On 15.12.2009 19:11 Luis R. Rodriguez wrote: > Hm, it seems best to just add the capability bit that way userspace > can stuff what it wishes and the kernel will only set what is supported. > As is now this would lead to -EOPNOTSUPP but we'd have no way of knowing > from userspace what failed. I could add the capability bit, but I don't understand you completely - do you mean the kernel should just ignore unsupported attributes and leave the checks to userspace? Lukas Turek
Attachment:
signature.asc
Description: This is a digitally signed message part.
