Re: [PATCH v6] lib/crypto: blake2s: include as built-in

[Jason A. Donenfeld]

On 1/3/22, Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> wrote:
> At this point I think we should push this through crypto.  The
> changes are too invasive with respect to the crypto Kconfig files.

Ugh, can we please not? That will really make things much harder and
more annoying for me. I have an early pull planned, and you'll quickly
be able to rebase on top of it. It also doesn't appear to conflict
with anything you have queued up. Please, I would really appreciate
some straight forward linearity here, and I don't think my taking it
will negatively impact the flow.

>
>> diff --git a/crypto/Kconfig b/crypto/Kconfig
>> index 285f82647d2b..b7a2e50dcbc8 100644
>> --- a/crypto/Kconfig
>> +++ b/crypto/Kconfig
>> @@ -702,7 +702,7 @@ config CRYPTO_BLAKE2S
>>  	  See https://blake2.net for further information.
>>
>>  config CRYPTO_BLAKE2S_X86
>> -	tristate "BLAKE2s digest algorithm (x86 accelerated version)"
>> +	bool "BLAKE2s digest algorithm (x86 accelerated version)"
>>  	depends on X86 && 64BIT
>>  	select CRYPTO_LIB_BLAKE2S_GENERIC
>>  	select CRYPTO_ARCH_HAVE_LIB_BLAKE2S
>
> This will break when CRYPTO is disabled because the x86 crypto
> glue code depends on the crypto subsystem.

That snippet is inside an 'if CRYPTO' block, so it can't be selected
without CRYPTO being enabled.