On Wed, Jul 03, 2024 at 06:58:12AM +0200, Christoph Hellwig wrote: > On Tue, Jul 02, 2024 at 09:55:39PM -0700, Darrick J. Wong wrote: > > > Good question. As far as I can tell there is no simply ioctl for it. > > > I really wonder if we need an extensible block topology ioctl that we > > > can keep adding files for new queue limits, to make it easy to query > > > them from userspace without all that sysfs mess.. > > > > Yeah. Or implement FS_IOC_GETSYSFSPATH for block devices? :P > > I know people like to fetishize file access (and to be honest from a > shell it is really nice), but from a C program would you rather do > one ioctl to find a sysfs base path, then do string manipulation to > find the actual attribute, then open + read + close it, or do a single > ioctl and read a bunch of values from a struct? Single ioctl and read from a struct. Or single ioctl and read a bunch of json (LOL) I wish the BLK* ioctls had kept pace with the spread of queue limits. > > > > That's going to take a day or so, I suspect. :/ > > > > > > No rush, just noticed it. Note that for the discard granularity > > > we should also look at the alignment and not just the size. > > > > <nod> AFAICT the xfs discard code doesn't check the alignment. Maybe > > the block layer does, but .. > > The block layer checks the alignment and silently skips anything not > matching it. So not adhering it isn't an error, it might just cause > pointless work. <nod> --D
