[dpdk-dev] [PATCH] ethdev: fix rte_eth_dev_owner_unset

[dpdk-dev] [PATCH] ethdev: fix rte_eth_dev_owner_unset

[Stephen Hemminger]

The rte_eth_dev_owner_unset function is unusable because
it always returns -EINVAL. This is because the magic (unowned)
value is flagged as not valid.

Move the validation of owner into set and unset as
separate calls.

Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
---
 lib/librte_ethdev/rte_ethdev.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
index 4c320250589a..9398550a1189 100644
--- a/lib/librte_ethdev/rte_ethdev.c
+++ b/lib/librte_ethdev/rte_ethdev.c
@@ -443,10 +443,6 @@ _rte_eth_dev_owner_set(const uint16_t port_id, const uint64_t old_owner_id,
 		return -ENODEV;
 	}
 
-	if (!rte_eth_is_valid_owner_id(new_owner->id) &&
-	    !rte_eth_is_valid_owner_id(old_owner_id))
-		return -EINVAL;
-
 	port_owner = &rte_eth_devices[port_id].data->owner;
 	if (port_owner->id != old_owner_id) {
 		RTE_ETHDEV_LOG(ERR,
@@ -475,6 +471,9 @@ rte_eth_dev_owner_set(const uint16_t port_id,
 {
 	int ret;
 
+	if (!rte_eth_is_valid_owner_id(owner->id))
+		return -EINVAL;
+
 	rte_eth_dev_shared_data_prepare();
 
 	rte_spinlock_lock(&rte_eth_dev_shared_data->ownership_lock);
@@ -492,6 +491,9 @@ rte_eth_dev_owner_unset(const uint16_t port_id, const uint64_t owner_id)
 			{.id = RTE_ETH_DEV_NO_OWNER, .name = ""};
 	int ret;
 
+	if (!rte_eth_is_valid_owner_id(owner_id))
+		return -EINVAL;
+
 	rte_eth_dev_shared_data_prepare();
 
 	rte_spinlock_lock(&rte_eth_dev_shared_data->ownership_lock);
-- 
2.18.0

Re: [dpdk-dev] [PATCH] ethdev: fix rte_eth_dev_owner_unset

[Matan Azrad]

Hi Stephen

From: Stephen Hemminger
> The rte_eth_dev_owner_unset function is unusable because it always
> returns -EINVAL. This is because the magic (unowned) value is flagged as not
> valid.
> 

It's OK to raise an error when you do unset for unowned device.
It means that unset owner should be called for owned device.

> Move the validation of owner into set and unset as separate calls.
> 
> Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
> ---
>  lib/librte_ethdev/rte_ethdev.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)
> 
> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
> index 4c320250589a..9398550a1189 100644
> --- a/lib/librte_ethdev/rte_ethdev.c
> +++ b/lib/librte_ethdev/rte_ethdev.c
> @@ -443,10 +443,6 @@ _rte_eth_dev_owner_set(const uint16_t port_id,
> const uint64_t old_owner_id,
>  		return -ENODEV;
>  	}
> 
> -	if (!rte_eth_is_valid_owner_id(new_owner->id) &&
> -	    !rte_eth_is_valid_owner_id(old_owner_id))
> -		return -EINVAL;
> -
>  	port_owner = &rte_eth_devices[port_id].data->owner;
>  	if (port_owner->id != old_owner_id) {
>  		RTE_ETHDEV_LOG(ERR,
> @@ -475,6 +471,9 @@ rte_eth_dev_owner_set(const uint16_t port_id,  {
>  	int ret;
> 
> +	if (!rte_eth_is_valid_owner_id(owner->id))
> +		return -EINVAL;
> +
>  	rte_eth_dev_shared_data_prepare();
> 
>  	rte_spinlock_lock(&rte_eth_dev_shared_data->ownership_lock);
> @@ -492,6 +491,9 @@ rte_eth_dev_owner_unset(const uint16_t port_id,
> const uint64_t owner_id)
>  			{.id = RTE_ETH_DEV_NO_OWNER, .name = ""};
>  	int ret;
> 
> +	if (!rte_eth_is_valid_owner_id(owner_id))
> +		return -EINVAL;
> +
>  	rte_eth_dev_shared_data_prepare();
> 
>  	rte_spinlock_lock(&rte_eth_dev_shared_data->ownership_lock);
> --
> 2.18.0

Re: [dpdk-dev] [PATCH] ethdev: fix rte_eth_dev_owner_unset

[Stephen Hemminger]

On Tue, 14 Aug 2018 05:52:20 +0000
Matan Azrad <matan@mellanox.com> wrote:

> Hi Stephen
> 
> From: Stephen Hemminger
> > The rte_eth_dev_owner_unset function is unusable because it always
> > returns -EINVAL. This is because the magic (unowned) value is flagged as not
> > valid.
> >   
> 
> It's OK to raise an error when you do unset for unowned device.
> It means that unset owner should be called for owned device.
> 

Original code was broken. The following would always fail.

	rte_eth_dev_owner_new(&owner.id);
	sprintf(owner.name, "example");
	rte_eth_dev_owner_set(port_id, &owner);
	rte_eth_dev_owner_unset(port_id, owner.id);

That is because of:
	rte_eth_dev_owner_unset(port_id, owner_id)
		_rte_eth_dev_owner_set(port_id, owner_id, &new_owner)  << new_owner.id == RTE_ETH_DEV_NO_OWNER (0)


	if (!rte_eth_is_valid_owner_id(new_owner->id) &&  << new_owner->id == RTE_ETH_DEV_NO_OWNER (which is flagged as invalid)
	    !rte_eth_is_valid_owner_id(old_owner_id))
		return -EINVAL;

The failsafe driver never checks the return value, and therefore doesn't see that it never clears ownership.

Re: [dpdk-dev] [PATCH] ethdev: fix rte_eth_dev_owner_unset

[Matan Azrad]

Hi Stephen

From: Stephen Hemminger
> On Tue, 14 Aug 2018 05:52:20 +0000
> Matan Azrad <matan@mellanox.com> wrote:
> 
> > Hi Stephen
> >
> > From: Stephen Hemminger
> > > The rte_eth_dev_owner_unset function is unusable because it always
> > > returns -EINVAL. This is because the magic (unowned) value is
> > > flagged as not valid.
> > >
> >
> > It's OK to raise an error when you do unset for unowned device.
> > It means that unset owner should be called for owned device.
> >
> 
> Original code was broken. The following would always fail.
> 
> 	rte_eth_dev_owner_new(&owner.id);
> 	sprintf(owner.name, "example");
> 	rte_eth_dev_owner_set(port_id, &owner);
> 	rte_eth_dev_owner_unset(port_id, owner.id);
> 
> That is because of:
> 	rte_eth_dev_owner_unset(port_id, owner_id)
> 		_rte_eth_dev_owner_set(port_id, owner_id, &new_owner)
> << new_owner.id == RTE_ETH_DEV_NO_OWNER (0)
> 
> 
> 	if (!rte_eth_is_valid_owner_id(new_owner->id) &&  <<
> new_owner->id == RTE_ETH_DEV_NO_OWNER (which is flagged as invalid)
> 	    !rte_eth_is_valid_owner_id(old_owner_id))
> 		return -EINVAL;
> 

But both should be invalid the new owner and the old owner(&&) to raise an EINVAL error.

In the aforementioned check above the old owner should be valid.

> The failsafe driver never checks the return value, and therefore doesn't see
> that it never clears ownership.

[dpdk-dev] [PATCH 0/2] ethdev: minor ownership changes

[Stephen Hemminger]

While debugging something else noticed that unset always logged
an error.

Stephen Hemminger (2):
  ethdev: silence error message on rte_eth_dev_owner_unset
  ethdev: make rte_eth_is_valid_owner_id return bool

 lib/librte_ethdev/rte_ethdev.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

-- 
2.18.0

[dpdk-dev] [PATCH 1/2] ethdev: silence error message on rte_eth_dev_owner_unset

[Stephen Hemminger]

The rte_eth_dev_owner_unset function always generates a log
message because the unset value for owner id is 0.

Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
---
 lib/librte_ethdev/rte_ethdev.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
index 4c320250589a..f09bf8bc8b01 100644
--- a/lib/librte_ethdev/rte_ethdev.c
+++ b/lib/librte_ethdev/rte_ethdev.c
@@ -393,11 +393,8 @@ static int
 rte_eth_is_valid_owner_id(uint64_t owner_id)
 {
 	if (owner_id == RTE_ETH_DEV_NO_OWNER ||
-	    rte_eth_dev_shared_data->next_owner_id <= owner_id) {
-		RTE_ETHDEV_LOG(ERR, "Invalid owner_id=%016"PRIx64"\n",
-			owner_id);
+	    rte_eth_dev_shared_data->next_owner_id <= owner_id)
 		return 0;
-	}
 	return 1;
 }
 
@@ -444,8 +441,12 @@ _rte_eth_dev_owner_set(const uint16_t port_id, const uint64_t old_owner_id,
 	}
 
 	if (!rte_eth_is_valid_owner_id(new_owner->id) &&
-	    !rte_eth_is_valid_owner_id(old_owner_id))
+	    !rte_eth_is_valid_owner_id(old_owner_id)) {
+		RTE_ETHDEV_LOG(ERR,
+			       "Invalid owner old_id=%016"PRIx64" new_id=%016"PRIx64"\n",
+			       old_owner_id, new_owner->id);
 		return -EINVAL;
+	}
 
 	port_owner = &rte_eth_devices[port_id].data->owner;
 	if (port_owner->id != old_owner_id) {
-- 
2.18.0

[dpdk-dev] [PATCH 2/2] ethdev: make rte_eth_is_valid_owner_id return bool

[Stephen Hemminger]

Function is boolean so use that.

Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
---
 lib/librte_ethdev/rte_ethdev.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
index f09bf8bc8b01..f0336736b7c1 100644
--- a/lib/librte_ethdev/rte_ethdev.c
+++ b/lib/librte_ethdev/rte_ethdev.c
@@ -389,13 +389,11 @@ rte_eth_dev_is_valid_port(uint16_t port_id)
 		return 1;
 }
 
-static int
+static bool
 rte_eth_is_valid_owner_id(uint64_t owner_id)
 {
-	if (owner_id == RTE_ETH_DEV_NO_OWNER ||
-	    rte_eth_dev_shared_data->next_owner_id <= owner_id)
-		return 0;
-	return 1;
+	return !(owner_id == RTE_ETH_DEV_NO_OWNER ||
+		 rte_eth_dev_shared_data->next_owner_id <= owner_id);
 }
 
 uint64_t
-- 
2.18.0

Re: [dpdk-dev] [PATCH 1/2] ethdev: silence error message on rte_eth_dev_owner_unset

[Matan Azrad]

Hi Stephen

From: Stephen Hemminger
> The rte_eth_dev_owner_unset function always generates a log message
> because the unset value for owner id is 0.
> 
> Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
> ---
>  lib/librte_ethdev/rte_ethdev.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
> index 4c320250589a..f09bf8bc8b01 100644
> --- a/lib/librte_ethdev/rte_ethdev.c
> +++ b/lib/librte_ethdev/rte_ethdev.c
> @@ -393,11 +393,8 @@ static int
>  rte_eth_is_valid_owner_id(uint64_t owner_id)  {
>  	if (owner_id == RTE_ETH_DEV_NO_OWNER ||
> -	    rte_eth_dev_shared_data->next_owner_id <= owner_id) {
> -		RTE_ETHDEV_LOG(ERR, "Invalid owner_id=%016"PRIx64"\n",
> -			owner_id);
> +	    rte_eth_dev_shared_data->next_owner_id <= owner_id)
>  		return 0;
> -	}
>  	return 1;
>  }

When you remove the log from this function you remove the error log also from
rte_eth_dev_owner_delete, so please add the missing error print there too.

> 
> @@ -444,8 +441,12 @@ _rte_eth_dev_owner_set(const uint16_t port_id,
> const uint64_t old_owner_id,
>  	}
> 
>  	if (!rte_eth_is_valid_owner_id(new_owner->id) &&
> -	    !rte_eth_is_valid_owner_id(old_owner_id))
> +	    !rte_eth_is_valid_owner_id(old_owner_id)) {
> +		RTE_ETHDEV_LOG(ERR,
> +			       "Invalid owner old_id=%016"PRIx64"
> new_id=%016"PRIx64"\n",
> +			       old_owner_id, new_owner->id);
>  		return -EINVAL;
> +	}
> 
>  	port_owner = &rte_eth_devices[port_id].data->owner;
>  	if (port_owner->id != old_owner_id) {
> --
> 2.18.0

Re: [dpdk-dev] [PATCH 2/2] ethdev: make rte_eth_is_valid_owner_id return bool

[Matan Azrad]

From: Stephen Hemminger
> Function is boolean so use that.

Ethdev is not using bool type, see also:
rte_eth_dev_is_valid_port
rte_eth_dev_is_removed
rte_eth_dev_pool_ops_supported

I think it should be a full solution to all.
 
> Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
> ---
>  lib/librte_ethdev/rte_ethdev.c | 8 +++-----
>  1 file changed, 3 insertions(+), 5 deletions(-)
> 
> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
> index f09bf8bc8b01..f0336736b7c1 100644
> --- a/lib/librte_ethdev/rte_ethdev.c
> +++ b/lib/librte_ethdev/rte_ethdev.c
> @@ -389,13 +389,11 @@ rte_eth_dev_is_valid_port(uint16_t port_id)
>  		return 1;
>  }
> 
> -static int
> +static bool
>  rte_eth_is_valid_owner_id(uint64_t owner_id)  {
> -	if (owner_id == RTE_ETH_DEV_NO_OWNER ||
> -	    rte_eth_dev_shared_data->next_owner_id <= owner_id)
> -		return 0;
> -	return 1;
> +	return !(owner_id == RTE_ETH_DEV_NO_OWNER ||
> +		 rte_eth_dev_shared_data->next_owner_id <= owner_id);
>  }
> 
>  uint64_t
> --
> 2.18.0

Re: [dpdk-dev] [PATCH 2/2] ethdev: make rte_eth_is_valid_owner_id return bool

[Stephen Hemminger]

On Tue, 21 Aug 2018 10:20:43 +0000
Matan Azrad <matan@mellanox.com> wrote:

> From: Stephen Hemminger
> > Function is boolean so use that.  
> 
> Ethdev is not using bool type, see also:
> rte_eth_dev_is_valid_port
> rte_eth_dev_is_removed
> rte_eth_dev_pool_ops_supported
> 
> I think it should be a full solution to all.
>  
> > Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>

I didn't want change type of visible (exported by ABI) functions.

[dpdk-dev] [PATCH v3] ethdev: silence error message on rte_eth_dev_owner_unset

[Stephen Hemminger]

From 74ad4c60262b1451a5a2fabf79a2df89c6c5373d Mon Sep 17 00:00:00 2001
From: Stephen Hemminger <stephen@networkplumber.org>
Date: Thu, 16 Aug 2018 15:37:14 -0700
Subject: [PATCH 1/5] ethdev: silence error message on rte_eth_dev_owner_unset

The rte_eth_dev_owner_unset function always generates a log
message because the unset value for owner id is 0.

Also, when rte_eth_dev_owner_delete is called with a valid
owner id, the log message should be at NOTICE not ERROR
severity.

Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
---
v3 - handle rte_eth_dev_owner_delete
     leave other cleanups to later patch series

 lib/librte_ethdev/rte_ethdev.c | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
index 4c320250589a..56dd9c2e83d9 100644
--- a/lib/librte_ethdev/rte_ethdev.c
+++ b/lib/librte_ethdev/rte_ethdev.c
@@ -393,11 +393,8 @@ static int
 rte_eth_is_valid_owner_id(uint64_t owner_id)
 {
 	if (owner_id == RTE_ETH_DEV_NO_OWNER ||
-	    rte_eth_dev_shared_data->next_owner_id <= owner_id) {
-		RTE_ETHDEV_LOG(ERR, "Invalid owner_id=%016"PRIx64"\n",
-			owner_id);
+	    rte_eth_dev_shared_data->next_owner_id <= owner_id)
 		return 0;
-	}
 	return 1;
 }
 
@@ -444,8 +441,12 @@ _rte_eth_dev_owner_set(const uint16_t port_id, const uint64_t old_owner_id,
 	}
 
 	if (!rte_eth_is_valid_owner_id(new_owner->id) &&
-	    !rte_eth_is_valid_owner_id(old_owner_id))
+	    !rte_eth_is_valid_owner_id(old_owner_id)) {
+		RTE_ETHDEV_LOG(ERR,
+			       "Invalid owner old_id=%016"PRIx64" new_id=%016"PRIx64"\n",
+			       old_owner_id, new_owner->id);
 		return -EINVAL;
+	}
 
 	port_owner = &rte_eth_devices[port_id].data->owner;
 	if (port_owner->id != old_owner_id) {
@@ -516,9 +517,13 @@ rte_eth_dev_owner_delete(const uint64_t owner_id)
 			if (rte_eth_devices[port_id].data->owner.id == owner_id)
 				memset(&rte_eth_devices[port_id].data->owner, 0,
 				       sizeof(struct rte_eth_dev_owner));
-		RTE_ETHDEV_LOG(ERR,
+		RTE_ETHDEV_LOG(NOTICE,
 			"All port owners owned by %016"PRIx64" identifier have removed\n",
 			owner_id);
+	} else {
+		RTE_ETHDEV_LOG(ERR,
+			       "Invalid owner id=%016"PRIx64"\n",
+			       owner_id);
 	}
 
 	rte_spinlock_unlock(&rte_eth_dev_shared_data->ownership_lock);
-- 
2.18.0

Re: [dpdk-dev] [PATCH 2/2] ethdev: make rte_eth_is_valid_owner_id return bool

[Matan Azrad]

Hi

From: Stephen Hemminger
> On Tue, 21 Aug 2018 10:20:43 +0000
> Matan Azrad <matan@mellanox.com> wrote:
> 
> > From: Stephen Hemminger
> > > Function is boolean so use that.
> >
> > Ethdev is not using bool type, see also:
> > rte_eth_dev_is_valid_port
> > rte_eth_dev_is_removed
> > rte_eth_dev_pool_ops_supported
> >
> > I think it should be a full solution to all.
> >
> > > Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
> 
> I didn't want change type of visible (exported by ABI) functions.
> 
Since ethdev now is not using bool type I think it's better not to change it only for this API.

Re: [dpdk-dev] [PATCH v3] ethdev: silence error message on rte_eth_dev_owner_unset

[Matan Azrad]

Hi

From: Stephen Hemminger
> From 74ad4c60262b1451a5a2fabf79a2df89c6c5373d Mon Sep 17 00:00:00 2001
> From: Stephen Hemminger <stephen@networkplumber.org>
> Date: Thu, 16 Aug 2018 15:37:14 -0700
> Subject: [PATCH 1/5] ethdev: silence error message on
> rte_eth_dev_owner_unset
> 
> The rte_eth_dev_owner_unset function always generates a log message
> because the unset value for owner id is 0.
> 
> Also, when rte_eth_dev_owner_delete is called with a valid owner id, the
> log message should be at NOTICE not ERROR severity.
> 
> Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
> ---
> v3 - handle rte_eth_dev_owner_delete
>      leave other cleanups to later patch series
> 
>  lib/librte_ethdev/rte_ethdev.c | 17 +++++++++++------
>  1 file changed, 11 insertions(+), 6 deletions(-)
> 
> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
> index 4c320250589a..56dd9c2e83d9 100644
> --- a/lib/librte_ethdev/rte_ethdev.c
> +++ b/lib/librte_ethdev/rte_ethdev.c
> @@ -393,11 +393,8 @@ static int
>  rte_eth_is_valid_owner_id(uint64_t owner_id)  {
>  	if (owner_id == RTE_ETH_DEV_NO_OWNER ||
> -	    rte_eth_dev_shared_data->next_owner_id <= owner_id) {
> -		RTE_ETHDEV_LOG(ERR, "Invalid owner_id=%016"PRIx64"\n",
> -			owner_id);
> +	    rte_eth_dev_shared_data->next_owner_id <= owner_id)
>  		return 0;
> -	}
>  	return 1;
>  }
> 
> @@ -444,8 +441,12 @@ _rte_eth_dev_owner_set(const uint16_t port_id,
> const uint64_t old_owner_id,
>  	}
> 
>  	if (!rte_eth_is_valid_owner_id(new_owner->id) &&
> -	    !rte_eth_is_valid_owner_id(old_owner_id))
> +	    !rte_eth_is_valid_owner_id(old_owner_id)) {
> +		RTE_ETHDEV_LOG(ERR,
> +			       "Invalid owner old_id=%016"PRIx64"
> new_id=%016"PRIx64"\n",
> +			       old_owner_id, new_owner->id);
>  		return -EINVAL;
> +	}
> 
>  	port_owner = &rte_eth_devices[port_id].data->owner;
>  	if (port_owner->id != old_owner_id) {
> @@ -516,9 +517,13 @@ rte_eth_dev_owner_delete(const uint64_t
> owner_id)
>  			if (rte_eth_devices[port_id].data->owner.id ==
> owner_id)
>  				memset(&rte_eth_devices[port_id].data-
> >owner, 0,
>  				       sizeof(struct rte_eth_dev_owner));
> -		RTE_ETHDEV_LOG(ERR,
> +		RTE_ETHDEV_LOG(NOTICE,
>  			"All port owners owned by %016"PRIx64" identifier
> have removed\n",
>  			owner_id);
> +	} else {
> +		RTE_ETHDEV_LOG(ERR,
> +			       "Invalid owner id=%016"PRIx64"\n",
> +			       owner_id);
>  	}
> 
>  	rte_spinlock_unlock(&rte_eth_dev_shared_data->ownership_lock);
> --
> 2.18.0
> 

I think the title should be:
ethdev: fix port ownership logs

while adding the fixes commits (at list 2 because of the NOTICE change)
and updating stable.

Besides that,
Acked-by: Matan Azrad <matan@mellanox.com>

Re: [dpdk-dev] [PATCH 2/2] ethdev: make rte_eth_is_valid_owner_id return bool

[Stephen Hemminger]

On Tue, 21 Aug 2018 15:48:19 +0000
Matan Azrad <matan@mellanox.com> wrote:

> Hi
> 
> From: Stephen Hemminger
> > On Tue, 21 Aug 2018 10:20:43 +0000
> > Matan Azrad <matan@mellanox.com> wrote:
> >   
> > > From: Stephen Hemminger  
> > > > Function is boolean so use that.  
> > >
> > > Ethdev is not using bool type, see also:
> > > rte_eth_dev_is_valid_port
> > > rte_eth_dev_is_removed
> > > rte_eth_dev_pool_ops_supported
> > >
> > > I think it should be a full solution to all.
> > >  
> > > > Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>  
> > 
> > I didn't want change type of visible (exported by ABI) functions.
> >   
> Since ethdev now is not using bool type I think it's better not to change it only for this API.

I hate to pick nits but there is already a bool usage in internal
function (static) in ethdev.


static bool
is_allocated(const struct rte_eth_dev *ethdev)
{
	return ethdev->data->name[0] != '\0';
}

Using bool functions doesn't really generate different code. It is is more
about using modern C conventions.

Re: [dpdk-dev] [PATCH] ethdev: fix rte_eth_dev_owner_unset

[Ferruh Yigit]

On 8/14/2018 8:46 PM, Matan Azrad wrote:
> Hi Stephen
> 
> From: Stephen Hemminger
>> On Tue, 14 Aug 2018 05:52:20 +0000
>> Matan Azrad <matan@mellanox.com> wrote:
>>
>>> Hi Stephen
>>>
>>> From: Stephen Hemminger
>>>> The rte_eth_dev_owner_unset function is unusable because it always
>>>> returns -EINVAL. This is because the magic (unowned) value is
>>>> flagged as not valid.
>>>>
>>>
>>> It's OK to raise an error when you do unset for unowned device.
>>> It means that unset owner should be called for owned device.
>>>
>>
>> Original code was broken. The following would always fail.
>>
>> 	rte_eth_dev_owner_new(&owner.id);
>> 	sprintf(owner.name, "example");
>> 	rte_eth_dev_owner_set(port_id, &owner);
>> 	rte_eth_dev_owner_unset(port_id, owner.id);
>>
>> That is because of:
>> 	rte_eth_dev_owner_unset(port_id, owner_id)
>> 		_rte_eth_dev_owner_set(port_id, owner_id, &new_owner)
>> << new_owner.id == RTE_ETH_DEV_NO_OWNER (0)
>>
>>
>> 	if (!rte_eth_is_valid_owner_id(new_owner->id) &&  <<
>> new_owner->id == RTE_ETH_DEV_NO_OWNER (which is flagged as invalid)
>> 	    !rte_eth_is_valid_owner_id(old_owner_id))
>> 		return -EINVAL;
>>
> 
> But both should be invalid the new owner and the old owner(&&) to raise an EINVAL error.
> 
> In the aforementioned check above the old owner should be valid.

It looks rte_eth_dev_owner_unset() works, updating patch as rejected.

Re: [dpdk-dev] [PATCH] ethdev: fix rte_eth_dev_owner_unset

[Stephen Hemminger]

On Fri, 24 Aug 2018 14:55:56 +0100
Ferruh Yigit <ferruh.yigit@intel.com> wrote:

> On 8/14/2018 8:46 PM, Matan Azrad wrote:
> > Hi Stephen
> > 
> > From: Stephen Hemminger  
> >> On Tue, 14 Aug 2018 05:52:20 +0000
> >> Matan Azrad <matan@mellanox.com> wrote:
> >>  
> >>> Hi Stephen
> >>>
> >>> From: Stephen Hemminger  
> >>>> The rte_eth_dev_owner_unset function is unusable because it always
> >>>> returns -EINVAL. This is because the magic (unowned) value is
> >>>> flagged as not valid.
> >>>>  
> >>>
> >>> It's OK to raise an error when you do unset for unowned device.
> >>> It means that unset owner should be called for owned device.
> >>>  
> >>
> >> Original code was broken. The following would always fail.
> >>
> >> 	rte_eth_dev_owner_new(&owner.id);
> >> 	sprintf(owner.name, "example");
> >> 	rte_eth_dev_owner_set(port_id, &owner);
> >> 	rte_eth_dev_owner_unset(port_id, owner.id);
> >>
> >> That is because of:
> >> 	rte_eth_dev_owner_unset(port_id, owner_id)
> >> 		_rte_eth_dev_owner_set(port_id, owner_id, &new_owner)
> >> << new_owner.id == RTE_ETH_DEV_NO_OWNER (0)
> >>
> >>
> >> 	if (!rte_eth_is_valid_owner_id(new_owner->id) &&  <<
> >> new_owner->id == RTE_ETH_DEV_NO_OWNER (which is flagged as invalid)
> >> 	    !rte_eth_is_valid_owner_id(old_owner_id))
> >> 		return -EINVAL;
> >>  
> > 
> > But both should be invalid the new owner and the old owner(&&) to raise an EINVAL error.
> > 
> > In the aforementioned check above the old owner should be valid.  
> 
> It looks rte_eth_dev_owner_unset() works, updating patch as rejected.
> 

The issue was the incorrect log message, addressed in later patches.

Re: [dpdk-dev] [PATCH 2/2] ethdev: make rte_eth_is_valid_owner_id return bool

[Matan Azrad]

From: Stephen Hemminger 
> On Tue, 21 Aug 2018 15:48:19 +0000
> Matan Azrad <matan@mellanox.com> wrote:
> 
> > Hi
> >
> > From: Stephen Hemminger
> > > On Tue, 21 Aug 2018 10:20:43 +0000
> > > Matan Azrad <matan@mellanox.com> wrote:
> > >
> > > > From: Stephen Hemminger
> > > > > Function is boolean so use that.
> > > >
> > > > Ethdev is not using bool type, see also:
> > > > rte_eth_dev_is_valid_port
> > > > rte_eth_dev_is_removed
> > > > rte_eth_dev_pool_ops_supported
> > > >
> > > > I think it should be a full solution to all.
> > > >
> > > > > Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
> > >
> > > I didn't want change type of visible (exported by ABI) functions.
> > >
> > Since ethdev now is not using bool type I think it's better not to change it
> only for this API.
> 
> I hate to pick nits but there is already a bool usage in internal function (static)
> in ethdev.
> 
> 
> static bool
> is_allocated(const struct rte_eth_dev *ethdev) {
> 	return ethdev->data->name[0] != '\0';
> }
> 
> Using bool functions doesn't really generate different code. It is is more
> about using modern C conventions.

Agree, but I think it should be the same API at least as  rte_eth_dev_is_valid_port, just for ethdev convention.

Let's give to the maintainer the decision.

Re: [dpdk-dev] [PATCH v3] ethdev: silence error message on rte_eth_dev_owner_unset

[Ferruh Yigit]

On 8/21/2018 4:53 PM, Matan Azrad wrote:
> Hi
> 
> From: Stephen Hemminger
>> From 74ad4c60262b1451a5a2fabf79a2df89c6c5373d Mon Sep 17 00:00:00 2001
>> From: Stephen Hemminger <stephen@networkplumber.org>
>> Date: Thu, 16 Aug 2018 15:37:14 -0700
>> Subject: [PATCH 1/5] ethdev: silence error message on
>> rte_eth_dev_owner_unset
>>
>> The rte_eth_dev_owner_unset function always generates a log message
>> because the unset value for owner id is 0.
>>
>> Also, when rte_eth_dev_owner_delete is called with a valid owner id, the
>> log message should be at NOTICE not ERROR severity.
>>
>> Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
>> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>

<...>

> I think the title should be:
> ethdev: fix port ownership logs
> 
> while adding the fixes commits (at list 2 because of the NOTICE change)
> and updating stable.
> 
> Besides that,
> Acked-by: Matan Azrad <matan@mellanox.com>

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

(Used suggested title.)