Post Go back to editing

2xadrv9009 on ZCU102



We are currently working on testing sync using 2xadrv9009 dev boards on the ZCU102. 

We have set up the clock routing and I believe this is working ok. 

I am currently looking at getting the linux side sorted using the adrv9009-zu11eg branch from github. 

I am hitting probe errors when initializing the XCVR IP which I think may be due to coordinating start up of the two chips.

I'm trying to understand the main adrv9009 probe for two devices. In the Talise SOM device tree one of the devices is designated as adrv9009-x2 and the other just as adrv9009. It seems though that when using a combined link layer, it won't be possible to get through the full probe as only one of the Talise chips will have been initialised during the first probe, and so the JESD links of the other one will not have been initialised. Is the order of the two adrv9009 entries in the device tree important?

It maybe that the driver support for 2xadrv9009 isn't working yet, which would be useful to know.



Top Replies

  • Hey,

    Sorry for the late reply.

    Is the order of the two adrv9009 entries in the device tree important?

    Normally it shouldn't be, but sometimes it is important.

    This is usually due to how Linux works. If we find that something depends on the order in the device-tree, we try to fix it in the driver.

    The adrv9009-zu11eg branch from github has been merged to master.

    Please try that as a reference.

    If you encounter any issues, feel free to tell us.

    We will do our best to look into it.