patches for DPDK stable branches
 help / color / mirror / Atom feed
* [dpdk-stable] [Patch v2] net/vdev_netvsc: prevent alarm loss on failed device probe
@ 2020-10-22  8:11 Long Li
  2020-10-26  7:14 ` Matan Azrad
  0 siblings, 1 reply; 3+ messages in thread
From: Long Li @ 2020-10-22  8:11 UTC (permalink / raw)
  To: Matan Azrad; +Cc: dev, Long Li, stable

From: Long Li <longli@microsoft.com>

If a device probe fails, the alarm is canceled and will no longer work for
previously probed devices.

Fix this by checking if alarm is necessary at the end of each device probe.
Reset the alarm if there are vdev_netvsc_ctx created.

Change log:
v2: removed lock and flags, use counter to decide if alarm should be reset

Cc: stable@dpdk.org
Signed-off-by: Long Li <longli@microsoft.com>
---
 drivers/net/vdev_netvsc/vdev_netvsc.c | 27 +++++++++++++--------------
 1 file changed, 13 insertions(+), 14 deletions(-)

diff --git a/drivers/net/vdev_netvsc/vdev_netvsc.c b/drivers/net/vdev_netvsc/vdev_netvsc.c
index a865a82811..9ed8a4949d 100644
--- a/drivers/net/vdev_netvsc/vdev_netvsc.c
+++ b/drivers/net/vdev_netvsc/vdev_netvsc.c
@@ -684,6 +684,7 @@ vdev_netvsc_vdev_probe(struct rte_vdev_device *dev)
 	int ret;
 
 	DRV_LOG(DEBUG, "invoked as \"%s\", using arguments \"%s\"", name, args);
+	rte_eal_alarm_cancel(vdev_netvsc_alarm, NULL);
 	if (!kvargs) {
 		DRV_LOG(ERR, "cannot parse arguments list");
 		goto error;
@@ -699,17 +700,13 @@ vdev_netvsc_vdev_probe(struct rte_vdev_device *dev)
 			 !strcmp(pair->key, VDEV_NETVSC_ARG_MAC))
 			++specified;
 	}
-	if (ignore) {
-		if (kvargs)
-			rte_kvargs_free(kvargs);
-		return 0;
-	}
+	if (ignore)
+		goto ignore;
 	if (specified > 1) {
 		DRV_LOG(ERR, "More than one way used to specify the netvsc"
 			" device.");
 		goto error;
 	}
-	rte_eal_alarm_cancel(vdev_netvsc_alarm, NULL);
 	/* Gather interfaces. */
 	ret = vdev_netvsc_foreach_iface(vdev_netvsc_netvsc_probe, 1, name,
 					kvargs, specified, &matched);
@@ -730,17 +727,19 @@ vdev_netvsc_vdev_probe(struct rte_vdev_device *dev)
 		}
 		DRV_LOG(WARNING, "non-netvsc device was probed as netvsc");
 	}
-	ret = rte_eal_alarm_set(VDEV_NETVSC_PROBE_MS * 1000,
-				vdev_netvsc_alarm, NULL);
-	if (ret < 0) {
-		DRV_LOG(ERR, "unable to schedule alarm callback: %s",
-			rte_strerror(-ret));
-		goto error;
-	}
 error:
+	++vdev_netvsc_ctx_inst;
+ignore:
 	if (kvargs)
 		rte_kvargs_free(kvargs);
-	++vdev_netvsc_ctx_inst;
+	/* Reset alarm if there are device context created */
+	if (vdev_netvsc_ctx_count) {
+		ret = rte_eal_alarm_set(VDEV_NETVSC_PROBE_MS * 1000,
+					vdev_netvsc_alarm, NULL);
+		if (ret < 0)
+			DRV_LOG(ERR, "unable to schedule alarm callback: %s",
+				rte_strerror(-ret));
+	}
 	return 0;
 }
 
-- 
2.25.1


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [dpdk-stable] [Patch v2] net/vdev_netvsc: prevent alarm loss on failed device probe
  2020-10-22  8:11 [dpdk-stable] [Patch v2] net/vdev_netvsc: prevent alarm loss on failed device probe Long Li
@ 2020-10-26  7:14 ` Matan Azrad
  2020-10-29 20:10   ` Ferruh Yigit
  0 siblings, 1 reply; 3+ messages in thread
From: Matan Azrad @ 2020-10-26  7:14 UTC (permalink / raw)
  To: Long Li, Matan Azrad; +Cc: dev, NBU-Contact-longli, stable

Hi Long

From: Long Li <longli@linuxonhyperv.com>
> If a device probe fails, the alarm is canceled and will no longer work for
> previously probed devices.
> 
> Fix this by checking if alarm is necessary at the end of each device probe.
> Reset the alarm if there are vdev_netvsc_ctx created.
> 
> Change log:
> v2: removed lock and flags, use counter to decide if alarm should be reset
> 
> Cc: stable@dpdk.org
> Signed-off-by: Long Li <longli@microsoft.com>

I suggest the next title:
net/vdev_netvsc: fix device probing error flow

and the next fixes line:
Fixes: e7dc5d7becc5 ("net/vdev_netvsc: implement core functionality")

Acked-by: Matan Azrad <matan@nvidia.com>

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [dpdk-stable] [Patch v2] net/vdev_netvsc: prevent alarm loss on failed device probe
  2020-10-26  7:14 ` Matan Azrad
@ 2020-10-29 20:10   ` Ferruh Yigit
  0 siblings, 0 replies; 3+ messages in thread
From: Ferruh Yigit @ 2020-10-29 20:10 UTC (permalink / raw)
  To: Matan Azrad, Long Li, Matan Azrad; +Cc: dev, NBU-Contact-longli, stable

On 10/26/2020 7:14 AM, Matan Azrad wrote:
> Hi Long
> 
> From: Long Li <longli@linuxonhyperv.com>
>> If a device probe fails, the alarm is canceled and will no longer work for
>> previously probed devices.
>>
>> Fix this by checking if alarm is necessary at the end of each device probe.
>> Reset the alarm if there are vdev_netvsc_ctx created.
>>
>> Change log:
>> v2: removed lock and flags, use counter to decide if alarm should be reset
>>
>> Cc: stable@dpdk.org
>> Signed-off-by: Long Li <longli@microsoft.com>
> 
> I suggest the next title:
> net/vdev_netvsc: fix device probing error flow
> 
> and the next fixes line:
> Fixes: e7dc5d7becc5 ("net/vdev_netvsc: implement core functionality")
> 
> Acked-by: Matan Azrad <matan@nvidia.com>
> 

Applied to dpdk-next-net/main, thanks.

(Used suggested title)

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2020-10-29 20:10 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-10-22  8:11 [dpdk-stable] [Patch v2] net/vdev_netvsc: prevent alarm loss on failed device probe Long Li
2020-10-26  7:14 ` Matan Azrad
2020-10-29 20:10   ` Ferruh Yigit

patches for DPDK stable branches

This inbox may be cloned and mirrored by anyone:

	git clone --mirror http://inbox.dpdk.org/stable/0 stable/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 stable stable/ http://inbox.dpdk.org/stable \
		stable@dpdk.org
	public-inbox-index stable

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.stable


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git