Re: [PATCH v6 0/9] Fixes and rework for paes_s390 and phmac_s390

Re: [PATCH v6 0/9] Fixes and rework for paes_s390 and phmac_s390

[Vasily Gorbik]

Hi Herbert and David,

Harald's
"[PATCH v4 0/3] Fixes and rework for aes_s390" [1] and
"[PATCH v6 0/9] Fixes and rework for paes_s390 and phmac_s390" [2]

grew out of Heiko's
"[PATCH 0/3] s390/crypto: Enable CONTEXT_ANALYSIS" [3],
which was part of the broader s390 context analysis enablement work for
v7.3. Routing it through the s390 tree together with the rest of that
work was also briefly discussed [4]. Harald's series then grew with
additional fixes based on Sashiko findings.

The resulting series still only modify arch/s390/crypto, but several
patches now deal with interfaces and semantics maintained by the crypto
subsystem. Holger has reviewed both and valid findings reported by Sashiko
have been addressed. Could you please have a look and ACK taking both
series through the s390 tree if there are no concerns from the crypto
side? Or let me know if you would like to handle it another way.

Thank you.

Harald, could you please address/Cc all relevant maintainers and mailing
lists for such changes in future. The AES series has Cc: linux-crypto,
but the PAES/PHMAC series lost it along the way with v4. If routing
through a specific tree is preferred, please state this in the cover
letter.

[1] https://lore.kernel.org/r/20260814123711.61145-1-freude@linux.ibm.com
[2] https://lore.kernel.org/r/20260821084648.23222-1-freude@linux.ibm.com
[3] https://lore.kernel.org/r/20260804113736.1423667-1-hca@linux.ibm.com
[4] https://lore.kernel.org/r/an-56jLbR6bfPwvk@gondor.apana.org.au

On Fri, Aug 21, 2026 at 10:46:39AM +0200, Harald Freudenberger wrote:
> Fix and rework some issues around arch/s390/paes_s390.c and
> arch/s390/phmac_s390.c:
> 
> - Fix skcipher_walk return code handling in paes_s390
> - Add scrub of some temp buffers
> - Shift from using a mutex to using a semaphore in PAES CTR
>   Surprisingly clang code analysis is able to deal with semaphores and
>   thus the shift also fixes the issue with CONTEXT_ANALYSIS enabled.

This is not accurate. The semaphore conversion fixes the atomic context
problem. But semaphores are not a supported context analysis primitive
and down_trylock()/up() have no acquire/release annotations. So a
warning free build does not mean that the analyzer validated it.

> - And some more fixes related to paes and phmac (see changelog).

...

> Harald Freudenberger (8):
>   s390/crypto: Fix return code handling at skcipher_walk_done in PAES
>     algorithms
>   s390/crypto: Fix missing scrub of temp buffers with PAES algorithm
>   s390/crypto: Fix use of mutex in atomic context in PAES
>   s390/crypto: Fix missing cra_flags in paes_s390
>   s390/crypto: Fix handling of EBUSY in PAES when req is pushed to
>     crypto engine
>   s390/crypto: Fix handling of EBUSY in PHMAC when req is pushed to
>     crypto engine
>   s390/crypto: Fix wrong return code to engine in asynch callbacks
>   s390/crypto: Map EBUSY to EIO when key conversion fails repeatedly
> 
> Heiko Carstens (1):
>   s390/crypto: Enable CONTEXT_ANALYSIS
> 
>  arch/s390/crypto/Makefile     |   2 +
>  arch/s390/crypto/paes_s390.c  | 100 +++++++++++++++++++++++-----------
>  arch/s390/crypto/phmac_s390.c |  36 ++++++++----
>  3 files changed, 97 insertions(+), 41 deletions(-)