On Tue, 31 Dec 2019 at 04:21, Eric Biggers <ebiggers@xxxxxxxxxx> wrote: > > The CRYPTO_TFM_RES_* flags are pointless since they are never checked > anywhere. And it's not really possible for anyone to start using them > without a lot of work, since many drivers aren't setting them or are > setting them when they shouldn't. > > Also, if we ever actually need to start distinguishing ->setkey() errors > better (which is somewhat unlikely, as it's been a long time with no one > caring), we'd probably be much better off just using different return > values, like -EINVAL if the key is invalid for the algorithm vs. > -EKEYREJECTED if the key was rejected by a policy like "no weak keys". > That would be much simpler, less error-prone, and easier to test. > > So let's just remove these flags for now. This gets rid of a lot of > pointless boilerplate code. > > Patches 6 and 8 are a bit large since they touch so many drivers, though > the changes are straightforward and it would seem overkill to do this as > a series of 70 separate patches. But let me know if it's needed. > > Eric Biggers (8): > crypto: chelsio - fix writing tfm flags to wrong place > crypto: artpec6 - return correct error code for failed setkey() > crypto: atmel-sha - fix error handling when setting hmac key > crypto: remove unused tfm result flags > crypto: remove CRYPTO_TFM_RES_BAD_BLOCK_LEN > crypto: remove CRYPTO_TFM_RES_BAD_KEY_LEN > crypto: remove CRYPTO_TFM_RES_WEAK_KEY > crypto: remove propagation of CRYPTO_TFM_RES_* flags > ... > 108 files changed, 218 insertions(+), 917 deletions(-) > For the series: Acked-by: Ard Biesheuvel <ardb@xxxxxxxxxx>
