[ofw] IB Atomic COMPARE_AND_SWAP returns IB_AE_WQ_REQ_ERR?

[Tang, Changqing]

How about remote READ and ATOMIC when you register memory ?
 
--CQ


________________________________

	From: ofw-bounces at lists.openfabrics.org
[mailto:ofw-bounces at lists.openfabrics.org] On Behalf Of Smith, Stan
	Sent: Thursday, October 11, 2007 10:04 AM
	To: Yossi Leybovich
	Cc: ofw at lists.openfabrics.org
	Subject: RE: [ofw] IB Atomic COMPARE_AND_SWAP returns
IB_AE_WQ_REQ_ERR?
	
	
________________________________

	From: Yossi Leybovich [mailto:sleybo at mellanox.co.il] 
	Sent: Thursday, October 11, 2007 3:11 AM
	To: Smith, Stan
	Cc: ofw at lists.openfabrics.org
	Subject: RE: [ofw] IB Atomic COMPARE_AND_SWAP returns
IB_AE_WQ_REQ_ERR?
	
	

	One simple and quick check

	Do you specify ATOMIC_ACCESS when you open the QP ? 

	 

	yes - IB_AC_ATOMIC | IB_AC_RDMA_READ | IB_AC_RDMA_WRITE |
IB_AC_LOCAL_WRITE are all set on the QP.

	 

	 

	 

	From: Smith, Stan [mailto:stan.smith at intel.com] 
	Sent: Wednesday, October 10, 2007 6:31 PM
	To: Yossi Leybovich
	Cc: ofw at lists.openfabrics.org
	Subject: RE: [ofw] IB Atomic COMPARE_AND_SWAP returns
IB_AE_WQ_REQ_ERR?

	 

	Additional HW info - Ven 15b3 Dev 25208 Hw a0 Fw 4.8.200 Drv
1.0.0000.614 (08/03/2006)

	 

________________________________

	From: ofw-bounces at lists.openfabrics.org
[mailto:ofw-bounces at lists.openfabrics.org] On Behalf Of Smith, Stan
	Sent: Wednesday, October 10, 2007 3:26 PM
	To: Yossi Leybovich; ofw at lists.openfabrics.org
	Subject: RE: [ofw] IB Atomic COMPARE_AND_SWAP returns
IB_AE_WQ_REQ_ERR?

	Hello,

________________________________

	From: Yossi Leybovich [mailto:sleybo at mellanox.co.il] 
	Sent: Wednesday, October 10, 2007 3:13 PM
	To: Smith, Stan; ofw at lists.openfabrics.org
	Subject: RE: [ofw] IB Atomic COMPARE_AND_SWAP returns
IB_AE_WQ_REQ_ERR?

	what HW do you use?  InfiniHost MT25208 - mthca.sys from
openib-windows svn.849, x64 on WinSvr 2003 (SP2); local process with RC
connection to another local process; should work?

	 

	if its mellanox HW there should be vendor syndrome in the EQE
entry , it should be in the same offset of the vendor syndrome of the
CQE.

	You can change the driver to report it - this will might help us
to identify the problem  

	 

	Thanks for the tip, will instrument the driver to display
information and get back to you.

	 

	stan.

	 

________________________________

	From: ofw-bounces at lists.openfabrics.org on behalf of Smith, Stan
	Sent: Wed 10/10/2007 17:51
	To: ofw at lists.openfabrics.org
	Subject: [ofw] IB Atomic COMPARE_AND_SWAP returns
IB_AE_WQ_REQ_ERR?

	Hello all,
	  I'm working on DAT v2 extensions which include an Atomic
	Compare-n-swap operation.
	The comp-n-swap operation generates an async QP error callback
with
	IB_AE_WR_REQ_ERR?
	The IB_AE_WR_REQ_ERR code is defined as 'an Opcode violation @
the
	responder'?
	
	Any clues as to what might generate an IB_AE_WR_REQ_ERR or what
an
	opcode violation @ responder implies?
	
	Anyone have an IBAL based compare-n-swap test program they would
like to
	share?
	
	Thanks,
	
	Stan.
	_______________________________________________
	ofw mailing list
	ofw at lists.openfabrics.org
	http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ofw

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/ofw/attachments/20071011/b171afe5/attachment.html>