Environment:
- Router (DHCP source)
- Raspberry Pi (as server)
- three custom devices with u-boot on board
Desired system behavior:
When the server is running. I'm executing reset at the same time (i.e. by one switch) on all three devices.
- Devices are booting up
- u-boot is starting
- dhcp command is executed
- all three devices are receiving IP's
- In next step "run bootcmd_pxe" command is executed automatically
- Kernel and Device tree are loaded and system is starting
Problem description:
When I'm executing the reset operation on a single board all is working fine in 99% of situations, so it is fine for me.
The problem is when I'm trying to do reset operation on several boards simultaneously.?All the above steps with "run bootcmd_pxe" are working properly. The problem is in the last step, because hardly ever instead of always (i.e. in 1/~50 of power cycles/resets) zImage and DT are loaded properly (almost "parallel"). Then the system is getting up similarly at the same time and "Starting kernel ..." is occur - and that is the desired behavior. Unfortunately, it works only in around in 1/50 of resets cycles.
In most cases, one device is getting up properly and the rest of them are trying to load zImage or it loading it with timeouts. After that, booting up process is stopping and the devices are going to u-boot command prompt mode.
Schematic:
question from:
https://stackoverflow.com/questions/65853565/tftp-configuration-tftp-nfs-u-boot-problem-with-parallel-boot-up-of-two 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…