Re: Weird Kconfig condition (IPV6 [=n] || IPV6 [=n]=n)

From: Andreas Ziegler
Date: Mon Sep 14 2015 - 06:57:28 EST


Hi,

On 09/14/2015 10:41, Mason wrote:
> Hello,
>
> I did make menuconfig in v4.2
>
> In the help for CONFIG_CRYPTO_HMAC, I have
>
> â Selected by: IP_SCTP [=n] && NET [=y] && INET [=y] && (IPV6 [=n] || IPV6 [=n]=n) || INET_AH [=n] && NET [=y] && INET [=y] || \ â
> â INET_ESP [=n] && NET [=y] && INET [=y] || INET6_AH [=n] && NET [=y] && INET [=y] && IPV6 [=n] || INET6_ESP [=n] && NET [=y] && INET [=y]\ â
> â && IPV6 [=n] || SCTP_COOKIE_HMAC_MD5 [=n] && NET [=y] && IP_SCTP [=n] && SCTP_COOKIE_HMAC_MD5 [=n] || SCTP_COOKIE_HMAC_SHA1 [=n] && \ â
> â NET [=y] && IP_SCTP [=n] && SCTP_COOKIE_HMAC_SHA1 [=n] || CIFS [=n] && NETWORK_FILESYSTEMS [=y] && INET [=y] || TRUSTED_KEYS [=n] && \ â
> â KEYS [=y] && TCG_TPM [=n] || ENCRYPTED_KEYS [=n] && KEYS [=y] || IMA [=n] && INTEGRITY [=n] || EVM [=n] && INTEGRITY [=n] || \ â
> â CRYPTO_DRBG_HMAC [=y] && CRYPTO [=y] && CRYPTO_DRBG_MENU [=m] || CRYPTO_DEV_OMAP_SHAM [=n] && CRYPTO [=y] && CRYPTO_HW [=y] && \ â
> â ARCH_OMAP2PLUS [=n] || CRYPTO_DEV_QAT [=n] && CRYPTO [=y] && CRYPTO_HW [=y] â
>
>
> What does (IPV6 [=n] || IPV6 [=n]=n) mean?
> Is this a bug in my configuration?

No, this comes from the definition of IP_SCTP in net/sctp/Kconfig.
menuconfig IP_SCTP
tristate "The SCTP Protocol"
depends on INET
depends on IPV6 || IPV6=n
^^^^^^^^^^^^^^

This weird-looking condition has the effect that:
- if IPV6 is disabled, IP_SCTP can be anything ("y","m" or "n")
- if IPV6 is enabled to "y", IP_SCTP can also be set to anything.
- (and this is the reason why it's there:) if IPV6 is a module ("m"),
IP_SCTP can only be "m" or "n".

In a way, the "depends on" does not only reference the symbol, but also
keeps track of the 'state' of the symbol it references, such that if the
dependency (here: IPV6) is to be built as a loadable module only, the
dependant feature can also only be built as a loadable module.

I can't find the Documentation for that at the moment, but I'm sure that
I read this at some point, maybe also only on a mailing list...

Regards,

Andreas

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/