|
|
Log in / Subscribe / Register

BPF: what's good, what's coming, and what's needed

BPF: what's good, what's coming, and what's needed

Posted May 12, 2019 5:14 UTC (Sun) by alison (subscriber, #63752)
Parent article: BPF: what's good, what's coming, and what's needed

> A file like /dev/bpf could be used for access control, but that's still > pretty coarse; perhaps what is needed is a hierarchy of files
> describing the different program types and their access
> permissions.

Is there some reason why sysfs is not the obvious answer? Maybe /sys/kernel/debug/bpf, or /sys/kernel/bpf? That would certainly clarify the ABI stability question! The solution would be to keep new BPF interfaces in staging branch until we're sure we'll keep them.


to post comments

BPF: what's good, what's coming, and what's needed

Posted May 12, 2019 19:12 UTC (Sun) by jkowalski (guest, #131304) [Link]

Yeah, either way a hierarchy of objects instead of a device node (and then having to use ioctl to lookup an object by name or something uglier) is a much nicer model (and one could argue should have been for a lot of other kernel interfaces), as it helps to use the descriptor to select the object in question, and access control can be done through the filesystem itself.


Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds