From: Mike Christie <michael.christie@oracle.com>
To: target-devel@vger.kernel.org, linux-scsi@vger.kernel.org,
stefanha@redhat.com, jasowang@redhat.com, mst@redhat.com,
sgarzare@redhat.com, virtualization@lists.linux-foundation.org
Cc: Mike Christie <michael.christie@oracle.com>
Subject: [PATCH 5/5] vhost-scsi: Reduce vhost_scsi_mutex use
Date: Wed, 22 Feb 2023 18:19:49 -0600 [thread overview]
Message-ID: <20230223001949.2884-6-michael.christie@oracle.com> (raw)
In-Reply-To: <20230223001949.2884-1-michael.christie@oracle.com>
Now that vhost_scsi_clear_endpoint will prevent new IO and flush running
IO before dropping it's refcounts on the tpg, we only need to hold the
vhost_scsi_mutex while looping over tpgs and taking a refcount, and when
manipulating the tpg list.
This removes the vhost_scsi_mutex from vhost_scsi_clear_endpoint and moves
the vhost_scsi_mutex use in vhost_scsi_set_endpoint so it's only taken
while looping over the tpgs and taking a refcount. We can then now avoid
issues where vhost_scsi_set_endpoint has to wait for another device's
vhost_scsi_clear_endpoint call.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
---
drivers/vhost/scsi.c | 11 +++++------
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/drivers/vhost/scsi.c b/drivers/vhost/scsi.c
index c405ab5c232a..75ea24f1c571 100644
--- a/drivers/vhost/scsi.c
+++ b/drivers/vhost/scsi.c
@@ -1540,7 +1540,7 @@ static int vhost_scsi_setup_vq_cmds(struct vhost_virtqueue *vq, int max_cmds)
* vhost_scsi_tpg with an active struct vhost_scsi_nexus
*
* The lock nesting rule is:
- * vhost_scsi_mutex -> vs->dev.mutex -> tpg->tv_tpg_mutex -> vq->mutex
+ * vs->dev.mutex -> vhost_scsi_mutex -> tpg->tv_tpg_mutex -> vq->mutex
*/
static int
vhost_scsi_set_endpoint(struct vhost_scsi *vs,
@@ -1554,7 +1554,6 @@ vhost_scsi_set_endpoint(struct vhost_scsi *vs,
int index, ret, i, len;
bool match = false;
- mutex_lock(&vhost_scsi_mutex);
mutex_lock(&vs->dev.mutex);
/* Verify that ring has been setup correctly. */
@@ -1575,6 +1574,7 @@ vhost_scsi_set_endpoint(struct vhost_scsi *vs,
if (vs->vs_tpg)
memcpy(vs_tpg, vs->vs_tpg, len);
+ mutex_lock(&vhost_scsi_mutex);
list_for_each_entry(tpg, &vhost_scsi_list, tv_tpg_list) {
mutex_lock(&tpg->tv_tpg_mutex);
if (!tpg->tpg_nexus) {
@@ -1590,6 +1590,7 @@ vhost_scsi_set_endpoint(struct vhost_scsi *vs,
if (!strcmp(tv_tport->tport_name, t->vhost_wwpn)) {
if (vs->vs_tpg && vs->vs_tpg[tpg->tport_tpgt]) {
mutex_unlock(&tpg->tv_tpg_mutex);
+ mutex_unlock(&vhost_scsi_mutex);
ret = -EEXIST;
goto undepend;
}
@@ -1604,6 +1605,7 @@ vhost_scsi_set_endpoint(struct vhost_scsi *vs,
if (ret) {
pr_warn("target_depend_item() failed: %d\n", ret);
mutex_unlock(&tpg->tv_tpg_mutex);
+ mutex_unlock(&vhost_scsi_mutex);
goto undepend;
}
tpg->tv_tpg_vhost_count++;
@@ -1613,6 +1615,7 @@ vhost_scsi_set_endpoint(struct vhost_scsi *vs,
}
mutex_unlock(&tpg->tv_tpg_mutex);
}
+ mutex_unlock(&vhost_scsi_mutex);
if (match) {
memcpy(vs->vs_vhost_wwpn, t->vhost_wwpn,
@@ -1667,7 +1670,6 @@ vhost_scsi_set_endpoint(struct vhost_scsi *vs,
kfree(vs_tpg);
out:
mutex_unlock(&vs->dev.mutex);
- mutex_unlock(&vhost_scsi_mutex);
return ret;
}
@@ -1683,7 +1685,6 @@ vhost_scsi_clear_endpoint(struct vhost_scsi *vs,
int index, ret, i;
u8 target;
- mutex_lock(&vhost_scsi_mutex);
mutex_lock(&vs->dev.mutex);
/* Verify that ring has been setup correctly. */
for (index = 0; index < vs->dev.nvqs; ++index) {
@@ -1765,12 +1766,10 @@ vhost_scsi_clear_endpoint(struct vhost_scsi *vs,
vs->vs_tpg = NULL;
WARN_ON(vs->vs_events_nr);
mutex_unlock(&vs->dev.mutex);
- mutex_unlock(&vhost_scsi_mutex);
return 0;
err_dev:
mutex_unlock(&vs->dev.mutex);
- mutex_unlock(&vhost_scsi_mutex);
return ret;
}
--
2.25.1
prev parent reply other threads:[~2023-02-23 0:20 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-23 0:19 [PATCH 0/5] vhost-scsi: Fix management operation hangs Mike Christie
2023-02-23 0:19 ` [PATCH 1/5] vhost-scsi: Hold tv_tpg_mutex when decrementing tv_tpg_vhost_count Mike Christie
2023-02-23 0:19 ` [PATCH 2/5] vhost-scsi: Drop vhost_scsi_flush during endpoint clearing Mike Christie
2023-02-23 0:19 ` [PATCH 3/5] vhost-scsi: Remove vhost_scsi_mutex from port link/unlink Mike Christie
2023-03-19 15:48 ` Mike Christie
2023-02-23 0:19 ` [PATCH 4/5] vhost-scsi: Delay releasing our refcount on the tpg Mike Christie
2023-02-23 0:19 ` Mike Christie [this message]
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=20230223001949.2884-6-michael.christie@oracle.com \
--to=michael.christie@oracle.com \
--cc=jasowang@redhat.com \
--cc=linux-scsi@vger.kernel.org \
--cc=mst@redhat.com \
--cc=sgarzare@redhat.com \
--cc=stefanha@redhat.com \
--cc=target-devel@vger.kernel.org \
--cc=virtualization@lists.linux-foundation.org \
/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;
as well as URLs for NNTP newsgroup(s).