On 28 Feb 2022, at 10:16, Trond Myklebust wrote: > On Mon, 2022-02-28 at 09:43 -0500, Benjamin Coddington wrote: >> On 27 Feb 2022, at 18:50, NeilBrown wrote: >>> I liked this patch at first, but I no longer think it is a good >>> idea. >>> >>> It is quite possible today for containers to provide sufficient >>> uniqueness simply by setting a unique hostname before the first NFS >>> mount. >>> Quite possibly some containers already do this, and are working >>> perfectly. >>> >>> If we add new randomness, the they will get a different identifier >>> on >>> each boot. This is bad for exactly the same reason that it is bad >>> for >>> "net == &init_net". >>> >>> i.e. I believe this patch could cause a regression for working >>> sites. >>> The regression may not be immediately obvious, but it may still be >>> there. >>> For this reason, I think this patch should be dropped. >> >> I agree, Trond please drop this patch. >> >> > Since it was already pushed to linux-next, it will have to be reverted. > I'm not seeing it on any branch on linux-next. I presume no one's pulled your linux-next branch. Are you not able to drop it from your linux-next branch? Ben
