sqm and hardware flow-offloading does not work, sqm and software flow-offloading is technically working, but you lose most of the benefit of the software flow-offloading that way - and this hardware isn't fast enough to do sqm/cake at 1 GBit/s line-speed, far from it.
filogic 820/ 830 might do, x86_64 (e.g. alderlake-n n100) certainly will - RPi4 or the faster rockchip SOCs (the newer Nano-Pis) should do.