RE: [PATCH V2 RFC 0/3] svcrdma: refactor marshalling logic

["Steve Wise" <swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>]

> -----Original Message-----
> From: linux-nfs-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org [mailto:linux-nfs-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org] On Behalf
> Of Devesh Sharma
> Sent: Monday, May 19, 2014 2:07 PM
> To: Steve Wise; J. Bruce Fields
> Cc: linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; tom-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org
> Subject: RE: [PATCH V2 RFC 0/3] svcrdma: refactor marshalling logic
> 
> While testing with ocrdma driver I am finding server side SQ full. Following is the log,
yet to
> identify why it's happening. Once this is reported Client side crashes due to some
reason.
> My kdump is not working properly therefore I am not able to analyze the situation
properly.
> 
> May 19 23:47:02 neo01-el64 kernel: svcrdma: RDMA_WRITE rmr=8008b12, to=45a2d790c,
> xdr_off=0, write_len=68, vec->sge=ffff88086cb4a0c8, vec->count=2
> May 19 23:47:02 neo01-el64 kernel: svcrdma: send_reply returns 0
> May 19 23:47:02 neo01-el64 kernel: svc: server ffff88086409a000 waiting for data (to =
> 3600000)
> May 19 23:47:02 neo01-el64 kernel: svc: transport ffff88087dfa2400 served by daemon
> ffff88086409a000
> May 19 23:47:02 neo01-el64 kernel: svc: server ffff88086409a000, pool 0, transport
> ffff88087dfa2400, inuse=18
> May 19 23:47:02 neo01-el64 kernel: svcrdma: rqstp=ffff88086409a000
> May 19 23:47:02 neo01-el64 kernel: svcrdma: processing ctxt=ffff880866754540 on
> xprt=ffff88087dfa2400, rqstp=ffff88086409a000, status=0
> May 19 23:47:02 neo01-el64 kernel: svcrdma: failed to post SQ WR rc=-22, sc_sq_count=0,
> sc_sq_depth=128
> May 19 23:47:02 neo01-el64 kernel: svcrdma: Error -22 posting RDMA_READ

Hey Deevesh,

Looking ocrdma_post_send(),-22 (-EINVAL) is returned when the QP is not in RTS.  If the SQ
is full, -ENOMEM is returned.  So I think the send error is a downstream error because the
connection got knocked down.  You should try and figure out what kicked the QP out of RTS.


Steve.

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html