Nice find, thank you @slh... I will need to read through that in detail to pick out the network kmods. From the skimming I just did, reply #80 calls out:
@iplaywithtoys - I do not need wifi on the box but I suppose it would be nice to have for just scanning purposes but I saw your recent replies in that thread indicating (as of 26-Jun-2023) no support for the WiFi stuff (RTL8821CE). Would you mind posting the output of the following on yours assuming it is up and running?
It's not up and running right now, but I did reset it to factory defaults after posting in the other thread. So whatever kmods are installed are those which ship with the x86_64 stable 22.03.5 image. And those are good enough for the on-board Ethernet and on-board storage to work. And, I assume, for USB storage to work, since I booted the thing from a USB stick containing OpenWRT before writing OpenWRT to the internal SSD.
I am OK with having the image run via squashfs so I can avoid fsck issues on a live ext4 partition. What are pros/cons of using openwrt-x86-64-generic-squashfs-combined-efi.img.gz vs openwrt-x86-64-generic-squashfs-combined.img.gz for this hardware?
OK, so it doesn't matter. Next question is I am building my own image with custom packages and with CONFIG_TARGET_ROOTFS_PARTSIZE=512 which despite being squashfs, should give me 0.5 GB of root partition space. Any downsides to not using the ext4 "live" filesystem? I believe using squashfs will protect the media from potential FS damage in a situation of ungraceful shutdown.
I will reinstall it from 0 and with no "extra" packets , now i had installed OpenWRT in that TP-Link RE200, configured it with the wifi hotspot and then connecting it to the x86 (eth0) and then connected my laptop to eth1. all working, so... ill reinstall all, not for nothing, just to not have S*it in that system.
I got the same unit. Booted from USB to check performance. Basic routing seems underwhelming.
For those knowledgeable: should running from USB cause performance issues? I had assumed all is in RAM memory, so the boot medium would not be a bottleneck once the system is up…
Huh. That's disappointing. I haven't actually put mine into my network yet; I've been fiddling with different "can I get this to work?" scenarios so I've not tested its performance yet.
Still, thanks for the heads-up. I'll see if I can set aside some time to put it through its paces at the weekend.
On what do you base this statement? I don't have enough time to get mine setup for my snort3 testing, but iperf3 seems fine.
EDIT: I removed the iperf3 results here because I realized that the test was flawed since the RPi4B does not go through an extra switch as the t8plus does. I need to get the t8plus setup properly and then replace it physically where the RPi4B is for an accurate comparison.
I am curious what data @fodiator generated to support the "underwhelming" assertion.