Hi, Mathieu, On 2/5/20 3:21 PM, Mathieu Poirier wrote: > Hi Andrew, > > On Wed, Feb 05, 2020 at 02:24:07AM -0800, Andrew P. Lentvorski wrote: >> I seem to be missing something obvious: Where do I get the values for >> src and dst in the pru_rpmsg_send() function? >> >> The TI examples do a pru_rpmsg_receive() first and then echo back to the >> same src and dst. That seems ... odd. What's the point of having a >> variable src and dst if they are effectively hardwired to a single >> communication channel? > > Since pru_rpmsg_send/receive() don't exist in the mainline kernel you will have > to be more specific... What TI examples are you referring to? Where can we find > them? TI keeps the examples here: https://git.ti.com/cgit/pru-software-support-package/pru-software-support-package/ This is a good example of the rpmsg system being used with the echo I talked about: https://git.ti.com/cgit/pru-software-support-package/pru-software-support-package/tree/examples/am335x/PRU_RPMsg_Echo_Interrupt1 That example works on the Beaglebone Black (an am3359 based Linux board). > There are other rpmsg_sendxyz() functions that allow you to send messages to > different addresses[1]. Given the amount of information, I'm afraid there isn't > much else I can add at this time. > > Thanks, > Mathieu > > [1]. https://elixir.bootlin.com/linux/latest/source/include/linux/rpmsg.h#L125 Thanks for the pointers. It looks like I need to drag the src/dst out of the rpmsg_device or rpmsg_channel_info on the Linux side and then pass it to the PRU side by some method. An initial message is as good as any and better than most. I completely flaked that this wouldn't be in mainline. Even a simple thought about the name "*pru*_rpmsg_send" should have set off alarm bells. Sorry about that. Thanks for the help, -a
