On Mon, Jun 03, 2019 at 11:38:40AM +0200, Johannes Thumshirn wrote: > On Mon, May 27, 2019 at 07:19:54PM +0200, David Sterba wrote: > > 1-5 are reviewed and ok, 6 and 13 should be reworked, 7-12 is ok. I > > can't put the branch to next yet due to the csum formatting "issues" but > > will do once you resend. Should be ok just 6 and 13 as they're > > independent. > > I'd still like to hold back 13/13. SHA-256 doesn't seem to be well received by > the community as the "slow" hash and using a plain SHA-256 is not sufficient > for the dm-verity/fs-verity like approach I intend to implement in subsequent > patches. > > For the record, the current idea is to use a HMAC(SHA-256) as checksum > algorithm with a key provided at mkfs and mount time. The patch actually adding the new hash won't be merged to any to-be-released branch until we have the final list, but for testing purposes the patch will be in for-next and available via linux-next.
