[ofa-general] Re: problems using smpdump

[Or Gerlitz]

Sasha Khapyorsky <sashak at voltaire.com> wrote:

> It is a bug in smpdump (not just print error), it tries to set sl explicitly to 0xffff
>  when using LID routed MADs (so I suppose direct routed MAD will work
> ). And seems this bug was from day "0" of smpdump.

Yes, it works fine with direct route, thanks for the quick resolution!

> Guess that this should be a patch:

I don't have the spec in front of me, but looking on the LRH structure
in include/rdma/ib_pack.h the SL field spans eight bit, so I think
there may be another bug somewhere in the umad code which lets the
consumer set a value larger then 0xff.

Or.