Re: [Freeipmi-devel] assemble_ipmi_kcs_pkt

[Albert Chu]

Cool, I've checked in my changes for this, and also for
assemble_rmcp_hdr.  Updated error checks in the appropriate places. 
Although, many places to do not have an error check on
assemble_ipmi_kcs_pkt.  We may want to fix that later.

A;

--
Albert Chu
address@hidden
Lawrence Livermore National Laboratory

----- Original Message -----
From: Anand Babu <address@hidden>
Date: Saturday, March 13, 2004 3:15 pm
Subject: Re: [Freeipmi-devel] assemble_ipmi_kcs_pkt

> Hi Al,
> I always had macros to calculate the length. And I mostly combining
> error code and result in return value. 
> 
> But you are correct, 
> - returning length makes more sense in assemble-X function.
> - uniformity among lower level driver APIs.
> - length cannot be negative. So both len and error-code can be
>   combined.
> 
> Happy Hacking,
> --ab
> 
> ,----[ Albert Chu <address@hidden> ]
> | Hey AB,
> | 
> | Is there a reason why this function should return 0 instead of the
> | length of data copied into the buffer??  To me, assemble_X functions
> | should hide the details of how packets are assembled.  This is 
> what we
> | do with assemble_ipmi_lan.  Let the function handle the packet
> | assembly, then inform the user of how much data they should send out
> | through ipmi_lan_send or whatever send function they wish.
> | 
> | Al
> `----
> 
> -- 
> _.|_ 
> (_||_)
> Free as in Freedom <www.gnu.org>
>