On Sat, Oct 24, 2015 at 1:57 PM, Eric Wong <[email protected]> wrote: > > On Thu, Oct 15, 2015 at 20:56:46, Edward Wingate wrote: > >> On Thu, Oct 15, 2015 at 6:25 AM, Mike Looijmans <mike.looijmans at > topic.nl> wrote: > >> As the datasheet should have explained, "0" is not a valid PHY address. > So > >> that does not leave many options. "0" is actually the "broadcast" > address on > >> the MDIO bus, so all phys will respond to it. Which makes it work with > just > >> one PHY, but with multiple PHYs on the MDIO bus this will not work. > > > >The datasheet itself never mentions 0 as being the broadcast address. > >I wonder if the GMII-to_RGMII IP responds to 0, or only its assigned > >PHY address. > > We have an almost identical configuration to yours as you've been > describing it and were able to get the 2nd ethernet port working by > adding this to the sclr-clkc device tree: > clocks = <&clkc 17>; > clock-names = "gem1_emio_clk"; >
A few months back, I mentioned the above, that adding those 2 lines to sclr/clkc device tree made the 2nd ethernet port work. Can anyone tell me exactly what they're doing? If I weren't using Linux and its device tree, what would be the bare-metal Xilinx SDK equivalent way to configure whatever those two lines are doing? Appreciate any help! > It is working with both PHY addresses being 0, and the converter IP > block has PHY address of 3. However, the PL converter IP block had to > be configured to use an external clock with the FPGA2 clock feeding > it. The DT entry above will then allow the FPGA2 clock frequency to > be set when eth1 interface is initialized. I don't know why an > external clock had to be used vs. the default clock in that's already > going to the converter IP block, but just couldn't get it to work with > the default clock. >
-- _______________________________________________ meta-xilinx mailing list [email protected] https://lists.yoctoproject.org/listinfo/meta-xilinx
