[dpdk-dev] [PATCH] net/failsafe: safer subdev iterator

[dpdk-dev] [PATCH] net/failsafe: safer subdev iterator

[Gaetan Rivet]

The sub_device iterator macro should follow the general gist of the
tailq API for an easier understanding and safer use.

Once the loop has finished, the iterator should be set to NULL.
If no sub_device was iterated upon, the iterator should still be NULL.

Signed-off-by: Gaetan Rivet <gaetan.rivet@6wind.com>
---
 drivers/net/failsafe/failsafe_private.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/failsafe/failsafe_private.h b/drivers/net/failsafe/failsafe_private.h
index 0361cf4..6fa8de5 100644
--- a/drivers/net/failsafe/failsafe_private.h
+++ b/drivers/net/failsafe/failsafe_private.h
@@ -222,6 +222,7 @@ extern int mac_from_arg;
  */
 #define FOREACH_SUBDEV_STATE(s, i, dev, state)				\
 	for (i = fs_find_next((dev), 0, state);				\
+	     ((s = NULL) == NULL) &&					\
 	     i < PRIV(dev)->subs_tail && (s = &PRIV(dev)->subs[i]);	\
 	     i = fs_find_next((dev), i + 1, state))
 
-- 
2.1.4

Re: [dpdk-dev] [PATCH] net/failsafe: safer subdev iterator

[Stephen Hemminger]

On Thu, 17 Aug 2017 17:52:09 +0200
Gaetan Rivet <gaetan.rivet@6wind.com> wrote:

> The sub_device iterator macro should follow the general gist of the
> tailq API for an easier understanding and safer use.
> 
> Once the loop has finished, the iterator should be set to NULL.
> If no sub_device was iterated upon, the iterator should still be NULL.
> 
> Signed-off-by: Gaetan Rivet <gaetan.rivet@6wind.com>
> ---
>  drivers/net/failsafe/failsafe_private.h | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/net/failsafe/failsafe_private.h b/drivers/net/failsafe/failsafe_private.h
> index 0361cf4..6fa8de5 100644
> --- a/drivers/net/failsafe/failsafe_private.h
> +++ b/drivers/net/failsafe/failsafe_private.h
> @@ -222,6 +222,7 @@ extern int mac_from_arg;
>   */
>  #define FOREACH_SUBDEV_STATE(s, i, dev, state)				\
>  	for (i = fs_find_next((dev), 0, state);				\
> +	     ((s = NULL) == NULL) &&					\
I assume you are trying to do assignment inside an expression and  just expect that
to always to be true and go onto next part of assignment. This is a convulted
way to put code in the loop iterator.

But this macro is way too complex. Please break it up into inline functions or
figure out how to simplify the logic better.

Re: [dpdk-dev] [PATCH] net/failsafe: safer subdev iterator

[Gaëtan Rivet]

On Thu, Aug 17, 2017 at 05:39:29PM -0700, Stephen Hemminger wrote:
> On Thu, 17 Aug 2017 17:52:09 +0200
> Gaetan Rivet <gaetan.rivet@6wind.com> wrote:
> 
> > The sub_device iterator macro should follow the general gist of the
> > tailq API for an easier understanding and safer use.
> > 
> > Once the loop has finished, the iterator should be set to NULL.
> > If no sub_device was iterated upon, the iterator should still be NULL.
> > 
> > Signed-off-by: Gaetan Rivet <gaetan.rivet@6wind.com>
> > ---
> >  drivers/net/failsafe/failsafe_private.h | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/drivers/net/failsafe/failsafe_private.h b/drivers/net/failsafe/failsafe_private.h
> > index 0361cf4..6fa8de5 100644
> > --- a/drivers/net/failsafe/failsafe_private.h
> > +++ b/drivers/net/failsafe/failsafe_private.h
> > @@ -222,6 +222,7 @@ extern int mac_from_arg;
> >   */
> >  #define FOREACH_SUBDEV_STATE(s, i, dev, state)				\
> >  	for (i = fs_find_next((dev), 0, state);				\
> > +	     ((s = NULL) == NULL) &&					\
> I assume you are trying to do assignment inside an expression and  just expect that
> to always to be true and go onto next part of assignment. This is a convulted
> way to put code in the loop iterator.
> 
> But this macro is way too complex. Please break it up into inline functions or
> figure out how to simplify the logic better.

I think you are right, thanks for the wake-up call.

-- 
Gaëtan Rivet
6WIND

[dpdk-dev] [PATCH v2] net/failsafe: safer subdev iterator

[Gaetan Rivet]

The sub_device iterator macro should follow the general gist of the
tailq API for an easier understanding and safer use.

Once the loop has finished, the iterator should be set to NULL.
If no sub_device was iterated upon, the iterator should still be NULL.

Signed-off-by: Gaetan Rivet <gaetan.rivet@6wind.com>
---

v2:

  - Rewrite fs_find_next to simplify the macro using it.

 drivers/net/failsafe/failsafe_private.h | 32 ++++++++++++++++++++------------
 1 file changed, 20 insertions(+), 12 deletions(-)

diff --git a/drivers/net/failsafe/failsafe_private.h b/drivers/net/failsafe/failsafe_private.h
index 0361cf4..19d36ab 100644
--- a/drivers/net/failsafe/failsafe_private.h
+++ b/drivers/net/failsafe/failsafe_private.h
@@ -220,10 +220,10 @@ extern int mac_from_arg;
  * dev:   (struct rte_eth_dev *), fail-safe ethdev
  * state: (enum dev_state), minimum acceptable device state
  */
-#define FOREACH_SUBDEV_STATE(s, i, dev, state)				\
-	for (i = fs_find_next((dev), 0, state);				\
-	     i < PRIV(dev)->subs_tail && (s = &PRIV(dev)->subs[i]);	\
-	     i = fs_find_next((dev), i + 1, state))
+#define FOREACH_SUBDEV_STATE(s, i, dev, state)		\
+	for (s = fs_find_next((dev), 0, state, &i);	\
+	     s != NULL;					\
+	     s = fs_find_next((dev), i + 1, state, &i))
 
 /**
  * Iterator construct over fail-safe sub-devices:
@@ -294,18 +294,26 @@ extern int mac_from_arg;
 
 /* inlined functions */
 
-static inline uint8_t
-fs_find_next(struct rte_eth_dev *dev, uint8_t sid,
-		enum dev_state min_state)
+static inline struct sub_device *
+fs_find_next(struct rte_eth_dev *dev,
+	     uint8_t sid,
+	     enum dev_state min_state,
+	     uint8_t *sid_out)
 {
-	while (sid < PRIV(dev)->subs_tail) {
-		if (PRIV(dev)->subs[sid].state >= min_state)
+	struct sub_device *subs;
+	uint8_t tail;
+
+	subs = PRIV(dev)->subs;
+	tail = PRIV(dev)->subs_tail;
+	while (sid < tail) {
+		if (subs[sid].state >= min_state)
 			break;
 		sid++;
 	}
-	if (sid >= PRIV(dev)->subs_tail)
-		return PRIV(dev)->subs_tail;
-	return sid;
+	*sid_out = sid;
+	if (sid >= tail)
+		return NULL;
+	return &subs[sid];
 }
 
 /*
-- 
2.1.4

Re: [dpdk-dev] [PATCH v2] net/failsafe: safer subdev iterator

[Ferruh Yigit]

On 8/23/2017 2:05 PM, Gaetan Rivet wrote:
> The sub_device iterator macro should follow the general gist of the
> tailq API for an easier understanding and safer use.
> 
> Once the loop has finished, the iterator should be set to NULL.
> If no sub_device was iterated upon, the iterator should still be NULL.
> 
> Signed-off-by: Gaetan Rivet <gaetan.rivet@6wind.com>

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