[TUHS] DEQNA / DELQA

[Lawrence Stewart via TUHS]

You might check

https://github.com/calmsacibis995/ultrix11-src/blob/main/sys/net/if_qe.c
which seems to be the Ultrix-11 DEQNA driver.



> On Apr 25, 2026, at 2:44 AM, Folkert van Heusden via TUHS <tuhs at tuhs.org> wrote:
> 
> Hello,
> 
> Am currently working on emulating the DEQNA / DELQA Ethernet card in my 70 emulator.
> 
> My sources of reference are:
> 
> * https://www.vaxhaven.com/images/5/50/EK-DEQNA-UG-001.pdf
> * https://github.com/RetroBSD/2.11BSD/ for the (431 patchlevel) BSD 2.11 sourceces
> 
> Am I right that  the DEQNA/DELQA keep scanning the ring-buffers by itself for packets to transmit?
> * If so: wouldn't that slow down the whole system? (assuming a shared bus with the cpu) And how many times per second?
> * If not: how does BSD (or whatever other OS runs on the PDP) notify the card that there is something queued? An LLM told me that for the DELQUA, there is a bit (12 I believe, that would be "QE_POWERUP 0x1000") in the CSR to tell the card to start transmitting, but I don't see that happening in the bsd driver code (it only detects if it s a DEQNA or DELQA card to print a message on the console, but nothing more).
> 
> Furthermore I use https://www.retro11.de/data/oc_w11/oskits/211bsd_rpethset.tgz from https://wfjm.github.io/home/w11/inst/systems.html#h_os_kits to test it. I changed the kernel to also support the deqna and when looking at my traces, I see that bsd sets the receive buffer addres and then polls CSR twice. Then nothing for a second or so and then the whole procedure of configuring the device happens again so it expects something to happen, but what? An interrupt?
> 
> regards
> 
> -- 
> www.vanheusden.com [1]
> 
> Links:
> ------
> [1] http://www.vanheusden.com