I believe the issue (not certain - need more test time) was caused by not
specifying all vlarb entries,
even those unused. So, QoS on, but incomplete specification. OFED HowTo
examples do this, so I figured it
was OK.
Its also possible that some software may have been specifying an un-configured
SL - without my permission ;)
Maybe vl 0 was loosing its turn in arbitration and got starved for > +3%/-51%
400,000 symbols when the HCA
was passing back flow control update?
Unfortunately - it usually worked, except under load.
Think its working now.
bob
btw Ira, you get to trade every ! into a beer the next time I see you.
!!! - there, an even 6 pack
On 05/21/2012 10:08 AM, Ira Weiny wrote:
On Fri, 18 May 2012 11:27:40 -0400
Hal Rosenstock<[email protected]> wrote:
Ira,
On 5/18/2012 11:10 AM, Ira Weiny wrote:
In our situation the FW in the switch had a bug.
Do you recall what the bug was, which firmware version had the bug and
which one fixed it ?
I don't recall but it was not Mellanox hardware.
Ira
-- Hal
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
