On 2026-02-09 10:43:55 [+0000], Vadim Fedorenko wrote: > On 09/02/2026 09:06, Sebastian Andrzej Siewior wrote: > > On 2026-02-08 11:25:40 [-0500], Willem de Bruijn wrote: > > > > > > But it's more like a question to maintainers whether it is > > > > > > acceptable > > > > > > way of "fixing" drivers or it's no-go solution > > > > > > > > > > Requiring OPT_TSONLY unless CAP_NET_RAW would break legacy users. > > > > > > > > Well, they are kinda broken already. Without OPT_TSONLY and CAP_NET_RAW > > > > all TX > > > > timestamps are silently dropped. > > > > > > Are you referring to sysctl_tstamp_allow_data? > > > > > > That is enabled by default. > > > > Yes. If so, then we don't need the check below which requires > > sk_callback_lock. > > > > Are SIOCSHWTSTAMP the legacy users or the ones which do not set > > OPT_TSONLY? > > > > I would suggest to move the CAP_NET_RAW check to the point where > > timestamping is getting enabled. > > Also if ndo_hwtstamp_set is the preferred method of getting things done, > > I could check how many old ones are can be easily converted… > > Looks like you are mixing things. SIOCSHWTSTAMP/ndo_hwtstamp_set are HW > configuration calls while OPT_TSONLY is socket option, which is setup via > setsockopt, you can find points searching for > SOF_TIMESTAMPING_OPT_TSONLY in the sources, basically > sock_set_timestamping() is the function to check
Yeah, but what is the legacy user here? If you enable HW-timestamps but never set OPT_TSONLY and the sysctl is also 0 then you reply on the CAP_NET_RAW later on. Right? I just try to justify the CAP_NET_RAW check and if it is required to move it earlier (where HW timestamps are enabled). And if the sysctl check is enough then maybe it is not needed. > > > > To receive these timestamps users have to get > > > > CAP_NET_RAW permission, and it will work with the updated logic as > > > > well... Sebastian
