Re: [RFC PATCH] RDMA/iwcm: allow aborting an active connect awaiting CONNECT_REPLY

[Bernard Metzler <bernard.metzler@linux.dev>]

On 01.09.2026 18:50, Stefan Metzmacher wrote:
> Hi Leon,
> 
>> On Wed, Aug 05, 2026 at 02:02:00AM +0200, Yunseong Kim wrote:
>>> After a successful connect downcall, iw_cm_connect() returns with
>>> IWCM_F_CONNECT_WAIT still set and the cm_id in IW_CM_STATE_CONN_SENT.
>>> The only thing that clears the flag is the provider delivering
>>> IW_CM_EVENT_CONNECT_REPLY (cm_conn_rep_handler()).  Until that event
>>> arrives, iw_cm_disconnect() and destroy_cm_id() sleep uninterruptibly
>>> in
>>>
>>>     wait_event(cm_id_priv->connect_wait,
>>>            !test_bit(IWCM_F_CONNECT_WAIT, &cm_id_priv->flags));
>>>
>>> and both state machines treat IW_CM_STATE_CONN_SENT as BUG(), so the
>>> API has no way to cancel a pending active connect.  If the provider
>>> never generates the reply, because the peer died in the middle of
>>> connection setup or because of a provider bug, every teardown path
>>> (rdma_disconnect(), rdma_destroy_id()) blocks in D state forever and
>>> the ULP cannot recover: there is no way to disconnect after
>>> rdma_connect() was called without risking an unbounded hang.
>>>
>>> This class of problem is not theoretical.  The pending smbdirect
>>> change "smb: smbdirect: bound the disconnect wait in destroy_sync" [1]
>>> had to work around it on the ULP side:
>>> smbdirect_socket_destroy_sync() waited unbounded for the socket to
>>> reach SMBDIRECT_SOCKET_DISCONNECTED, a transition that depends on an
>>> asynchronous RDMA CM disconnect event, and when the peer died abruptly
>>> (a killed client, or Soft-RoCE/RXE where no graceful disconnect
>>> completes) that event never arrived.  The destroy ran on the single
>>> ksmbd-conn-release workqueue, every later connection release queued
>>> behind it in D state, and the whole server wedged until hung_task
>>> fired.  That change bounded the wait and drove the socket state
>>> machine to DISCONNECTED locally on timeout; ULPs should not have to
>>> resort to that, the CM should offer a teardown they can rely on.
>>>
>>> IWCM_F_CONNECT_WAIT currently guards two different windows:
>>>
>>>   * the connect/accept downcall into the provider being in progress;
>>>     teardown must keep waiting for that, it is short and bounded;
>>>
>>>   * an issued active connect waiting for CONNECT_REPLY, which is
>>>     potentially unbounded.
>>>
>>> Mark the second window with a new flag, IWCM_F_CONNECT_SENT: set by
>>> iw_cm_connect() once the downcall has returned successfully, cleared
>>> by cm_conn_rep_handler().  The teardown waits now complete when either
>>> the downcall has finished (!IWCM_F_CONNECT_WAIT, as before) or the
>>> pending-reply window has been entered (IWCM_F_CONNECT_SENT), and the
>>> previously BUG() CONN_SENT cases become:
>>>
>>>   * iw_cm_disconnect(): return -ENOTCONN; there is no established
>>>     connection to disconnect, aborting is the destroy path's job;
>>>
>>>   * destroy_cm_id(): abort the pending connect locally by moving to
>>>     DESTROYING and putting the QP into error so the provider tears the
>>>     connection attempt down.
>>>
>>> A CONNECT_REPLY that arrives after the abort is dropped: either
>>> cm_work_handler() sees IWCM_F_DROP_EVENTS, or cm_conn_rep_handler()
>>> now recognizes IW_CM_STATE_DESTROYING (the abort and the reply
>>> serialize on cm_id_priv->lock) and frees the event without touching
>>> the QP that destroy_cm_id() already released.  The cm_id memory stays
>>> valid for such a late reply because the provider holds its own
>>> reference (cm_id->add_ref) for as long as it can deliver events.
>>>
>>> The passive side has a sibling gap, where after a successful accept
>>> downcall the flag stays set until the provider's ESTABLISHED event
>>> arrives, which this patch deliberately does not change.
>>>
>>> [1] https://github.com/smfrench/smb3-kernel/commit/26d0f82a02c8a9c9c8cdfc138acb7ed0bf8e01a9
>>>
>>> Suggested-by: Stefan Metzmacher <metze@samba.org>
>>> Signed-off-by: Yunseong Kim <yunseong.kim@est.tech>
>>> ---
>>>   drivers/infiniband/core/iwcm.c | 76 +++++++++++++++++++++++++++++-----
>>>   drivers/infiniband/core/iwcm.h |  1 +
>>>   2 files changed, 67 insertions(+), 10 deletions(-)
>>
>> I'm not sure what to do with this patch, as the iWARP folks have
>> remained silent.
>>
>> Stefan,
>>
>> Do you still need this patch? Have you tested it?
> 
> I'll test it soon.
> 
> But the problem is real and I hit it very often in the past and the only option
> was a reboot.
> 
> I added Bernard explicitly...
> 
> metze
> 
I follow Yunseong's and Stefan's argumentation.

Thanks,
Bernard.