DPDK patches and discussions
 help / color / mirror / Atom feed
From: Jeff Guo <jia.guo@intel.com>
To: konstantin.ananyev@intel.com, anatoly.burakov@intel.com,
	thomas@monjalon.net, bernard.iremonger@intel.com,
	jingjing.wu@intel.com, wenzhuo.lu@intel.com
Cc: ferruh.yigit@intel.com, dev@dpdk.org, jia.guo@intel.com,
	helin.zhang@intel.com, matan@mellanox.com, shaopeng.he@intel.com
Subject: [dpdk-dev] [PATCH V2 3/3] app/testpmd: fix callback issue for hot-unplug
Date: Thu, 15 Nov 2018 17:18:24 +0800	[thread overview]
Message-ID: <1542273504-127387-5-git-send-email-jia.guo@intel.com> (raw)
In-Reply-To: <1542273504-127387-1-git-send-email-jia.guo@intel.com>

Because the user's callback is invoked in eal interrupt callback, the
interrupt callback need to be finished before it can be unregistered
when detaching device. So finish callback soon and use a deferred
removal to detach device is need.

It is a workaround, once the device detaching be moved into the eal in
the future, the deferred removal could be deleted. This patch aim to
add this workaround and refine the function name and the description to
be more explicit and comment the limitation.

Fixes: 2049c5113fe8 ("app/testpmd: use hotplug failure handler")
Signed-off-by: Jeff Guo <jia.guo@intel.com>
---
v2->v1:
rename the function, refine the doc, and show the limitation.
---
 app/test-pmd/testpmd.c | 34 ++++++++++++++++++++++++++--------
 1 file changed, 26 insertions(+), 8 deletions(-)

diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index 9c0edca..4c75587 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testpmd.c
@@ -506,7 +506,7 @@ static void check_all_ports_link_status(uint32_t port_mask);
 static int eth_event_callback(portid_t port_id,
 			      enum rte_eth_event_type type,
 			      void *param, void *ret_param);
-static void eth_dev_event_callback(const char *device_name,
+static void dev_event_callback(const char *device_name,
 				enum rte_dev_event_type type,
 				void *param);
 
@@ -2434,7 +2434,7 @@ pmd_test_exit(void)
 		}
 
 		ret = rte_dev_event_callback_unregister(NULL,
-			eth_dev_event_callback, NULL);
+			dev_event_callback, NULL);
 		if (ret < 0) {
 			RTE_LOG(ERR, EAL,
 				"fail to unregister device event callback.\n");
@@ -2516,8 +2516,14 @@ check_all_ports_link_status(uint32_t port_mask)
 	}
 }
 
+/*
+ * This callback is for remove a port for a device. It has limitation because
+ * it is not for multiple port removal for a device.
+ * TODO: the device detach invoke will plan to be removed from user side to
+ * eal. And convert all PMDs to free port resources on ether device closing.
+ */
 static void
-rmv_event_callback(void *arg)
+rmv_port_callback(void *arg)
 {
 	int need_to_start = 0;
 	int org_no_link_check = no_link_check;
@@ -2565,7 +2571,7 @@ eth_event_callback(portid_t port_id, enum rte_eth_event_type type, void *param,
 		if (port_id_is_invalid(port_id, DISABLED_WARN))
 			break;
 		if (rte_eal_alarm_set(100000,
-				rmv_event_callback, (void *)(intptr_t)port_id))
+				rmv_port_callback, (void *)(intptr_t)port_id))
 			fprintf(stderr, "Could not set up deferred device removal\n");
 		break;
 	default:
@@ -2598,7 +2604,7 @@ register_eth_event_callback(void)
 
 /* This function is used by the interrupt thread */
 static void
-eth_dev_event_callback(const char *device_name, enum rte_dev_event_type type,
+dev_event_callback(const char *device_name, enum rte_dev_event_type type,
 			     __rte_unused void *arg)
 {
 	uint16_t port_id;
@@ -2612,7 +2618,7 @@ eth_dev_event_callback(const char *device_name, enum rte_dev_event_type type,
 
 	switch (type) {
 	case RTE_DEV_EVENT_REMOVE:
-		RTE_LOG(ERR, EAL, "The device: %s has been removed!\n",
+		RTE_LOG(DEBUG, EAL, "The device: %s has been removed!\n",
 			device_name);
 		ret = rte_eth_dev_get_port_by_name(device_name, &port_id);
 		if (ret) {
@@ -2620,7 +2626,19 @@ eth_dev_event_callback(const char *device_name, enum rte_dev_event_type type,
 				device_name);
 			return;
 		}
-		rmv_event_callback((void *)(intptr_t)port_id);
+		/*
+		 * Because the user's callback is invoked in eal interrupt
+		 * callback, the interrupt callback need to be finished before
+		 * it can be unregistered when detaching device. So finish
+		 * callback soon and use a deferred removal to detach device
+		 * is need. It is a workaround, once the device detaching be
+		 * moved into the eal in the future, the deferred removal could
+		 * be deleted.
+		 */
+		if (rte_eal_alarm_set(100000,
+				rmv_port_callback, (void *)(intptr_t)port_id))
+			RTE_LOG(ERR, EAL,
+				"Could not set up deferred device removal\n");
 		break;
 	case RTE_DEV_EVENT_ADD:
 		RTE_LOG(ERR, EAL, "The device: %s has been added!\n",
@@ -3167,7 +3185,7 @@ main(int argc, char** argv)
 		}
 
 		ret = rte_dev_event_callback_register(NULL,
-			eth_dev_event_callback, NULL);
+			dev_event_callback, NULL);
 		if (ret) {
 			RTE_LOG(ERR, EAL,
 				"fail  to register device event callback\n");
-- 
2.7.4

      parent reply	other threads:[~2018-11-15  9:14 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-11-06  6:07 [dpdk-dev] [PATCH 0/3] fix vfio hot-unplug issue Jeff Guo
2018-11-06  6:07 ` [dpdk-dev] [PATCH 1/3] eal: fix lock issue for hot-unplug Jeff Guo
2018-11-06  6:22   ` Matan Azrad
2018-11-07  5:49     ` Jeff Guo
2018-11-08  7:08       ` Matan Azrad
2018-11-06  6:07 ` [dpdk-dev] [PATCH 2/3] vfio: fix to add handler lock " Jeff Guo
2018-11-06  6:23   ` Matan Azrad
2018-11-07  6:15     ` Jeff Guo
2018-11-08  7:20       ` Matan Azrad
2018-11-08  8:30         ` Jeff Guo
2018-11-06  6:07 ` [dpdk-dev] [PATCH 3/3] app/testpmd: fix callback issue " Jeff Guo
2018-11-06  6:36   ` Matan Azrad
2018-11-07  7:30     ` Jeff Guo
2018-11-07 11:05       ` Ananyev, Konstantin
2018-11-08  7:28         ` Matan Azrad
2018-11-08  8:49           ` Jeff Guo
2018-11-08  9:35             ` Matan Azrad
2018-11-09  3:55               ` Jeff Guo
2018-11-09  5:24                 ` Matan Azrad
2018-11-09  6:17                   ` Jeff Guo
     [not found]                     ` <AM0PR0502MB401938411A7E2BA9E76576A2D2C00@AM0PR0502MB4019.eurprd05.prod.outlook.com>
2018-11-12  1:35                       ` Thomas Monjalon
2018-11-14  9:32                         ` Jeff Guo
2018-11-15  9:18 ` [dpdk-dev] [PATCH V2 0/3] fix vfio hot-unplug issue Jeff Guo
2018-11-15  9:18   ` Jeff Guo
2018-11-18 16:19     ` Thomas Monjalon
2018-11-15  9:18   ` [dpdk-dev] [PATCH V2 1/3] eal: fix lock issue for hot-unplug Jeff Guo
2018-11-15  9:18   ` [dpdk-dev] [PATCH V2 2/3] vfio: fix to add handler lock " Jeff Guo
2018-11-15  9:18   ` Jeff Guo [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=1542273504-127387-5-git-send-email-jia.guo@intel.com \
    --to=jia.guo@intel.com \
    --cc=anatoly.burakov@intel.com \
    --cc=bernard.iremonger@intel.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=helin.zhang@intel.com \
    --cc=jingjing.wu@intel.com \
    --cc=konstantin.ananyev@intel.com \
    --cc=matan@mellanox.com \
    --cc=shaopeng.he@intel.com \
    --cc=thomas@monjalon.net \
    --cc=wenzhuo.lu@intel.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;
as well as URLs for NNTP newsgroup(s).