Hi all together,
Firstly, despite having long experience in developing embedded systems, this is my very first project with OPENWRT and I never build an OPENWRT image up to now.
Secondly - let me describe my goals:
I am working for a company, which is building airborne geophysical exploration devices and i am mainly developing hard and software for this kind of systems.
For an autonomous and spatially distributed sensor network which would use ETHERNET/POE for communication with and powering the connected sensors, we where looking for an "intelligent" POE-switch with a certain computing power which enables it to be the central data acquisition device for the sensor network running our own acquisition software.
The question was whether to develop our own hardware or to find an off-the-shelve device which suits (most of) our requirements.
The closest fit in hardware was the RB5009 with its powerful MARVELL ARMADA, the 120W POE option, USB 3.1 storage and the opportunity to run OPENWRT as LINUX OS which could also be able to run our own software.
So we ordered one devices to test the flashing and to test the USB storage speed.
I must say, it was relatively easy to get OPENWRT (release 24.10.0) running and i was a bit pleased, how easy it is to install additional software packages using OPKG.
At this point, i must say, that USB storage would not be our preferred storage option but as the speed is good enough, we would come around with that.
Much more interesting would be a PCIe/NVME disk and as i have seen that there is an unpopulated NVME header i was quite happy
So i ordered the appropriate connector and put it onto the board adding some R's and C's to be ready for an NVME test.
But unluckily, there is obviously no PCIe/NVME support in the kernel
So my big questions would be:
- How can i get PCIe/NVME support (for ARMADA 7040) in the kernel image?
- Or, is there someone who could compile me an image with this stuff in?
- Has anybody else tried to get the NVME on RB5009 working?
Any help or any hint is appreciated!
Many thanks in advance and best regards,
Marco.