On 12/25/17 05:02, Moni Shoua wrote:
1. I will do my best to add more tests to RXE regression. However, it
may take a while.
2. Differences in behavior doesn't necessarily mean that at least one
implementation is wrong. In what you describe it is hard to understand
what you think is wrong with RXE, If I understand it right the script
tried to delete a directory that ib_srpt owns (configs or such?) and
this operation waits for a completion. If this is right do you know
who is expected to call complete()? It sound unlikely that rxe is the
one.
3. Despite that, let's try this: when script hangs, can you run echo t
> /proc/sysrq-trigger and see if you something in dmesg that can
explain the hang? Maybe a trace that rdma_rxe is a part of it?
Hello Moni,
The ib_srpt driver uses zero-length writes to trigger the completion
handler if either an RTU event is received or an RDMA channel is being
closed. In the log I saw the message "queued zerolength write" appear
but not "srpt_zerolength_write_done: wc->status = ..." when the hang was
observed. That made me wonder whether the rxe driver perhaps suppresses
completions for zero-length writes if the queue pair state is changed
into IB_QPS_ERR? I think it is required by the IB spec to queue an error
completion for pending work requests upon the transition to IB_QPS_ERR.
Thanks,
Bart.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
