On Thu, 29 Aug 2024 at 19:41, Aleksandr Mikhalitsyn <aleksandr.mikhalitsyn@xxxxxxxxxxxxx> wrote: > Let's think about it a bit more and if you confirm that we want to go > this way, then I'll rework my patches. And ACK from Christian would be good. I don't see why this would be a radical change: idmapping changes the meaning of fsuid/fsgid, so we now can't send those to the server in the non-create case. The only thing that changes is that an "invalid ID" value is introduced into the protocol. Probably makes sense to explicitly define this valiue in <uapi/linux/fuse.h>. Thanks, Miklos
