* Double RESET -> INIT transitions
@ 2021-04-06 15:01 Haakon Bugge
2021-04-06 15:28 ` Jason Gunthorpe
0 siblings, 1 reply; 3+ messages in thread
From: Haakon Bugge @ 2021-04-06 15:01 UTC (permalink / raw)
To: Doug Ledford, Jason Gunthorpe, OFED mailing list
Cc: Leon Romanovsky, Mark Haywood
Running rdma_server (RC transport) on an IB device that supports abi_ver 4, I see two ib_modify_qp() calls to the INIT state with mask 0x39 (STATE, ACCESS_FLAGS, PKEY_INDEX, and PORT) and of course, state == 1 (IB_QPS_INIT).
Presumable, from ucma_init_conn_qp(), then again in ucma_modify_qp_rtr().
Slip of the keyboard?
Thxs, Håkon
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Double RESET -> INIT transitions
2021-04-06 15:01 Double RESET -> INIT transitions Haakon Bugge
@ 2021-04-06 15:28 ` Jason Gunthorpe
2021-04-06 15:45 ` Haakon Bugge
0 siblings, 1 reply; 3+ messages in thread
From: Jason Gunthorpe @ 2021-04-06 15:28 UTC (permalink / raw)
To: Haakon Bugge
Cc: Doug Ledford, OFED mailing list, Leon Romanovsky, Mark Haywood
On Tue, Apr 06, 2021 at 03:01:41PM +0000, Haakon Bugge wrote:
> Running rdma_server (RC transport) on an IB device that supports
> abi_ver 4, I see two ib_modify_qp() calls to the INIT state with
> mask 0x39 (STATE, ACCESS_FLAGS, PKEY_INDEX, and PORT) and of course,
> state == 1 (IB_QPS_INIT).
>
> Presumable, from ucma_init_conn_qp(), then again in ucma_modify_qp_rtr().
>
> Slip of the keyboard?
The control flow is complicated here and I worry parts of this are ABI
now. Would it be OK to return from rdma_create_qp_ex() with an
incompletely initialized QP? I don't know.
Jason
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Double RESET -> INIT transitions
2021-04-06 15:28 ` Jason Gunthorpe
@ 2021-04-06 15:45 ` Haakon Bugge
0 siblings, 0 replies; 3+ messages in thread
From: Haakon Bugge @ 2021-04-06 15:45 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: Doug Ledford, OFED mailing list, Leon Romanovsky, Mark Haywood
> On 6 Apr 2021, at 17:28, Jason Gunthorpe <jgg@nvidia.com> wrote:
>
> On Tue, Apr 06, 2021 at 03:01:41PM +0000, Haakon Bugge wrote:
>
>> Running rdma_server (RC transport) on an IB device that supports
>> abi_ver 4, I see two ib_modify_qp() calls to the INIT state with
>> mask 0x39 (STATE, ACCESS_FLAGS, PKEY_INDEX, and PORT) and of course,
>> state == 1 (IB_QPS_INIT).
>>
>> Presumable, from ucma_init_conn_qp(), then again in ucma_modify_qp_rtr().
>>
>> Slip of the keyboard?
>
> The control flow is complicated here and I worry parts of this are ABI
> now. Would it be OK to return from rdma_create_qp_ex() with an
> incompletely initialized QP? I don't know.
Two transitions to INIT with same mask and data is idempotent. Should be possible to remove the second for this case. Let me discuss internally with Mark and get back.
Thxs, Håkon
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2021-04-06 15:45 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-04-06 15:01 Double RESET -> INIT transitions Haakon Bugge
2021-04-06 15:28 ` Jason Gunthorpe
2021-04-06 15:45 ` Haakon Bugge
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).