Hi, thanks for the response.
Have learnt a bit about the perils of using an access point in 'client' mode/wireless bridge. Apparently (from openwrt, best explanation:
https://wiki.openwrt.org/doc/howto/clientmode#bridgedclientmodeissues) 802.11 doesn't nicely handle MACs across more than 1 access point. Now got two additional wireless access points (TPlink 801nd and 701nd) set up one with master and other as slave in 'WDS' mode - this properly handles MACs and now any device plugged into the slave AP is properly acting with regards to dhcp and so on.
I had managed some success behind the original setup if I set my board up to 'receive all' ie MAC filtering in hardware disabled. This would then happily work so long as I set a static IP as DHCP responses still weren't getting passed back. Now I have the above wireless AP problem solved I have disabled receive all, back to default HAL driver settings.
I am now focussing back on LWIP - device is pingable and a very simple single page web server (using netconn API) is working fine. Only issue I'm having now is a cyclical ping response - every 6 pings increase in value ie:
10/22/29/57/70/88/7/22/38/45/68/78. I can't see any reason in my code - freeRTOS tasks ethif is running at highest priority, and cpu usage is stable at about 21%.