On Wed, Mar 15, 2023 at 12:55:31PM -0600, Jens Axboe wrote: > On 3/15/23 12:54 PM, Eric Biggers wrote: > > On Wed, Mar 15, 2023 at 12:50:45PM -0600, Jens Axboe wrote: > >> On 3/15/23 12:39 PM, Eric Biggers wrote: > >>> From: Eric Biggers <ebiggers@xxxxxxxxxx> > >>> > >>> To avoid hiding information, pass on the error code from > >>> blk_crypto_rq_get_keyslot() instead of always using BLK_STS_IOERR. > >> > >> Maybe just fold this with the previous patch? > > > > I'd prefer to keep the behavior change separate from the cleanup. > > OK fair enough, not a big deal to me. Series looks fine as far as > I'm concerned. Not loving the extra additions in the completion path, > but I suppose there's no way around that. Well, my first patch didn't have that: https://lore.kernel.org/linux-block/20230226203816.207449-1-ebiggers@xxxxxxxxxx But it didn't fix the keyslot refcounting work as expected, which Nathan didn't like (and I agree with that). - Eric
![]() |