FW: IPMI kernel module errors on 6.x

[John Baldwin]

On Tuesday 14 November 2006 16:09, Raymond T. Sundland wrote:
> Hi John,
> 
> I don't see any of the additional messages that should be output by 
> KCS_DEBUG.  I see the code (ifdef's) everywhere, but I don't actually 
> see the output.  is this something that should be in dmesg or 
> elsewhere?  I even removed the ifdef's to see if the output would be 
> displayed in the dmesg without the check, and I see nothing.  I also 
> tried loading the module with kldload -v.

The 'failed to start write' messages are under the #ifdef.  You'll need to add 
new printfs to dig into this further though, possibly dumping the status read 
inside the kcs_error() function, or what the status is when kcs_start_write() 
fails.

> John Baldwin wrote:
> > On Monday 13 November 2006 21:50, Raymond T. Sundland wrote:
> >   
> >> This is the output with KCS_DEBUG enabled:
> >>
> >>     ipmi0: <IPMI System Interface> on isa0
> >>     ipmi0: KCS mode found at mem 0xca2 alignment 0x4 on isa
> >>     ipmi0: KCS: initial state: 00
> >>     ipmi0: KCS: Failed to start write
> >>     ipmi0: KCS Error retry exhausted
> >>     ipmi0: KCS: Failed to start write
> >>     ipmi0: KCS Error retry exhausted
> >>     ipmi0: KCS: Failed to start write
> >>     ipmi0: KCS Error retry exhausted
> >>     ipmi0: Timed out waiting for GET_DEVICE_ID
> >>     
> >
> > Hmmm.  It appears to not be working at all.  You can add more printf's in 
> > kcs_start_write() and kcs_error() to see exactly which conditions are 
> > failing.  If the status byte was just always stuck as 0x00 kcs_error() 
should 
> > succeed, so it is doing something.
> >
> >   
> >> John Baldwin wrote:
> >>     
> >>> On Monday 13 November 2006 13:48, Raymond Sundland wrote:
> >>>   
> >>>       
> >>>> I sent this to -questions earlier, but thought I would get a more 
> >>>>         
> > complete
> >   
> >>>> response from either -hackers or -amd64.
> >>>>     
> >>>>         
> >>> Try enabling KCS_DEBUG in sys/dev/ipmi/ipmi_kcs.c and save the output to 
a 
> >>> file (it will be a lot!).
> >>>
> >>>   
> >>>       
> >
> >   
> 

-- 
John Baldwin