On 25/05/2022 12:11, Hans Schultz wrote: > On ons, maj 25, 2022 at 11:38, Nikolay Aleksandrov <razor@xxxxxxxxxxxxx> wrote: >> On 25/05/2022 11:34, Hans Schultz wrote: >>> On ons, maj 25, 2022 at 11:06, Nikolay Aleksandrov <razor@xxxxxxxxxxxxx> wrote: >>>> On 24/05/2022 19:21, Hans Schultz wrote: >>>>>> >>>>>> Hi Hans, >>>>>> So this approach has a fundamental problem, f->dst is changed without any synchronization >>>>>> you cannot rely on it and thus you cannot account for these entries properly. We must be very >>>>>> careful if we try to add any new synchronization not to affect performance as well. >>>>>> More below... >>>>>> >>>>>>> @@ -319,6 +326,9 @@ static void fdb_delete(struct net_bridge *br, struct net_bridge_fdb_entry *f, >>>>>>> if (test_bit(BR_FDB_STATIC, &f->flags)) >>>>>>> fdb_del_hw_addr(br, f->key.addr.addr); >>>>>>> >>>>>>> + if (test_bit(BR_FDB_ENTRY_LOCKED, &f->flags) && !test_bit(BR_FDB_OFFLOADED, &f->flags)) >>>>>>> + atomic_dec(&f->dst->locked_entry_cnt); >>>>>> >>>>>> Sorry but you cannot do this for multiple reasons: >>>>>> - f->dst can be NULL >>>>>> - f->dst changes without any synchronization >>>>>> - there is no synchronization between fdb's flags and its ->dst >>>>>> >>>>>> Cheers, >>>>>> Nik >>>>> >>>>> Hi Nik, >>>>> >>>>> if a port is decoupled from the bridge, the locked entries would of >>>>> course be invalid, so maybe if adding and removing a port is accounted >>>>> for wrt locked entries and the count of locked entries, would that not >>>>> work? >>>>> >>>>> Best, >>>>> Hans >>>> >>>> Hi Hans, >>>> Unfortunately you need the correct amount of locked entries per-port if you want >>>> to limit their number per-port, instead of globally. So you need a >>>> consistent >>> >>> Hi Nik, >>> the used dst is a port structure, so it is per-port and not globally. >>> >>> Best, >>> Hans >>> >> >> Yeah, I know. :) That's why I wrote it, if the limit is not a feature requirement I'd suggest >> dropping it altogether, it can be enforced externally (e.g. from user-space) if needed. >> >> By the way just fyi net-next is closed right now due to merge window. And one more >> thing please include a short log of changes between versions when you send a new one. >> I had to go look for v2 to find out what changed. >> > > Okay, I will drop the limit in the bridge module, which is an easy thing > to do. :) (It is mostly there to ensure against DOS attacks if someone > bombards a locked port with random mac addresses.) > I have a similar limitation in the driver, which should then probably be > dropped too? > That is up to you/driver, I'd try looking for similar problems in other switch drivers and check how those were handled. There are people in the CC above that can directly answer that. :) > The mayor difference between v2 and v3 is in the mv88e6xxx driver, where > I now keep an inventory of locked ATU entries and remove them based on a > timer (mv88e6xxx_switchcore.c). > ack > I guess the mentioned log should be in the cover letter part? > Yep, usually a short mention of what changed to make it easier for reviewers. Some people also add the patch-specific changes to each patch under the --- so they're not included in the log, but I'm fine either way as long as I don't have to go digging up the old versions. > >>>> fdb view with all its attributes when changing its dst in this case, which would >>>> require new locking because you have multiple dependent struct fields and it will >>>> kill roaming/learning scalability. I don't think this use case is worth the complexity it >>>> will bring, so I'd suggest an alternative - you can monitor the number of locked entries >>>> per-port from a user-space agent and disable port learning or some similar solution that >>>> doesn't require any complex kernel changes. Is the limit a requirement to add the feature? >>>> >>>> I have an idea how to do it and to minimize the performance hit if it really is needed >>>> but it'll add a lot of complexity which I'd like to avoid if possible. >>>> >>>> Cheers, >>>> Nik
