On Wed, 26 May 2004, Ann Smith wrote:

> IBM configured our PCICA card to our VM LPAR that runs linux but we see
> the following errors:
> 600 May 26 08:03:24 hostl1 kernel: z90crypt: Hardware error
>     601 May 26 08:03:24 hostl1 kernel: z90crypt: Type 82 Message Header:
> 0082000010000000
>     602 May 26 08:03:24 hostl1 kernel: z90crypt: Device 14 returned
> hardware error 13
>     603 May 26 08:03:43 hostl1 kernel: z90crypt: Error in
> get_crypto_request_buffer: 16

Not sure if I ever saw exactly this error, but I can suggest to make sure
that you have the CRYPTO APVIRT statement specified in the Linux guest's
directory entry.  You do not need the CRYPTO parameter on the CPU
statement, this is just for MVS guests to give them access to the CCF (I
think).

You should be able to issue "#CP Q CRYPTO" (Q VIRT CRYPTO) in your Linux
guest and see the virtual queue to the PCICA.  The same command issued
by your suitably-privileged non-G user (a real Q CRYPTO) should show you a
little bit more info.

There are dependencies in LPAR activation and definition as well:

 * The crypto can only be accessed by one LPAR at a time.  You can define
a given crypto in the activation profile of more than one LPAR, but the
first LPAR activated will get that crypto (to the exclusion of all others
that may have it in their profile).

 * PCICAs and PCICCs have CP/IFL affinity, apparently: even if you get the
above right, if the crypto's affinity is to a CP/IFL that is not defined
or available to the LPAR, the LPAR will not be able to use the crypto[1].

Cheers,
Vic Cross


[1] I may be wrong on this -- this was a piece of advice we encountered
along our way while diagnosing problems with crypto on a z800, but we
later found that the PCICA in question failed diagnostics...

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to