Re: [Freeipmi-users] IBM eServer 325 workaround

[Albert Chu]

Hey AB,

I ended up implementing a "--force-permsg-auth" option for this isue.

The problem with implementing what you suggest is what if IBM (I've sent
e-mails to some engineering folks there) fixes the problem in a later
firmware release?  Or perhaps we don't know what range of products from
IBM have this problem.  It may only be some of them.

Al

--
Albert Chu
address@hidden
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory


----- Original Message -----
From: Anand Babu <address@hidden>
Date: Tuesday, December 13, 2005 5:05 pm
Subject: Re: [Freeipmi-users] IBM eServer 325 workaround

> ,----[ Albert Chu <address@hidden> ]
> | Hi Gerald,
> | >
> | Carson Gaspar, a fellow I found through the ipmitool mailing list,
> | was able to determine what the problem was with the IBM eServer325.
> | I was able to test a patched version of 'ipmipower' and it worked.
> | >
> | Short story. there is a flag called "per msg authentication" in the
> | beginning of IPMI protocol.  The flag, to some degree, directs how
> | later IPMI packets should be generated.  The IBM eServer sets the
> | flag but later in the protocol assumes it isn't set.
> | >
> | I'm not entirely sure of the best way to deal with this right now.
> | This is the first vendor workaround that I'll have to do the
> | opposite of what the protocol tells me to do. I'm thinking of a
> | "--ibmeserver325" option.
> `----
> Instead of --ibmeserver325 option, you should use the manf-id/prod-id
> from "get_device_info" command  and automatically handle it from
> inside the libfreeipmi.
> 
> -- 
> Anand Babu 
> GPG Key ID: 0x62E15A31
> Blog [http://ab.freeshell.org]              
> The GNU Operating System [http://www.gnu.org]  
>