On 04/10/2018 04:27 PM, Martin Braun wrote:
> On 04/10/2018 09:14 AM, Rob Kossler via USRP-users wrote:
>> Hi,
>> Within the last couple of weeks, the N310 tree went from having 1 front
>> end to 2 front ends per daughterboard (A,B,C, or D).  What is the reason
>> for this?  Why would the user want to choose subdev spec "A0" as opposed
>> to "A1" and same for other paths?
>> I attached the uhd_usrp_probe results from the most recent uhd version
>> (2c9087d) as well as one from a couple of weeks ago (a1b5c4a) showing
>> the difference in the trees.
>> Rob
> Hey Rob,
> you found a bug. Let me clarify:
> The subdev spec is a bit of an artefact from the past, so we're trying
> to bridge the USRP stone age with the N310 future here. For that reason,
> the four possible values for your subdev spec on the N310 are:
> A:0 A:1 B:0 B:1
> ...and they correspond to the front panel labels:
> RF0 RF1 RF2 RF3
> The A and B signify both the physical daughterboard, as well as the
> corresponding RFNoC block and RFNoC block controller (in software).
> When we first released, we had A:0 B:0 C:0 D:0 for various reasons.
> However, we wanted to change it to what we have now mostly because it
> helps remind people that A:0 and A:1 are on the same hardware, share an
> LO, and so on. When we did change it, we made a mistake that wasn't
> caught until now.
> We're working on a fix, as well as a clarification in the manual right now.

Actually, I need to add some more clarifications:

- When I said we moved from "A:0 B:0 C:0 D:0" to "A:0 A:1 B:0 B:1", I
meant on master branch. On maint branch, the former is still correct
(and will stay like that for the entirety of the 3.11 release cycle).
- There still is a bug on maint that'll produce what you saw (and we
have a fix for it)


USRP-users mailing list

Reply via email to