Re: [PATCH 4/10] crypto: hash - Count error stats differently

[Eric Biggers <ebiggers@kernel.org>]

On Wed, Feb 15, 2023 at 05:25:13PM +0800, Herbert Xu wrote:
> Move all stat code specific to hash into the hash code.
> 
> While we're at it, change the stats so that bytes and counts
> are always incremented even in case of error.  This allows the
> reference counting to be removed as we can now increment the
> counters prior to the operation.
> 
> After the operation we simply increase the error count if necessary.
> This is safe as errors can only occur synchronously (or rather,
> the existing code already ignored asynchronous errors which are
> only visible to the callback function).
> 
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
> ---
> 
>  crypto/ahash.c            |   82 +++++++++++++++++------------
>  crypto/algapi.c           |   24 --------
>  crypto/crypto_user_stat.c |   38 -------------
>  crypto/hash.h             |   37 +++++++++++++
>  crypto/shash.c            |  128 ++++++++++++++++++++++++++++++++++++----------
>  include/crypto/hash.h     |   85 +++++++++++++++++++++++-------
>  include/linux/crypto.h    |   20 -------
>  7 files changed, 250 insertions(+), 164 deletions(-)

This patch is causing the 'BUG_ON(refcount_read(&alg->cra_refcnt) != 1);' in
crypto_unregister_alg() to be hit at boot time.

Most likely related to the way the algorithm structs are being rearranged (which
really ought to be explained in the commit message).

- Eric