Le 13/04/2017 à 15:29, Johannes Berg a écrit : > On Thu, 2017-04-13 at 15:27 +0200, Nicolas Dichtel wrote: >> >>> Yes, some - very few - families still insist on using attribute 0, >>> perhaps parsing by hand or so. Like you say though, the entire >>> infrastructure makes that hard and undesirable, so I don't really >>> see >>> why we need to invest the extra code/work into making it work >>> *here*, >>> especially since it's such a corner case as I described in my other >>> email. >> >> Here is an example: >> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/co >> mmit/?id=31e20bad8d58 >> >> I also see one in openvswitch (I will send a similar patch), but >> there are probably some others. > > Yeah. I'm not really sure what the point of such a patch is though - > the API is set now, and can't really be changed. The goal is to avoid copy and paste error, like it was done in diag subsystem. > > Anyway, the ones you point out are only used for *output* by the > kernel, so wouldn't be affected by any "missing attribute" reporting > anyway. Sure. It was just to mention that attribute 0 exists somewhere. The other 0 attribute is OVS_TUNNEL_KEY_ATTR_ID. But I agree with you that it remains a corner case. Nicolas
