Hi Guy,
Sorry, I confused Storage NetFn with S/E, so the implementation of Ethereal
is right!
Thanks for your detailed explanation.
Regards,
Joseph
"Guy Harris"
<guy@xxxxxxxxxxxx To: "Joseph Hsiao" <Joseph.Hsiao@xxxxxxx>
> cc: ethereal-users@xxxxxxxxxxxx
Subject: Re: [Ethereal-users] Bug reporting about IPMI protocol
12/13/2003 07:38
AM
On Dec 12, 2003, at 1:15 AM, Joseph Hsiao wrote:
> I found a bug in Ethereal IPMI protocol that some IPMI command names
> do not
> match those in IPMI specification:
>
> NetFn CMD CMD name in spec. CMD name in
> Ethereal
> ==================================================================
> S/E 20h Get Device SDR Info Get SDR
> Repository Info
> S/E 23h Get Sensor Reading Factors Get SDR
Actually, that bug doesn't appear to be in the current CVS version.
Command codes are decoded by "get_netfn_cmd_text()", which takes a
NetFn value and a CMD value as arguments.
If the NetFn value is 0x04 or 0x05, the CMD value is decoded with the
"ipmi_se_cmd_vals[]" table, which has "Get Device SDR Info" for 0x20
and "Get Sensor Reading Factors" for 0x23.
Only if the NetFn value is 0x0a or 0x0b does it use the
"ipmi_storage_cmd_vals[]" table, which has "Get SDR Repository Info"
for 0x20 and "Get SDR" for 0x23, as it should, at least according to
the 1.5 rev 1.1 IPMI spec.
If you have a capture that appears to show the wrong command name,
check that the NetFn value is 0x04 or 0x05; if it is, send it to us, so
we can figure out why the wrong table is being used.
