On Sat, 2023-09-02 at 12:08 -0700, Linus Torvalds wrote: > On Sat, 2 Sept 2023 at 00:39, James Bottomley > <James.Bottomley@xxxxxxxxxxxxxxxxxxxxx> wrote: > > > > Updates to the usual drivers (ufs, lpfc, qla2xxx, mpi3mr, libsas) > > and the usual minor updates and bug fixes but no significant core > > changes. > > Removing 3000+ lines for UFS HPB support wasn't even worth > mentioning? Sorry, my bias is showing. I always thought HPB was a useless marketing feature and was against including it, so removing it is just the world going back to being more correct. To be honest, I also didn't think you cared at all about UFS ... > I am happy to see it gone, and maybe as a technology it was a failure > not worth it, but as a "we gave up on it as being worthless" might > still have been worth a word or two.. > > Sadly, I see from the commit message that apparently the next stage > is going to involve zoned storage. Now *there* is a technology that > seems to be a complete failure, brought to us by the same kind of > failed hardware people who tried to convince us that we should care > about 64kB pages in SSD's. Well, we put HPB in in spite of most of us thinking it was a bad idea because there was a tiny chance it might work and it could be yanked out if it didn't. There's no reason not to extend the same courtesy to zoned storage. > Oh well. With enough thrust, even a pig will fly. I suspect that's > the motivating factor behind all those zoned storage things too. That's why RFC1925 exists, yes ... James
