On Mon, Oct 28, 2024 at 11:27:33AM -0700, Bart Van Assche wrote: > On 10/25/24 2:36 PM, Keith Busch wrote: > > When multiple partitions are used, you may want to enforce different > > subsets of the available write hints for each partition. Provide a > > bitmap attribute of the available write hints, and allow an admin to > > write a different mask to set the partition's allowed write hints. > > After /proc/irq/*/smp_affinity was introduced (a bitmask), > /proc/irq/*/smp_affinity_list (set of ranges) was introduced as a more > user-friendly alternative. Is the same expected to happen with the > write_hint_mask? If so, shouldn't we skip the bitmask user space > interface and directly introduce the more user friendly interface (set > of ranges)? I don't much of have an opinion either way. One thing I like for the bitmask representation is you write 0 to turn it off vs. the list type writes a null string. Writing 0 to disable just feels more natural to me, but not a big deal.
