Scott Gilbertson wrote: >> To make your local patch to rt_e100.c as small as possible, I would >> suggest to let the custom FPGA driver export some callback registration >> interface, and then simply exchange rtdm_irq_request/free with that one. >> I don't see that this callback requires task context, but if it really >> shall be like that, you will also need to patch the affected >> rtdm_lock_get/put in rt_e100 for use outside IRQ context. >> > > How about something (ioctl or whatever) in e100 that lets you get the > semaphore handle and free the IRQ, and something in the FPGA driver that > lets you pass in the semaphore handle. That way we don't care which starts > first (FPGA driver or e100). We'd make a little program you can run to flip > between using the e100 ISR or the FPGA ISR. When you first load e100, it > would always use its own ISR, so the only local patch would be the ioctl (or > whatever) that disables it. >
Well, it's up to you. I just sketched one possible path that is minimal invasive to a mainline rt_e100 so that you can also pull future RTnet releases and patch them easily to your scenario. Jan
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------------
_______________________________________________ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users
