Thank you, will follow the updated commit closely.

On SANICAP verification — that would be a valuable addition. The specific gap it addresses: a drive that does not support the requested sanitize action (CDW10=0x02, 0x01, or 0x04) can return an error that is easy to miss without pre-flight capability checking. A pre-flight SANICAP check before dispatching the first cycle would prevent issuing unsupported actions entirely.

AAD-50 documents the recommended SANICAP check as:

sudo nvme id-ctrl /dev/nvme0 | grep sanicap

The wait and repeat options together implement exactly the verification architecture RFC #3415 proposed — per-cycle confirmed completion via Log Page 0x81 with multi-cycle redundancy. Having both natively in nvme-cli means operators no longer need a separate tool for verified multi-cycle sanitization. That is a significant step forward for the ecosystem.

Hi, Just pushed the SANICAP verifiation changes also and some fixes for the repaat repeat option. Thank you.

Hi, Just pushed the SANICAP verifiation changes also and some fixes for the repaat repeat option. Thank you.

This means the tool now confirms capability before dispatching any cycle, preventing silent failures on drives that report supporting an action they do not actually implement.

PR #3438 now has the complete verification stack. Per-cycle SSTAT confirmation via --wait, multi-cycle redundancy via --repeat N, and SANICAP pre-flight capability check. That is everything RFC #3415 was asking for.
Looking forward to seeing it merged.