image atf-2 ... sha256+ OK ## Checking hash(es) for Image atf-3 ... sha256+ OK INFO: Preloader serial: 2 NOTICE: BL31: v2.3():v2.3-868-g040d2de11:derrick.huang, fwver: v1.48 NOTICE: BL31: Built : 15:02:44, Dec 19 2024 INFO: spec: 0x13 INFO: code: 0x88 INFO: ext 32k is not valid INFO: ddr: stride-en 4CH INFO: GICv3 without legacy support detected. INFO: ARM GICv3 driver initialized in EL3 INFO: valid_cpu_msk=0xff bcore0_rst = 0x0, bcore1_rst = 0x0 INFO: l3 cache partition cfg-0 INFO: system boots from cpu-hwid-0 INFO: disable memory repair INFO: idle_st=0x21fff, pd_st=0x11fff9, repair_st=0xfff70001 INFO: dfs DDR fsp_params[0].freq_mhz= 2112MHz INFO: dfs DDR fsp_params[1].freq_mhz= 528MHz INFO: dfs DDR fsp_params[2].freq_mhz= 1068MHz INFO: dfs DDR fsp_params[3].freq_mhz= 1560MHz INFO: BL31: Initialising Exception Handling Framework INFO: BL31: Initializing runtime services WARNING: No OPTEE provided by BL2 boot loader, Booting device without OPTEE initialization. SMC`s destined for OPTEE K ERROR: Error initializing runtime service opteed_fast INFO: BL31: Preparing for EL3 exit to normal world INFO: Entry point address = 0x800000 INFO: SPSR = 0x3c9 U-Boot 2026.04_armbian-2026.04-S88dc-P9ef2-H95ba-Va0a8-B138f-R448a (Apr 25 2026 - 16:56:23 +0000) Model: Xunlong Orange Pi 5 SoC: RK3588S DRAM: 8 GiB Core: 726 devices, 33 uclasses, devicetree: separate MMC: mmc@fe2c0000: 0 Loading Environment from nowhere... OK In: serial@feb50000 Out: serial@feb50000 Err: serial@feb50000 Model: Xunlong Orange Pi 5 SoC: RK3588S Net: eth0: ethernet@fe1c0000 Hit any key to stop autoboot: 0 Scanning for bootflows in all bootdevs Seq Method State Uclass Part Name Filename --- ----------- ------ -------- ---- ------------------------ ---------------- Scanning global bootmeth 'efi_mgr': Cannot persist EFI variables without system partition 0 efi_mgr ready (none) 0 <NULL> ** Booting bootflow '<NULL>' with efi_mgr Loading Boot0000 'mmc 0' failed EFI boot manager: Cannot load any image Boot failed (err=-14) Scanning bootdev 'mmc@fe2c0000.bootdev': 1 script ready mmc 1 mmc@fe2c0000.bootdev.part /boot/boot.scr ** Booting bootflow 'mmc@fe2c0000.bootdev.part_1' with script Boot script loaded from mmc 0:1 244 bytes read in 5 ms (46.9 KiB/s) 31493216 bytes read in 2669 ms (11.3 MiB/s) 41200128 bytes read in 3487 ms (11.3 MiB/s) 165243 bytes read in 43 ms (3.7 MiB/s) Working FDT set to 12000000 Trying kaslrseed command... Info: Unknown command can be safely ignored since kaslrseed does not apply to all boards. Unknown command 'kaslrseed' - try 'help' ## Loading init Ramdisk from Legacy Image at 12180000 ... Image Name: uInitrd Image Type: AArch64 Linux RAMDisk Image (gzip compressed) Data Size: 31493152 Bytes = 30 MiB Load Address: 00000000 Entry Point: 00000000 Verifying Checksum ... OK ## Flattened Device Tree blob at 12000000 Booting using the fdt blob at 0x12000000 Working FDT set to 12000000 Loading Ramdisk to eadf7000, end ecbffc20 ... OK Loading Device Tree to 00000000ece04000, end 00000000ece94fff ... OK Working FDT set to ece04000 [image_setup_libfdt] called [image_setup_libfdt] calling fdt_fixup_ethernet [fdt_fixup_ethernet] called [fdt_fixup_ethernet] alias #0: name='gpio0', path='/pinctrl/gpio@fd8a0000' [fdt_fixup_ethernet] alias #1: name='gpio1', path='/pinctrl/gpio@fec20000' [fdt_fixup_ethernet] alias #2: name='gpio2', path='/pinctrl/gpio@fec30000' [fdt_fixup_ethernet] alias #3: name='gpio3', path='/pinctrl/gpio@fec40000' [fdt_fixup_ethernet] alias #4: name='gpio4', path='/pinctrl/gpio@fec50000' [fdt_fixup_ethernet] alias #5: name='i2c0', path='/i2c@fd880000' [fdt_fixup_ethernet] alias #6: name='i2c1', path='/i2c@fea90000' [fdt_fixup_ethernet] alias #7: name='i2c2', path='/i2c@feaa0000' [fdt_fixup_ethernet] alias #8: name='i2c3', path='/i2c@feab0000' [fdt_fixup_ethernet] alias #9: name='i2c4', path='/i2c@feac0000' [fdt_fixup_ethernet] alias #10: name='i2c5', path='/i2c@fead0000' [fdt_fixup_ethernet] alias #11: name='i2c6', path='/i2c@fec80000' [fdt_fixup_ethernet] alias #12: name='i2c7', path='/i2c@fec90000' [fdt_fixup_ethernet] alias #13: name='i2c8', path='/i2c@feca0000' [fdt_fixup_ethernet] alias #14: name='serial0', path='/serial@fd890000' [fdt_fixup_ethernet] alias #15: name='serial1', path='/serial@feb40000' [fdt_fixup_ethernet] alias #16: name='serial2', path='/serial@feb50000' [fdt_fixup_ethernet] alias #17: name='serial3', path='/serial@feb60000' [fdt_fixup_ethernet] alias #18: name='serial4', path='/serial@feb70000' [fdt_fixup_ethernet] alias #19: name='serial5', path='/serial@feb80000' [fdt_fixup_ethernet] alias #20: name='serial6', path='/serial@feb90000' [fdt_fixup_ethernet] alias #21: name='serial7', path='/serial@feba0000' [fdt_fixup_ethernet] alias #22: name='serial8', path='/serial@febb0000' [fdt_fixup_ethernet] alias #23: name='serial9', path='/serial@febc0000' [fdt_fixup_ethernet] alias #24: name='spi0', path='/spi@feb00000' [fdt_fixup_ethernet] alias #25: name='spi1', path='/spi@feb10000' [fdt_fixup_ethernet] alias #26: name='spi2', path='/spi@feb20000' [fdt_fixup_ethernet] alias #27: name='spi3', path='/spi@feb30000' [fdt_fixup_ethernet] alias #28: name='spi4', path='/spi@fecb0000' [fdt_fixup_ethernet] alias #29: name='ethernet0', path='/ethernet@fe1c0000' [fdt_fixup_ethernet] env var for alias 'ethernet0' is 'ethaddr', value='02:2a:1c:db:3c:a0' [fdt_fixup_ethernet] Patching node '/ethernet@fe1c0000' (offset 40388) with MAC 02:2a:1c:db:3c:a0 [fdt_fixup_ethernet] alias #30: name='mmc0', path='/mmc@fe2c0000' [fdt_fixup_ethernet] SUMMARY: aliases found=31, attempted=1, skipped=0, patched=1 Starting kernel ... 0 Quote
eselarm Posted 44 minutes ago Posted 44 minutes ago 56 minutes ago, zjd said: Starting kernel ... What happens after that? Do you see PCIe getting activated? Is the NVME SSD enumerated? What brand and type of NVME? 0 Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.