patches for DPDK stable branches
 help / color / mirror / Atom feed
From: Thomas Monjalon <thomas@monjalon.net>
To: matan@mellanox.com
Cc: stable@dpdk.org
Subject: [dpdk-stable] [PATCH 08/11] ethdev: fix port visibility before initialization
Date: Wed,  9 May 2018 11:31:02 +0200	[thread overview]
Message-ID: <20180509093105.25984-9-thomas@monjalon.net> (raw)
In-Reply-To: <20180509093105.25984-1-thomas@monjalon.net>

The port was set to the state ATTACHED during allocation.
The consequence was to iterate over ports which are not initialized.

The state ATTACHED is now set as the last step of probing.

The uniqueness of port name is now checked before the availability
of a port id for allocation (order reversed).

As the state is not set on allocation anymore, it is also not checked
in the function telling whether a port is allocated or not.
The name of the port is set on allocation, so it is enough as a check.

Fixes: 5588909af21b ("ethdev: add device iterator")
Cc: stable@dpdk.org

Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Signed-off-by: Matan Azrad <matan@mellanox.com>
---
 lib/librte_ethdev/rte_ethdev.c        | 18 +++++++++---------
 lib/librte_ethdev/rte_ethdev_driver.h |  2 ++
 2 files changed, 11 insertions(+), 9 deletions(-)

diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
index 357be2dca..91cd0db11 100644
--- a/lib/librte_ethdev/rte_ethdev.c
+++ b/lib/librte_ethdev/rte_ethdev.c
@@ -233,7 +233,7 @@ rte_eth_dev_allocated_lock_free(const char *name)
 	unsigned i;
 
 	for (i = 0; i < RTE_MAX_ETHPORTS; i++) {
-		if ((rte_eth_devices[i].state == RTE_ETH_DEV_ATTACHED) &&
+		if (rte_eth_devices[i].data != NULL &&
 		    strcmp(rte_eth_devices[i].data->name, name) == 0)
 			return &rte_eth_devices[i];
 	}
@@ -278,7 +278,6 @@ eth_dev_get(uint16_t port_id)
 	struct rte_eth_dev *eth_dev = &rte_eth_devices[port_id];
 
 	eth_dev->data = &rte_eth_dev_shared_data->data[port_id];
-	eth_dev->state = RTE_ETH_DEV_ATTACHED;
 
 	eth_dev_last_created_port = port_id;
 
@@ -296,16 +295,15 @@ rte_eth_dev_allocate(const char *name)
 	/* Synchronize port creation between primary and secondary threads. */
 	rte_spinlock_lock(&rte_eth_dev_shared_data->ownership_lock);
 
-	port_id = rte_eth_dev_find_free_port();
-	if (port_id == RTE_MAX_ETHPORTS) {
-		ethdev_log(ERR, "Reached maximum number of Ethernet ports");
+	if (rte_eth_dev_allocated_lock_free(name) != NULL) {
+		ethdev_log(ERR, "Ethernet device with name %s already allocated",
+				name);
 		goto unlock;
 	}
 
-	if (rte_eth_dev_allocated_lock_free(name) != NULL) {
-		ethdev_log(ERR,
-			"Ethernet Device with name %s already allocated!",
-			name);
+	port_id = rte_eth_dev_find_free_port();
+	if (port_id == RTE_MAX_ETHPORTS) {
+		ethdev_log(ERR, "Reached maximum number of Ethernet ports");
 		goto unlock;
 	}
 
@@ -3387,6 +3385,8 @@ rte_eth_dev_probing_finish(struct rte_eth_dev *dev)
 {
 	if (dev == NULL)
 		return;
+
+	dev->state = RTE_ETH_DEV_ATTACHED;
 }
 
 int
diff --git a/lib/librte_ethdev/rte_ethdev_driver.h b/lib/librte_ethdev/rte_ethdev_driver.h
index 3821f0b1d..3640dff68 100644
--- a/lib/librte_ethdev/rte_ethdev_driver.h
+++ b/lib/librte_ethdev/rte_ethdev_driver.h
@@ -106,6 +106,8 @@ int _rte_eth_dev_callback_process(struct rte_eth_dev *dev,
  * This is the last step of device probing.
  * It must be called after a port is allocated and initialized successfully.
  *
+ * The state is set as RTE_ETH_DEV_ATTACHED.
+ *
  * @param dev
  *  New ethdev port.
  */
-- 
2.16.2

  parent reply	other threads:[~2018-05-09  9:31 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20180509093105.25984-1-thomas@monjalon.net>
2018-05-09  9:30 ` [dpdk-stable] [PATCH 01/11] ethdev: fix debug log of owner id Thomas Monjalon
2018-05-09 10:44   ` Luca Boccassi
2018-05-09 12:15     ` Thomas Monjalon
2018-05-09 13:59       ` Luca Boccassi
2018-05-09 14:52         ` Thomas Monjalon
2018-05-09  9:30 ` [dpdk-stable] [PATCH 02/11] net/failsafe: fix sub-device visibility Thomas Monjalon
2018-05-09  9:30 ` [dpdk-stable] [PATCH 04/11] drivers/net: use higher level of probing helper for PCI Thomas Monjalon
2018-05-09  9:30 ` [dpdk-stable] [PATCH 05/11] ethdev: add probing finish function Thomas Monjalon
2018-05-09  9:31 ` [dpdk-stable] [PATCH 06/11] ethdev: allow ownership operations on unused port Thomas Monjalon
2018-05-09  9:31 ` [dpdk-stable] [PATCH 07/11] ethdev: add lock to port allocation check Thomas Monjalon
2018-05-09  9:31 ` Thomas Monjalon [this message]
2018-05-09  9:31 ` [dpdk-stable] [PATCH 09/11] ethdev: fix port probing notification Thomas Monjalon
2018-05-09  9:31 ` [dpdk-stable] [PATCH 10/11] net/failsafe: fix sub-device ownership race Thomas Monjalon
2018-05-09  9:31 ` [dpdk-stable] [PATCH 11/11] ethdev: fix port removal notification timing Thomas Monjalon
     [not found] <20180509094337.26112-1-thomas@monjalon.net>
2018-05-09  9:43 ` [dpdk-stable] [PATCH 08/11] ethdev: fix port visibility before initialization Thomas Monjalon
2018-05-09 18:03   ` Ferruh Yigit
2018-05-09 19:08     ` Thomas Monjalon

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=20180509093105.25984-9-thomas@monjalon.net \
    --to=thomas@monjalon.net \
    --cc=matan@mellanox.com \
    --cc=stable@dpdk.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).