Re: [PATCH v2 12/13] s390/kexec: refactor for kernel/Kconfig.kexec

On Wed, Jun 21, 2023 at 12:10:49PM -0500, Eric DeVolder wrote:
Hi Eric,
...

quoted

quoted

NOTE: The original Kconfig has a KEXEC_SIG which depends on
MODULE_SIG_FORMAT. However, attempts to keep the MODULE_SIG_FORMAT
dependency (using the strategy outlined in this series, and other
techniques) results in 'error: recursive dependency detected'
on CRYPTO. This occurs due to any path through KEXEC_SIG
attempting to select CRYPTO is ultimately dependent upon CRYPTO:

  CRYPTO
   <- ARCH_SUPPORTS_KEXEC_FILE
      <- KEXEC_FILE
         <- KEXEC_SIG

Therefore, the solution is to drop the MODULE_SIG_FORMAT dependency
for KEXEC_SIG. In practice, however, MODULE_SIG_FORMAT is still
configured-in as the use of KEXEC_SIG is in step with the use of
SYSTEM_DATA_VERIFICATION, which does select MODULE_SIG_FORMAT.

No, it is actually the other way around.
Could you please provide the correct explanation?

AFAICT the MODULE_SIG_FORMAT dependency was introduced with commit
c8424e776b09 ("MODSIGN: Export module signature definitions") and
in fact was not necessary, since s390 did/does not use mod_check_sig()
anyway. So the SYSTEM_DATA_VERIFICATION could have left intact.

Thomas, would the correct explanation be simply indicating that
MODULE_SIG_FORMAT isn't needed as it is not used by s390 (crediting your
summary above)?

I guess, you asked me? Anyway, I will try to answer as if I were Thomas :)

MODULE_SIG_FORMAT is needed to select SYSTEM_DATA_VERIFICATION.
But SYSTEM_DATA_VERIFICATION is also selected by FS_VERITY*, so
dropping MODULE_SIG_FORMAT does not hurt.

Thanks!