diff mbox

[dpdk-dev,v3,1/6] ethdev: add devop to check removal status

Message ID 1513703415-29145-2-git-send-email-matan@mellanox.com (mailing list archive)
State Superseded, archived
Delegated to: Ferruh Yigit
Headers

Checks

Context Check Description
ci/checkpatch success coding style OK
ci/Intel-compilation success Compilation OK

Commit Message

Matan Azrad Dec. 19, 2017, 5:10 p.m. UTC 
  There is time between the physical removal of the device until PMDs get
a RMV interrupt. At this time DPDK PMDs and applications still don't
know about the removal.

Current removal detection is achieved only by registration to device RMV
event and the notification comes asynchronously. So, there is no option
to detect a device removal synchronously.
Applications and other DPDK entities may want to check a device removal
synchronously and to take an immediate decision accordingly.

Add new dev op called is_removed to allow DPDK entities to check an
Ethernet device removal status immediately.

Signed-off-by: Matan Azrad <matan@mellanox.com>
---
 lib/librte_ether/rte_ethdev.c           | 28 +++++++++++++++++++++++++---
 lib/librte_ether/rte_ethdev.h           | 17 +++++++++++++++++
 lib/librte_ether/rte_ethdev_version.map |  7 +++++++
 3 files changed, 49 insertions(+), 3 deletions(-)

Comments

Stephen Hemminger Dec. 19, 2017, 5:20 p.m. UTC | #1 
On Tue, 19 Dec 2017 17:10:10 +0000
Matan Azrad <matan@mellanox.com> wrote:

>  int
> +rte_eth_dev_is_removed(uint16_t port_id)
> +{
> +	struct rte_eth_dev *dev;
> +	int ret;
> +
> +	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
> +
> +	dev = &rte_eth_devices[port_id];
> +
> +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->is_removed, 0);
> +
> +	if (dev->state == RTE_ETH_DEV_REMOVED)
> +		return 1;
> +
> +	ret = dev->dev_ops->is_removed(dev);
> +	if (ret != 0)
> +		dev->state = RTE_ETH_DEV_REMOVED;
> +
> +	return ret;
> +}
> +

This looks good.
May be a candidate to use bool instead of int for return value?
Matan Azrad Dec. 19, 2017, 5:24 p.m. UTC | #2 
HI

> -----Original Message-----
> From: Stephen Hemminger [mailto:stephen@networkplumber.org]
> Sent: Tuesday, December 19, 2017 7:20 PM
> To: Matan Azrad <matan@mellanox.com>
> Cc: Adrien Mazarguil <adrien.mazarguil@6wind.com>; Thomas Monjalon
> <thomas@monjalon.net>; Gaetan Rivet <gaetan.rivet@6wind.com>;
> dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v3 1/6] ethdev: add devop to check removal
> status
> 
> On Tue, 19 Dec 2017 17:10:10 +0000
> Matan Azrad <matan@mellanox.com> wrote:
> 
> >  int
> > +rte_eth_dev_is_removed(uint16_t port_id) {
> > +	struct rte_eth_dev *dev;
> > +	int ret;
> > +
> > +	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
> > +
> > +	dev = &rte_eth_devices[port_id];
> > +
> > +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->is_removed, 0);
> > +
> > +	if (dev->state == RTE_ETH_DEV_REMOVED)
> > +		return 1;
> > +
> > +	ret = dev->dev_ops->is_removed(dev);
> > +	if (ret != 0)
> > +		dev->state = RTE_ETH_DEV_REMOVED;
> > +
> > +	return ret;
> > +}
> > +
> 
> This looks good.
> May be a candidate to use bool instead of int for return value?

Yes, I thought about it but didn't see any precedence for bool usage in ethdev APIs.
Guys, what do you think?
Thomas Monjalon Dec. 19, 2017, 8:51 p.m. UTC | #3 
19/12/2017 18:24, Matan Azrad:
> HI
> 
> > -----Original Message-----
> > From: Stephen Hemminger [mailto:stephen@networkplumber.org]
> > Sent: Tuesday, December 19, 2017 7:20 PM
> > To: Matan Azrad <matan@mellanox.com>
> > Cc: Adrien Mazarguil <adrien.mazarguil@6wind.com>; Thomas Monjalon
> > <thomas@monjalon.net>; Gaetan Rivet <gaetan.rivet@6wind.com>;
> > dev@dpdk.org
> > Subject: Re: [dpdk-dev] [PATCH v3 1/6] ethdev: add devop to check removal
> > status
> > 
> > On Tue, 19 Dec 2017 17:10:10 +0000
> > Matan Azrad <matan@mellanox.com> wrote:
> > 
> > >  int
> > > +rte_eth_dev_is_removed(uint16_t port_id) {
> > > +	struct rte_eth_dev *dev;
> > > +	int ret;
> > > +
> > > +	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
> > > +
> > > +	dev = &rte_eth_devices[port_id];
> > > +
> > > +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->is_removed, 0);
> > > +
> > > +	if (dev->state == RTE_ETH_DEV_REMOVED)
> > > +		return 1;
> > > +
> > > +	ret = dev->dev_ops->is_removed(dev);
> > > +	if (ret != 0)
> > > +		dev->state = RTE_ETH_DEV_REMOVED;
> > > +
> > > +	return ret;
> > > +}
> > > +
> > 
> > This looks good.
> > May be a candidate to use bool instead of int for return value?
> 
> Yes, I thought about it but didn't see any precedence for bool usage in ethdev APIs.
> Guys, what do you think?

I think this function can return error, isn't it?
(look at macros *_OR_ERR_RET used in the function)
Gaëtan Rivet Dec. 19, 2017, 10:13 p.m. UTC | #4 
On Tue, Dec 19, 2017 at 09:51:10PM +0100, Thomas Monjalon wrote:
> 19/12/2017 18:24, Matan Azrad:
> > HI
> > 
> > > -----Original Message-----
> > > From: Stephen Hemminger [mailto:stephen@networkplumber.org]
> > > Sent: Tuesday, December 19, 2017 7:20 PM
> > > To: Matan Azrad <matan@mellanox.com>
> > > Cc: Adrien Mazarguil <adrien.mazarguil@6wind.com>; Thomas Monjalon
> > > <thomas@monjalon.net>; Gaetan Rivet <gaetan.rivet@6wind.com>;
> > > dev@dpdk.org
> > > Subject: Re: [dpdk-dev] [PATCH v3 1/6] ethdev: add devop to check removal
> > > status
> > > 
> > > On Tue, 19 Dec 2017 17:10:10 +0000
> > > Matan Azrad <matan@mellanox.com> wrote:
> > > 
> > > >  int
> > > > +rte_eth_dev_is_removed(uint16_t port_id) {
> > > > +	struct rte_eth_dev *dev;
> > > > +	int ret;
> > > > +
> > > > +	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
> > > > +
> > > > +	dev = &rte_eth_devices[port_id];
> > > > +
> > > > +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->is_removed, 0);
> > > > +
> > > > +	if (dev->state == RTE_ETH_DEV_REMOVED)
> > > > +		return 1;
> > > > +
> > > > +	ret = dev->dev_ops->is_removed(dev);
> > > > +	if (ret != 0)
> > > > +		dev->state = RTE_ETH_DEV_REMOVED;
> > > > +
> > > > +	return ret;
> > > > +}
> > > > +
> > > 
> > > This looks good.
> > > May be a candidate to use bool instead of int for return value?
> > 
> > Yes, I thought about it but didn't see any precedence for bool usage in ethdev APIs.
> > Guys, what do you think?
> 
> I think this function can return error, isn't it?
> (look at macros *_OR_ERR_RET used in the function)
> 

But those macros are used to return 0.

While I think I see a logic behind it, I think it is surprising the
API user, which is not ideal.
Matan Azrad Dec. 20, 2017, 8:39 a.m. UTC | #5 
Hi

> -----Original Message-----
> From: Gaëtan Rivet [mailto:gaetan.rivet@6wind.com]
> Sent: Wednesday, December 20, 2017 12:13 AM
> To: Thomas Monjalon <thomas@monjalon.net>
> Cc: Matan Azrad <matan@mellanox.com>; Stephen Hemminger
> <stephen@networkplumber.org>; Adrien Mazarguil
> <adrien.mazarguil@6wind.com>; dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v3 1/6] ethdev: add devop to check removal
> status
> 
> On Tue, Dec 19, 2017 at 09:51:10PM +0100, Thomas Monjalon wrote:
> > 19/12/2017 18:24, Matan Azrad:
> > > HI
> > >
> > > > -----Original Message-----
> > > > From: Stephen Hemminger [mailto:stephen@networkplumber.org]
> > > > Sent: Tuesday, December 19, 2017 7:20 PM
> > > > To: Matan Azrad <matan@mellanox.com>
> > > > Cc: Adrien Mazarguil <adrien.mazarguil@6wind.com>; Thomas
> Monjalon
> > > > <thomas@monjalon.net>; Gaetan Rivet <gaetan.rivet@6wind.com>;
> > > > dev@dpdk.org
> > > > Subject: Re: [dpdk-dev] [PATCH v3 1/6] ethdev: add devop to check
> > > > removal status
> > > >
> > > > On Tue, 19 Dec 2017 17:10:10 +0000 Matan Azrad
> > > > <matan@mellanox.com> wrote:
> > > >
> > > > >  int
> > > > > +rte_eth_dev_is_removed(uint16_t port_id) {
> > > > > +	struct rte_eth_dev *dev;
> > > > > +	int ret;
> > > > > +
> > > > > +	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
> > > > > +
> > > > > +	dev = &rte_eth_devices[port_id];
> > > > > +
> > > > > +	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops-
> >is_removed, 0);
> > > > > +
> > > > > +	if (dev->state == RTE_ETH_DEV_REMOVED)
> > > > > +		return 1;
> > > > > +
> > > > > +	ret = dev->dev_ops->is_removed(dev);
> > > > > +	if (ret != 0)
> > > > > +		dev->state = RTE_ETH_DEV_REMOVED;
> > > > > +
> > > > > +	return ret;
> > > > > +}
> > > > > +
> > > >
> > > > This looks good.
> > > > May be a candidate to use bool instead of int for return value?
> > >
> > > Yes, I thought about it but didn't see any precedence for bool usage in
> ethdev APIs.
> > > Guys, what do you think?
> >
> > I think this function can return error, isn't it?
> > (look at macros *_OR_ERR_RET used in the function)
> >
> 
> But those macros are used to return 0.
> 
> While I think I see a logic behind it, I think it is surprising the API user, which is
> not ideal.
> 

The logic behind it is that "is" semantic is question which expects to yes\no answer.
Therefore, user who uses this API just expects to either True or False return value and doesn't need to check more errors options like -ENODEV or -ENOTSUP.
I decided that the return value will be only 0 or 1 to make it easier to user:
Removed - 1,
Present - 0,
No support - it makes sense that PMD which doesn't implement "is_removed" devop means that its underlying devices are not removable so they are always present and this function should return '0' for it.
No port - I think that '0' here will be better that '1'.

Stephen suggestion to replace '0' and '1' by 'false' and 'true' makes sense but I decided not to do it like this because of the next ideas:
1. No precedence for bool value in ethdev APIs.
2. Maybe it will be problematic to use *OR_ERR_RET defines to return bool value. 

I hope that this explanation helps you.

Thanks, 
Matan.
   
  


   
> --
> Gaëtan Rivet
> 6WIND
Thomas Monjalon Jan. 7, 2018, 9:53 a.m. UTC | #6 
19/12/2017 18:10, Matan Azrad:
> There is time between the physical removal of the device until PMDs get
> a RMV interrupt. At this time DPDK PMDs and applications still don't
> know about the removal.
> 
> Current removal detection is achieved only by registration to device RMV
> event and the notification comes asynchronously. So, there is no option
> to detect a device removal synchronously.
> Applications and other DPDK entities may want to check a device removal
> synchronously and to take an immediate decision accordingly.
> 
> Add new dev op called is_removed to allow DPDK entities to check an
> Ethernet device removal status immediately.
> 
> Signed-off-by: Matan Azrad <matan@mellanox.com>

Acked-by: Thomas Monjalon <thomas@monjalon.net>
diff mbox

Patch

diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
index 318af28..c759d0e 100644
--- a/lib/librte_ether/rte_ethdev.c
+++ b/lib/librte_ether/rte_ethdev.c
@@ -142,7 +142,8 @@  enum {
 rte_eth_find_next(uint16_t port_id)
 {
 	while (port_id < RTE_MAX_ETHPORTS &&
-	       rte_eth_devices[port_id].state != RTE_ETH_DEV_ATTACHED)
+	       rte_eth_devices[port_id].state != RTE_ETH_DEV_ATTACHED &&
+	       rte_eth_devices[port_id].state != RTE_ETH_DEV_REMOVED)
 		port_id++;
 
 	if (port_id >= RTE_MAX_ETHPORTS)
@@ -286,8 +287,7 @@  struct rte_eth_dev *
 rte_eth_dev_is_valid_port(uint16_t port_id)
 {
 	if (port_id >= RTE_MAX_ETHPORTS ||
-	    (rte_eth_devices[port_id].state != RTE_ETH_DEV_ATTACHED &&
-	     rte_eth_devices[port_id].state != RTE_ETH_DEV_DEFERRED))
+	    (rte_eth_devices[port_id].state == RTE_ETH_DEV_UNUSED))
 		return 0;
 	else
 		return 1;
@@ -1118,6 +1118,28 @@  struct rte_eth_dev *
 }
 
 int
+rte_eth_dev_is_removed(uint16_t port_id)
+{
+	struct rte_eth_dev *dev;
+	int ret;
+
+	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
+
+	dev = &rte_eth_devices[port_id];
+
+	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->is_removed, 0);
+
+	if (dev->state == RTE_ETH_DEV_REMOVED)
+		return 1;
+
+	ret = dev->dev_ops->is_removed(dev);
+	if (ret != 0)
+		dev->state = RTE_ETH_DEV_REMOVED;
+
+	return ret;
+}
+
+int
 rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
 		       uint16_t nb_rx_desc, unsigned int socket_id,
 		       const struct rte_eth_rxconf *rx_conf,
diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
index 341c2d6..3aa9d3f 100644
--- a/lib/librte_ether/rte_ethdev.h
+++ b/lib/librte_ether/rte_ethdev.h
@@ -1196,6 +1196,9 @@  struct rte_eth_dcb_info {
 typedef int (*eth_dev_reset_t)(struct rte_eth_dev *dev);
 /** <@internal Function used to reset a configured Ethernet device. */
 
+typedef int (*eth_is_removed_t)(struct rte_eth_dev *dev);
+/**< @internal Function used to detect an Ethernet device removal. */
+
 typedef void (*eth_promiscuous_enable_t)(struct rte_eth_dev *dev);
 /**< @internal Function used to enable the RX promiscuous mode of an Ethernet device. */
 
@@ -1525,6 +1528,8 @@  struct eth_dev_ops {
 	eth_dev_close_t            dev_close;     /**< Close device. */
 	eth_dev_reset_t		   dev_reset;	  /**< Reset device. */
 	eth_link_update_t          link_update;   /**< Get device link state. */
+	eth_is_removed_t           is_removed;
+	/**< Check if the device was physically removed. */
 
 	eth_promiscuous_enable_t   promiscuous_enable; /**< Promiscuous ON. */
 	eth_promiscuous_disable_t  promiscuous_disable;/**< Promiscuous OFF. */
@@ -1711,6 +1716,7 @@  enum rte_eth_dev_state {
 	RTE_ETH_DEV_UNUSED = 0,
 	RTE_ETH_DEV_ATTACHED,
 	RTE_ETH_DEV_DEFERRED,
+	RTE_ETH_DEV_REMOVED,
 };
 
 /**
@@ -1997,6 +2003,17 @@  int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
 void _rte_eth_dev_reset(struct rte_eth_dev *dev);
 
 /**
+ * Check if an Ethernet device was physically removed.
+ *
+ * @param port_id
+ *   The port identifier of the Ethernet device.
+ * @return
+ *   - 0 when the Ethernet device is removed, otherwise 1.
+ */
+int
+rte_eth_dev_is_removed(uint16_t port_id);
+
+/**
  * Allocate and set up a receive queue for an Ethernet device.
  *
  * The function allocates a contiguous block of memory for *nb_rx_desc*
diff --git a/lib/librte_ether/rte_ethdev_version.map b/lib/librte_ether/rte_ethdev_version.map
index e9681ac..78547ff 100644
--- a/lib/librte_ether/rte_ethdev_version.map
+++ b/lib/librte_ether/rte_ethdev_version.map
@@ -198,6 +198,13 @@  DPDK_17.11 {
 
 } DPDK_17.08;
 
+DPDK_18.02 {
+	global:
+
+	rte_eth_dev_is_removed;
+
+} DPDK_17.11;
+
 EXPERIMENTAL {
 	global: