On Tue, Mar 25, 2025 at 04:19:31PM -0600, Uday Shankar wrote: > There are a couple of places in the kublk selftests ublk server which > use the legacy ublk opcodes. These operations fail (with -EOPNOTSUPP) on > a kernel compiled without CONFIG_BLKDEV_UBLK_LEGACY_OPCODES set. We > could easily require it to be set as a prerequisite for these selftests, > but since new applications should not be using the legacy opcodes, use > the ioctl-encoded opcodes everywhere in kublk. Is it required to allow for the building of old userspace code (using legacy opcodes) against new kernel headers? Or do we only need to guarantee that old userspace code using legacy opcodes that is already compiled continues to work against newer ublk_drv? If it's the latter case, maybe we can consider removing the legacy opcode definitions from the userspace header as a follow-on change?
