[PATCH net-next v2 0/2] Add an assert in napi_consume_skb()

[PATCH net-next v2 0/2] Add an assert in napi_consume_skb()

[Yunsheng Lin]

This patch introduces a lockdep_assert_in_softirq() interface and
uses it to assert the case when napi_consume_skb() is not called in
the softirq context.

Changelog:
V2: Use lockdep instead of one-off Kconfig knob

Yunsheng Lin (2):
  lockdep: Introduce in_softirq lockdep assert
  net: Use lockdep_assert_in_softirq() in napi_consume_skb()

 include/linux/lockdep.h | 7 +++++++
 net/core/skbuff.c       | 2 ++
 2 files changed, 9 insertions(+)

-- 
2.8.1

[PATCH net-next v2 1/2] lockdep: Introduce in_softirq lockdep assert

[Yunsheng Lin]

The current semantic for napi_consume_skb() is that caller need
to provide non-zero budget when calling from NAPI context, and
breaking this semantic will cause hard to debug problem, because
_kfree_skb_defer() need to run in atomic context in order to push
the skb to the particular cpu' napi_alloc_cache atomically.

So add the lockdep_assert_in_softirq() to assert when the running
context is not in_softirq, in_softirq means softirq is serving or
BH is disabled. Because the softirq context can be interrupted by
hard IRQ or NMI context, so lockdep_assert_in_softirq() need to
assert about hard IRQ or NMI context too.

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
---
 include/linux/lockdep.h | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
index f559487..f5e3d81 100644
--- a/include/linux/lockdep.h
+++ b/include/linux/lockdep.h
@@ -594,6 +594,12 @@ do {									\
 		      this_cpu_read(hardirqs_enabled)));		\
 } while (0)
 
+#define lockdep_assert_in_softirq()					\
+do {									\
+	WARN_ON_ONCE(__lockdep_enabled			&&		\
+		     (!in_softirq() || in_irq() || in_nmi()));		\
+} while (0)
+
 #else
 # define might_lock(lock) do { } while (0)
 # define might_lock_read(lock) do { } while (0)
@@ -605,6 +611,7 @@ do {									\
 
 # define lockdep_assert_preemption_enabled() do { } while (0)
 # define lockdep_assert_preemption_disabled() do { } while (0)
+# define lockdep_assert_in_softirq() do { } while (0)
 #endif
 
 #ifdef CONFIG_PROVE_RAW_LOCK_NESTING
-- 
2.8.1

Re: [PATCH net-next v2 1/2] lockdep: Introduce in_softirq lockdep assert

[Peter Zijlstra]

On Sat, Nov 21, 2020 at 11:06:15AM +0800, Yunsheng Lin wrote:
> The current semantic for napi_consume_skb() is that caller need
> to provide non-zero budget when calling from NAPI context, and
> breaking this semantic will cause hard to debug problem, because
> _kfree_skb_defer() need to run in atomic context in order to push
> the skb to the particular cpu' napi_alloc_cache atomically.
> 
> So add the lockdep_assert_in_softirq() to assert when the running
> context is not in_softirq, in_softirq means softirq is serving or
> BH is disabled. Because the softirq context can be interrupted by
> hard IRQ or NMI context, so lockdep_assert_in_softirq() need to
> assert about hard IRQ or NMI context too.
> 
> Suggested-by: Jakub Kicinski <kuba@kernel.org>
> Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
> ---
>  include/linux/lockdep.h | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h
> index f559487..f5e3d81 100644
> --- a/include/linux/lockdep.h
> +++ b/include/linux/lockdep.h
> @@ -594,6 +594,12 @@ do {									\
>  		      this_cpu_read(hardirqs_enabled)));		\
>  } while (0)

Due to in_softirq() having a deprication notice (due to it being
awefully ambiguous), could we have a nice big comment here that explains
in detail understandable to !network people (me) why this is actually
correct?

I'm not opposed to the thing, if that his what you need, it's fine, but
please put on a comment that explains that in_softirq() is ambiguous and
when you really do need it anyway.

> +#define lockdep_assert_in_softirq()					\
> +do {									\
> +	WARN_ON_ONCE(__lockdep_enabled			&&		\
> +		     (!in_softirq() || in_irq() || in_nmi()));		\
> +} while (0)
> +
>  #else
>  # define might_lock(lock) do { } while (0)
>  # define might_lock_read(lock) do { } while (0)
> @@ -605,6 +611,7 @@ do {									\
>  
>  # define lockdep_assert_preemption_enabled() do { } while (0)
>  # define lockdep_assert_preemption_disabled() do { } while (0)
> +# define lockdep_assert_in_softirq() do { } while (0)
>  #endif
>  
>  #ifdef CONFIG_PROVE_RAW_LOCK_NESTING
> -- 
> 2.8.1
> 

Re: [PATCH net-next v2 1/2] lockdep: Introduce in_softirq lockdep assert

[Jakub Kicinski]

On Mon, 23 Nov 2020 15:27:25 +0100 Peter Zijlstra wrote:
> On Sat, Nov 21, 2020 at 11:06:15AM +0800, Yunsheng Lin wrote:
> > The current semantic for napi_consume_skb() is that caller need
> > to provide non-zero budget when calling from NAPI context, and
> > breaking this semantic will cause hard to debug problem, because
> > _kfree_skb_defer() need to run in atomic context in order to push
> > the skb to the particular cpu' napi_alloc_cache atomically.
> > 
> > So add the lockdep_assert_in_softirq() to assert when the running
> > context is not in_softirq, in_softirq means softirq is serving or
> > BH is disabled. Because the softirq context can be interrupted by
> > hard IRQ or NMI context, so lockdep_assert_in_softirq() need to
> > assert about hard IRQ or NMI context too.

> Due to in_softirq() having a deprication notice (due to it being
> awefully ambiguous), could we have a nice big comment here that explains
> in detail understandable to !network people (me) why this is actually
> correct?
> 
> I'm not opposed to the thing, if that his what you need, it's fine, but
> please put on a comment that explains that in_softirq() is ambiguous and
> when you really do need it anyway.

One liner would be:

	* Acceptable for protecting per-CPU resources accessed from BH
	
We can add:

	* Much like in_softirq() - semantics are ambiguous, use carefully. *


IIUC we basically want to protect the nc array and counter here:

static inline void _kfree_skb_defer(struct sk_buff *skb)
{
	struct napi_alloc_cache *nc = this_cpu_ptr(&napi_alloc_cache);

	/* drop skb->head and call any destructors for packet */
	skb_release_all(skb);

	/* record skb to CPU local list */
	nc->skb_cache[nc->skb_count++] = skb;

#ifdef CONFIG_SLUB
	/* SLUB writes into objects when freeing */
	prefetchw(skb);
#endif

	/* flush skb_cache if it is filled */
	if (unlikely(nc->skb_count == NAPI_SKB_CACHE_SIZE)) {
		kmem_cache_free_bulk(skbuff_head_cache, NAPI_SKB_CACHE_SIZE,
				     nc->skb_cache);
		nc->skb_count = 0;
	}
}

> > +#define lockdep_assert_in_softirq()					\
> > +do {									\
> > +	WARN_ON_ONCE(__lockdep_enabled			&&		\
> > +		     (!in_softirq() || in_irq() || in_nmi()));		\
> > +} while (0)

Re: [PATCH net-next v2 1/2] lockdep: Introduce in_softirq lockdep assert

[Peter Zijlstra]

On Mon, Nov 23, 2020 at 12:12:59PM -0800, Jakub Kicinski wrote:
> One liner would be:
> 
> 	* Acceptable for protecting per-CPU resources accessed from BH
> 	
> We can add:
> 
> 	* Much like in_softirq() - semantics are ambiguous, use carefully. *
> 
> 
> IIUC we basically want to protect the nc array and counter here:

Works for me, thanks!

Re: [PATCH net-next v2 1/2] lockdep: Introduce in_softirq lockdep assert

[Yunsheng Lin]

On 2020/11/24 16:11, Peter Zijlstra wrote:
> On Mon, Nov 23, 2020 at 12:12:59PM -0800, Jakub Kicinski wrote:
>> One liner would be:
>>
>> 	* Acceptable for protecting per-CPU resources accessed from BH
>> 	
>> We can add:
>>
>> 	* Much like in_softirq() - semantics are ambiguous, use carefully. *
>>
>>
>> IIUC we basically want to protect the nc array and counter here:
> 
> Works for me, thanks!

Will add the above comment in v3.

Thanks.

> .
> 

[PATCH net-next v2 2/2] net: Use lockdep_assert_in_softirq() in napi_consume_skb()

[Yunsheng Lin]

Use napi_consume_skb() to assert the case when it is not called
in a atomic softirq context.

Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
---
 net/core/skbuff.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index ffe3dcc..effa19d 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -902,6 +902,8 @@ void napi_consume_skb(struct sk_buff *skb, int budget)
 		return;
 	}
 
+	lockdep_assert_in_softirq();
+
 	if (!skb_unref(skb))
 		return;
 
-- 
2.8.1