On 3/8/20 9:19 AM, Yan Yankovskyi wrote: > On Sat, Mar 07, 2020 at 02:41:44PM -0500, Boris Ostrovsky wrote: > >> I also think that, given that this patch is trying to get types in >> order, find_virq() will need more changes: it is supposed to return >> evtchn_port_t. But then it also wants to return a (signed) error. > > As we don't care which error we got during find_virq call, we can just > return 0 in case of error, and port number otherwise. Port 0 is never > valid, so this approach can work for the other functions as well. > On the other hand, passing port using pointer and returning actual > error message, as it's done in xenbus_alloc_evtchn(), sounds like a > better approach overall. What do you think?
Either way should be fine. (BTW, you also need to update evtchn_from_irq(), I think you missed it). -boris _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel