On Wed, Oct 02, 2024 at 05:19:49PM +0200, Christoph Hellwig wrote: > > Nothing prevents future improvements in that direction. It just seems > > out of scope for what Kanchan is trying to enable for his customer use > > cases. This patch looks harmless. > > It's not really. Once we wire it up like this we mess up the ability > to use the feature in other ways. Additionally the per-I/O hints are > simply broken if you want a file system in the way as last time, > and if I don't misremember you acknowledged in person last week. You remember right. I also explained the use cases for per-io hints are to replace current passthrough users with generic read/write paths that have all the memory guard rails, read/write stats, and other features that you don't get with passthrough. This just lets you accomplish the same placement hints with the nvme raw block device that you get with the nvme char device.