freeipmi-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Freeipmi-devel] Interfacing IPMB using SSIF


From: Al Chu
Subject: Re: [Freeipmi-devel] Interfacing IPMB using SSIF
Date: Fri, 17 Jul 2009 13:24:17 -0700

Hey Fillod,

I haven't personally implemented one (obviously, since it's not in
FreeIPMI :-), but for some of your questions below ...

On Fri, 2009-07-17 at 19:27 +0200, Fillod Stephane wrote:
> Hi,
> 
> Does anyone have implemented an IPMB controller (BMC) with a Linux I2C
> stack?
> Rem: please let me know when I don't use the right terminology. 
> I'm quite new to this whole IPMI world :-)
> 
> The basic idea would be to allow freeipmi based applications to do
> in-band directly to IPMB. I don't know whether it makes a big
> difference,
> I have uTCA as the backplane in mind.
> 
> In order to implement a function ipmi_ssif_cmd_api_ipmb(), we would
> need the OS (Linux in this case) to be able to handle an I2C slave write
> IO, 
> in a multi-master fashion. IOW, when replying, some device on the IPMB
> does 
> an I2C write IO to the Linux node (which is not to be mistaken with an
> I2C 
> read), and the Linux node has to handle that IO. Maybe it could store it
> 
> until freeipmi retrieves it.
> So, on top of I2C_RDWR ioctl, we would need a kind of I2C_SLAVE_RECV.
> Does someone know whether such a thing exists? How best to define
> and code the ioctl interface if no such ioctl exists?

As you can see in the code, most of the time there is a generic "system
error" I return a SYSTEM_ERROR or INTERNAL_ERROR as needed.  If the
ioctl device just doesn't exist, we can just return an appropriate error
up the chain.

Al

> Maybe a question to the Linux I2C guys? http://*i2c.wiki.kernel.org
> 
> Thanks for your feedback
-- 
Albert Chu
address@hidden
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory





reply via email to

[Prev in Thread] Current Thread [Next in Thread]