On Tue, 25 Feb 2025 07:44:30 -0800, Christoph Hellwig wrote:
> this is the tip of the iceberg of some of the PI work I've done a while
> ago, and given the current discussions it might be a good time to send it
> out.
>
> The idea is to:
>
> a) make the auto-PI code stand out more clearly as it seems to lead to
> a lot of confusion
> b) optimize the size of the integrity payload to prepare for usage in
> file systems
> c) make sure the mempool backing actually works for auto-PI. We'll still
> need a mempool for the actual metadata buffer, but that is left for the
> next series.
>
> [...]
Applied, thanks!
[1/3] block: mark bounce buffering as incompatible with integrity
commit: e9945facd48d2d5da87fa247f5d6a19c23d935fd
[2/3] block: move the block layer auto-integrity code into a new file
commit: d2cfe5ceca59b74ef96bfe00632e3927d00c9918
[3/3] block: split struct bio_integrity_payload
commit: 1972a1faaa026eb34322a2b4fcbb3c28d68cc5e2
Best regards,
--
Jens Axboe