[ofa-general] smpquery vlarb ... dumps only 1 entry?

[Sasha Khapyorsky]

On 10:27 Sat 29 Mar     , Sasha Khapyorsky wrote:
> Hi,
> 
> On 14:44 Thu 27 Mar     , akepner at sgi.com wrote:
> > 
> > In ofed 1.2, "smpquery vlarb ... " would produce output like:
> > 
> > # VLArbitration tables: Lid 8 port 0 LowCap 8 HighCap 8
> > # Low priority VL Arbitration Table:
> > VL    : |0x0 |0x1 |0x2 |0x3 |0x4 |0x5 |0x6 |0x7 |
> > WEIGHT: |0x0 |0x0 |0x0 |0x0 |0x0 |0x0 |0x0 |0x0 |
> > # High priority VL Arbitration Table:
> > VL    : |0x0 |0x0 |0x0 |0x0 |0x1 |0x2 |0x3 |0x4 |
> > WEIGHT: |0xFF|0xFF|0xFF|0xFF|0xFF|0x7F|0x3F|0x1F|
> > 
> > which is what I expected.
> > 
> > But in ofed 1.3, I only get a single entry:
> > 
> > # VLArbitration tables: Lid 1 port 0 LowCap 8 HighCap 8
> > # Low priority VL Arbitration Table:
> > VL    : |0x3 |
> > WEIGHT: |0x3 |
> > # High priority VL Arbitration Table:
> > VL    : |0x3 |
> > WEIGHT: |0x8 |
> > 
> > Anyone know what changed here? 
> 
> I'm not aware about any of such changes. And it looks wrong (I cannot
> see this in my setup).
> 
> mad_dump_vlarbitration() (from libibmad/src/dump.c) has this:
> 
> 	sprintf(buf_line1, "%s0x%-2X|", buf_line1, vl);
> 
> Any chance that you have different sprintf() implementations with OFED
> 1.2 and 1.3?

Also how output of

  smpdump 1 0x18 0x00010000
  smpdump 1 0x18 0x00030000

looks?

Sasha