21 Nov
2024
21 Nov
'24
8:55 p.m.
On 21/11/2024 19:20, Sven Eckelmann wrote:
On Thursday, 21 November 2024 14:14:13 CET Antonio Quartulli wrote:
I am wondering if acquiring a lock around the entire section can make sense. Also in this case it would cover the case where we get an ADD in between.
What do you think?
So you would change the hash with its per-bucket spinlocks to a global spinlock? Just imagine that I am now locking angry at you. Especially because I still hear the people asking for a migration from the batadv hash the rhashtable.
argh - true. We have a per-bucket lock and not one lock for the entire table, hence my suggestion is not valid.
Ok, I take that back :)
--
Antonio Quartulli