On 8/24/23 11:05, Bart Van Assche wrote: > On 8/24/23 08:35, Bernard Metzler wrote: >> I spent some time testing the srp/002 blktest with siw, still >> trying to get it hanging. >> Looking closer into the logs: While most of the time RDMA CM >> connection setup works, I also see some connection rejects being >> created by the passive ULP side during setup: >> >> [16848.757937] scsi host11: ib_srp: REJ received >> [16848.757939] scsi host11: REJ reason 0xffffff98 >> >> This does not affect the overall success of the current test >> run, other connect attempts succeed etc. Is that connection >> rejection intended behavior of the test? > > Hi Bernard, > > In the logs I see that the SRP initiator (ib_srp) may try to log in before > the SRP target driver (ib_srpt) has finished associating with the configured > RDMA ports. I think this is why REJ messages appear in the logs. The retry > loop in the test script should be sufficient to deal with this. > > Bart. Thanks to both of you for taking the time to look at this. Bob
