Use an SD card to install a new firmware.
The following commands write the image from the host system to the SD card:
$ cd <BSP project>/platform-<platform_name>/images $ sudo dd if=@bsp_image_name@-<build Timestamp>.rootfs.wic of=/dev/sdc bs=1M conv=fsync #Assuming the SD card is assigned to /dev/sdc
The image must be written to the RAW device of the SD Card (e.g. /dev/sdc) and not into a partition (e.g. /dev/sdc1) !
To identify the SD card in Linux the shell command dmesg can be used:
$ dmesg | tail -n 15
$ dmesg | tail -n 15 #plug in SD Card into the reader $ dmesg | tail -n 15
One approach to update the firmare components U-Boot,Linux kernel and devicetree in a running system it to load them from a tftp server in U-Boot.
This process requires a running TFTP server in your network, please see the following page how to setup TFTP Server.
After setting up the TFTP server, the binaries to be updated must be copied into the TFTP directory.
1. Setup the Starterkit to boot from eMMC or SD
2. Connect the kit to the network with TFTP server supplying the binaries via ethernet on connector@eth_prime@
3. Power up the system and interrupt the boot process in U-Boot
Hit any key to stop autoboot: 0 =>
3. Setup the ethernet interface:
Set network settings using DHCP
4. Set U-Boot variable mmcdev according to the drive you want to write to.
=> setenv mmcdev <device_number>
5. Set the file name of the binary to update in U-Boot :
The binaries (U-Boot, devicetree, Linux kernel) that are generated when building the the BSP are listed in section Images Yocto.
6. Perform Update by running the update command:
=> run update_uboot
=> run update_fdt
=> run update_kernel
The U-Boot environment variables can be saved with the command saveenv to keep the values permanently.
2. Save the U-Boot environment vari with the saveenv command (optional):
=> saveenv
3. run the following command to start the update procedure:
=> run install_firmware
To boot the @mod_name@ from network you need a working bootloader in eMMC/SD-card or SPI-NOR (placement option on @mod_name@ ) which is able to get the kernel image over tftp and to provide the kernel with commandline settings for NFS. The dtb-file and kernel image have to be provided via tftp and the rootfs via nfs.
The bootloader environment needs to be modified to work with your tftp-server and your nfs-server.
1. Prepare network interface:
Set network settings using DHCP
2. set the Uboot variables for TFTP and NFS:
3. Run the uboot script to boot from nfs: