From: Lee Duncan <lduncan@suse.com>
To: Mike Christie <michael.christie@oracle.com>,
martin.petersen@oracle.com, linux-scsi@vger.kernel.org,
mrangankar@marvell.com, njavali@marvell.com,
GR-QLogic-Storage-Upstream@marvell.com, cleech@redhat.com,
liuzhengyuang521@gmail.com
Subject: Re: [PATCH 4/6] scsi: iscsi, ql4: Use per session workqueue for unbinding.
Date: Mon, 28 Feb 2022 10:19:05 -0800 [thread overview]
Message-ID: <fa265ba8-b0d4-0d02-ac32-27abab0e9f3d@suse.com> (raw)
In-Reply-To: <20220226230435.38733-5-michael.christie@oracle.com>
On 2/26/22 15:04, Mike Christie wrote:
> We currently allocate a workqueue per host and only use it for removing
> the target. For the session per host case we could be using this workqueue
> to be able to do recoveries (block, unblock, timeout handling) in
> parallel. To also allow offload drivers to do their session recoveries in
> parallel, this drops the per host workqueue and replaces it with a per
> session one.
>
> Signed-off-by: Mike Christie <michael.christie@oracle.com>
> ---
> drivers/scsi/qla4xxx/ql4_os.c | 2 +-
> drivers/scsi/scsi_transport_iscsi.c | 19 ++++++++++++++-----
> include/scsi/scsi_transport_iscsi.h | 2 ++
> 3 files changed, 17 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/scsi/qla4xxx/ql4_os.c b/drivers/scsi/qla4xxx/ql4_os.c
> index 0ae936d839f1..955d8cb675f1 100644
> --- a/drivers/scsi/qla4xxx/ql4_os.c
> +++ b/drivers/scsi/qla4xxx/ql4_os.c
> @@ -5096,7 +5096,7 @@ int qla4xxx_unblock_flash_ddb(struct iscsi_cls_session *cls_session)
> ql4_printk(KERN_INFO, ha, "scsi%ld: %s: ddb[%d]"
> " start scan\n", ha->host_no, __func__,
> ddb_entry->fw_ddb_index);
> - scsi_queue_work(ha->host, &ddb_entry->sess->scan_work);
> + queue_work(ddb_entry->sess->workq, &ddb_entry->sess->scan_work);
> }
> return QLA_SUCCESS;
> }
> diff --git a/drivers/scsi/scsi_transport_iscsi.c b/drivers/scsi/scsi_transport_iscsi.c
> index 05cd4bca979e..ecb592a70e03 100644
> --- a/drivers/scsi/scsi_transport_iscsi.c
> +++ b/drivers/scsi/scsi_transport_iscsi.c
> @@ -2032,19 +2032,27 @@ EXPORT_SYMBOL_GPL(iscsi_alloc_session);
>
> int iscsi_add_session(struct iscsi_cls_session *session, unsigned int target_id)
> {
> + struct Scsi_Host *shost = iscsi_session_to_shost(session);
> unsigned long flags;
> int id = 0;
> int err;
>
> session->sid = atomic_add_return(1, &iscsi_session_nr);
>
> + session->workq = alloc_workqueue("iscsi_ctrl_%d:%d",
> + WQ_SYSFS | WQ_MEM_RECLAIM | WQ_UNBOUND, 0,
> + shost->host_no, session->sid);
> + if (!session->workq)
> + return -ENOMEM;
> +
> if (target_id == ISCSI_MAX_TARGET) {
> id = ida_simple_get(&iscsi_sess_ida, 0, 0, GFP_KERNEL);
>
> if (id < 0) {
> iscsi_cls_session_printk(KERN_ERR, session,
> "Failure in Target ID Allocation\n");
> - return id;
> + err = id;
> + goto destroy_wq;
> }
> session->target_id = (unsigned int)id;
> session->ida_used = true;
> @@ -2078,7 +2086,8 @@ int iscsi_add_session(struct iscsi_cls_session *session, unsigned int target_id)
> release_ida:
> if (session->ida_used)
> ida_simple_remove(&iscsi_sess_ida, session->target_id);
> -
> +destroy_wq:
> + destroy_workqueue(session->workq);
> return err;
> }
> EXPORT_SYMBOL_GPL(iscsi_add_session);
> @@ -2177,6 +2186,8 @@ void iscsi_remove_session(struct iscsi_cls_session *session)
>
> transport_unregister_device(&session->dev);
>
> + destroy_workqueue(session->workq);
> +
> ISCSI_DBG_TRANS_SESSION(session, "Completing session removal\n");
> device_del(&session->dev);
> }
> @@ -3833,8 +3844,7 @@ iscsi_if_recv_msg(struct sk_buff *skb, struct nlmsghdr *nlh, uint32_t *group)
> case ISCSI_UEVENT_UNBIND_SESSION:
> session = iscsi_session_lookup(ev->u.d_session.sid);
> if (session)
> - scsi_queue_work(iscsi_session_to_shost(session),
> - &session->unbind_work);
> + queue_work(session->workq, &session->unbind_work);
> else
> err = -EINVAL;
> break;
> @@ -4707,7 +4717,6 @@ iscsi_register_transport(struct iscsi_transport *tt)
> INIT_LIST_HEAD(&priv->list);
> priv->iscsi_transport = tt;
> priv->t.user_scan = iscsi_user_scan;
> - priv->t.create_work_queue = 1;
>
> priv->dev.class = &iscsi_transport_class;
> dev_set_name(&priv->dev, "%s", tt->name);
> diff --git a/include/scsi/scsi_transport_iscsi.h b/include/scsi/scsi_transport_iscsi.h
> index 90b55db46d7c..7a0d24d3b916 100644
> --- a/include/scsi/scsi_transport_iscsi.h
> +++ b/include/scsi/scsi_transport_iscsi.h
> @@ -251,6 +251,8 @@ struct iscsi_cls_session {
> bool recovery_tmo_sysfs_override;
> struct delayed_work recovery_work;
>
> + struct workqueue_struct *workq;
> +
> unsigned int target_id;
> bool ida_used;
>
Reviewed-by: Lee Duncan <lduncan@suse.com>
next prev parent reply other threads:[~2022-02-28 18:34 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-02-26 23:04 [PATCH 0/6] iscsi: Speed up failover with lots of devices Mike Christie
2022-02-26 23:04 ` [PATCH 1/6] scsi: iscsi: Fix recovery and ublocking race Mike Christie
2022-02-27 19:49 ` Lee Duncan
2022-02-28 20:06 ` Chris Leech
2022-02-26 23:04 ` [PATCH 2/6] scsi: iscsi: Speed up session unblocking and removal Mike Christie
2022-02-28 16:05 ` Lee Duncan
2022-02-28 20:06 ` Chris Leech
2022-02-26 23:04 ` [PATCH 3/6] scsi: iscsi: Remove iscsi_scan_finished Mike Christie
2022-02-28 18:05 ` Lee Duncan
2022-02-28 20:39 ` Mike Christie
2022-02-28 20:07 ` Chris Leech
2022-02-26 23:04 ` [PATCH 4/6] scsi: iscsi, ql4: Use per session workqueue for unbinding Mike Christie
2022-02-28 18:19 ` Lee Duncan [this message]
2022-02-28 20:07 ` Chris Leech
2022-02-26 23:04 ` [PATCH 5/6] scsi: iscsi: Use the session workqueue for recovery Mike Christie
2022-02-28 20:08 ` Chris Leech
2022-02-28 20:09 ` Lee Duncan
2022-02-26 23:04 ` [PATCH 6/6] scsi: iscsi: Drop temp workq_name Mike Christie
2022-02-28 20:08 ` Chris Leech
2022-02-28 22:49 ` Lee Duncan
2022-02-28 15:53 ` [PATCH 0/6] iscsi: Speed up failover with lots of devices Mike Christie
2022-03-02 4:20 ` Martin K. Petersen
2022-03-09 4:14 ` Martin K. Petersen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=fa265ba8-b0d4-0d02-ac32-27abab0e9f3d@suse.com \
--to=lduncan@suse.com \
--cc=GR-QLogic-Storage-Upstream@marvell.com \
--cc=cleech@redhat.com \
--cc=linux-scsi@vger.kernel.org \
--cc=liuzhengyuang521@gmail.com \
--cc=martin.petersen@oracle.com \
--cc=michael.christie@oracle.com \
--cc=mrangankar@marvell.com \
--cc=njavali@marvell.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox