> From: Long Li [mailto:[email protected]] > Sent: Friday, 10 April 2026 23.14 > > > On Fri, 10 Apr 2026 16:34:56 +0200 > > Morten Brørup <[email protected]> wrote: > > > > > Is there a workaround until this patch is added, if I don't want > rte_eal_init() to > > probe any devices, but I want my application to probe devices later > by calling > > rte_bus_scan() followed by rte_bus_probe()? > > > > > > It's not something we really need, so if you don't know it off the > top of your > > head, don't spend time trying to figure it out. > > > > > > > > > Venlig hilsen / Kind regards, > > > -Morten Brørup > > > > We do need it because of the problems with vdev_netvsc. > > The issue is that vdev_netvsc tries to setup failsafe on all netvsc > devices it sees, > > and it determines which ones are used by OS based on somewhat fragile > OS > > probes with netlink. > > > > I would recommend getting rid of vdev_netvsc, or make the the auto > scanning > > there something that has to be forced by devargs. > > But that really is up to the netvsc maintainers. > > Hi Morten, > > Can you use --vdev=net_vdev_netvsc,ignore=1? The ignore devarg is > already supported by the vdev_netvsc PMD — when set, the probe function > skips all netvsc interface discovery and returns early. > > Long
Thanks for the suggestion, Long. I'm not looking for a netvsc solution, I'm looking for a general solution to not probe any drivers at EAL initialization, but at a later stage, controlled by the application. -Morten
