Re: [uml-user] Kernel config CRYPTO_AES vs CRYPTO_AES_586.

[Blaisorblade <bla...@ya...>]

On Friday 08 July 2005 14:28, Oleg Girko wrote:
> Hello!

> New configuration option CRYPTO_AES_586 first appeared in 2.6.12 kernel.
> This option can be activated when architecture is 32-bit X86 and it results
> in building X86-opitmised code for AES encryption. On the other hand, old
> CRYPTO_AES option is disabled when architectire is 32-bit X86. The code for
> X86-optimised AES module is located in "arch/i386/crypto" subdirectory and,
> hence, is built only if the value of the ARCH variable is equal to "i386".
No, UML build descend in that folder too, if needed. So your reasoning is 
wrong.
> But for unknown reason, CRYPTO_AES_586 option is enabled also if building
> UML for X86 architecture, and CRYPTO_AES option is unavailable in this
> case.

> There is no way to enable CRYPTO_AES option, but CRYPTO_AES_586 
> option does nothing,
How did you verify this?

Hmm, wait a moment. I understand the patch, but the current behaviour was done 
in order to allow using CRYPTO_AES_586 with UML, and it worked when I checked 
last time; and indeed, it was compiled and linked in (I checked this only).

> because there is no module for CRYPTO_AES_586 in 
> "arch/um" subdirectory, only in "arch/i386", which is irrelevant for
> ARCH=um.
No, it should be installed anyway. The module list is taken from the content 
of .tmp_versions.

However, I'm going to test this too.
-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade

		
___________________________________ 
Yahoo! Messenger: chiamate gratuite in tutto il mondo 
http://it.beta.messenger.yahoo.com