Hi,
On Mon, Jan 21, 2008 at 10:17:33AM -0600, Corey Minyard wrote:
> I'm not really sure what is wrong, the code looks ok to me.
>
> Can you turn on message tracing with DEBUG_MSG_ENABLE(); before you open
> the connection, then send me the output (assuming you can't figure it out
> from that :).
Did that some time ago and figured out that the last input
message never generated an event. Seems like the connection was
prematurely closed. Or something. Anyway, changing the loop
at the end of the setup_ipmi_con from
while (1) {
rv = sel_select(os_sel, NULL, 0, NULL, NULL);
if (gstatus != IPMI_RUNNING) {
break;
}
}
to
while (!op_done) {
rv = os_hnd->perform_one_op(os_hnd, NULL);
if (rv) {
break;
}
}
made it behave. The global op_done is set to 1 in the
rsp_handler. Not sure if changing from sel_select to
os_hnd->perform_one_op made a difference. Or was it having an
extra global variable to handle the exit condition (perhaps
gstatus got changed before rsp_handler).
Many thanks for looking into this.
Cheers,
Dejan
>
> -corey
>
> Dejan Muhamedagic wrote:
>> Hi,
>>
>> On Thu, Jan 17, 2008 at 05:23:15PM -0600, Corey Minyard wrote:
>>
>>> I know why you are getting the segfault, which I will fix, but I do not
>>> know why the initialization is failing and causing this.
>>>
>>> Can you step through ipmi_init() with the debugger and see which exact
>>> function is failing to initialize?
>>>
>>
>> Sorry for the delay. Did some changes in the meantime and now I
>> can't reproduce the coredumps anymore. I think that they were
>> caused by calling ipmi_init twice. If you want to pursue that
>> issue, I can retrieve an old version and try to reproduce the
>> segfault.
>>
>> But I still can't make the thing work. This is the output from
>> ipmilan:
>>
>> INFO: ipmi_lan.c(connection_up): Connection 0 to the BMC is up
>> INFO: ipmi_lan.c(connection_up): Connection to the BMC restored
>> 0f 07 00 01 ff ** INFO: IPMI request 4 failed: ff
>>
>> The openipmicmd with same parameters returns good status. What's
>> funny is that the stuff on the wire seems to be very similar in
>> both cases---I can see the very same response as with ipmicmd
>> when running ipmilan.
>>
>> Looked for a long time at the source of ipmicmd and ipmilan, but
>> can't find any problems. Walked through both with gdb, still
>> didn't get any wiser. If you could take a look at the code and
>> tell me what's wrong with it, I'd appreciate very much. It has
>> been done around openipmi release 1.0 by a guy from Intel, but
>> he's not maintaining it any more.
>>
>> Cheers,
>>
>> Dejan
>>
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Openipmi-developer mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/openipmi-developer
