Hi, On Fri, Sep 13, 2024 at 3:53 PM Stefan Schmidt <stefan@xxxxxxxxxxxxxxxxxx> wrote: > > Hello, > > On 9/13/24 12:59 AM, Alexander Aring wrote: > > Hi, > > > > On Mon, Sep 9, 2024 at 8:19 PM James Hanley <jhanley@xxxxxxxxxxxx> wrote: > >> > >> Has there been any effort to understand the changes needed to include/net/ieee802154.h and associated files within drivers/net/ieee802154 to support the ratification of 15.4-2020? One prime example is the "Extended PHR" bit which was previously reserved to now allow extend the PHR of 2 more octets giving bits 8-11 to be used for "Frame Length MSB" and bits 12-15 marked as "Reserved" - this in combination of the legacy PHR bits 0-6 labeled as "Frame Length LSB" now allows for a frame MTU of 2048 octets. > >> > >> The 802.15.4-2020 is available individually free of charge through the IEEE website through the IEEE Get Program. https://ieeexplore.ieee.org/document/9144691 > >> > >> Is there a way to prototype these new changes to the spec with the mac802154_hwsim driver? > >> > > > > mac802154_hwsim driver uses mac802154 the SoftMAC implementation. > > There are quite more changes necessary as the whole mac802154 stack > > deals with static 127 bytes MTU defines, etc. Unfortunately, it isn't > > just a driver change. > > I understand it the way that James actually wanted to try prototyping > stack changes and verify with hwsim. James, could you clarify? > Why shouldn't this be possible? Of course it should and I would definitely want to see it when adding any support for that for the rest that's being involved. The "rest" is probably most of the work. > To answer your question, we currently have no support for any of the > newer 802154 specs. :/ Bigger MTU was brought up before (IIRC in the > subGHz context) but nobody started to actually work on it. > > We are happy to take changes in, but currently we have no plans on our > side to get this going. > yea, just send patches. I am open to starting with hwsim to support kind of such a "bigger mtu" phy "type" as long it doesn't break the kernel and finding the spots that need to be somehow changed. - Alex
