[v2] ring: fix use after free in ring release
Checks
Commit Message
After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
If rte_ring_lookup is called at this time, it will cause a use-after-free
problem. This change prevents that from happening.
Fixes: 4e32101f9b01 ("ring: support freeing")
Cc: stable@dpdk.org
Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
---
v2: update code suggested by Honnappa Nagarahalli
---
lib/ring/rte_ring.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
Comments
> -----Original Message-----
> From: Yunjian Wang <wangyunjian@huawei.com>
> Sent: Thursday, April 20, 2023 1:44 AM
> To: dev@dpdk.org
> Cc: Honnappa Nagarahalli <Honnappa.Nagarahalli@arm.com>;
> konstantin.v.ananyev@yandex.ru; luyicai@huawei.com; Yunjian Wang
> <wangyunjian@huawei.com>; stable@dpdk.org
> Subject: [dpdk-dev] [PATCH v2] ring: fix use after free in ring release
>
> After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
> If rte_ring_lookup is called at this time, it will cause a use-after-free problem.
> This change prevents that from happening.
>
> Fixes: 4e32101f9b01 ("ring: support freeing")
> Cc: stable@dpdk.org
>
> Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
This is incorrect, this is not a suggestion from me. Please remove this.
> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Other than the above, the patch looks fine.
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> ---
> v2: update code suggested by Honnappa Nagarahalli
> ---
> lib/ring/rte_ring.c | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c index
> 8ed455043d..2755323b8a 100644
> --- a/lib/ring/rte_ring.c
> +++ b/lib/ring/rte_ring.c
> @@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
> return;
> }
>
> - if (rte_memzone_free(r->memzone) != 0) {
> - RTE_LOG(ERR, RING, "Cannot free memory\n");
> - return;
> - }
> -
> ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
> rte_mcfg_tailq_write_lock();
>
> @@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
>
> TAILQ_REMOVE(ring_list, te, next);
>
> + if (rte_memzone_free(r->memzone) != 0)
> + RTE_LOG(ERR, RING, "Cannot free memory\n");
> +
> rte_mcfg_tailq_write_unlock();
>
> rte_free(te);
> --
> 2.33.0
> After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
> If rte_ring_lookup is called at this time, it will cause a use-after-free
> problem. This change prevents that from happening.
>
> Fixes: 4e32101f9b01 ("ring: support freeing")
> Cc: stable@dpdk.org
>
> Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
> ---
> v2: update code suggested by Honnappa Nagarahalli
> ---
> lib/ring/rte_ring.c | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c
> index 8ed455043d..2755323b8a 100644
> --- a/lib/ring/rte_ring.c
> +++ b/lib/ring/rte_ring.c
> @@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
> return;
> }
>
> - if (rte_memzone_free(r->memzone) != 0) {
> - RTE_LOG(ERR, RING, "Cannot free memory\n");
> - return;
> - }
> -
> ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
> rte_mcfg_tailq_write_lock();
>
> @@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
>
> TAILQ_REMOVE(ring_list, te, next);
>
> + if (rte_memzone_free(r->memzone) != 0)
> + RTE_LOG(ERR, RING, "Cannot free memory\n");
> +
I nit: I think it is a bit better to first release the lock and
then free the memzone.
Apart from that, LGTM.
Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
> rte_mcfg_tailq_write_unlock();
>
> rte_free(te);
> --
> 2.33.0
> -----Original Message-----
> From: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
> Sent: Monday, May 1, 2023 7:32 AM
> To: wangyunjian@huawei.com
> Cc: dev@dpdk.org; Honnappa Nagarahalli
> <Honnappa.Nagarahalli@arm.com>; konstantin.v.ananyev@yandex.ru;
> luyicai@huawei.com; stable@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2] ring: fix use after free in ring release
>
>
>
> > After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
> > If rte_ring_lookup is called at this time, it will cause a
> > use-after-free problem. This change prevents that from happening.
> >
> > Fixes: 4e32101f9b01 ("ring: support freeing")
> > Cc: stable@dpdk.org
> >
> > Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> > Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
> > ---
> > v2: update code suggested by Honnappa Nagarahalli
> > ---
> > lib/ring/rte_ring.c | 8 +++-----
> > 1 file changed, 3 insertions(+), 5 deletions(-)
> >
> > diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c index
> > 8ed455043d..2755323b8a 100644
> > --- a/lib/ring/rte_ring.c
> > +++ b/lib/ring/rte_ring.c
> > @@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
> > return;
> > }
> >
> > - if (rte_memzone_free(r->memzone) != 0) {
> > - RTE_LOG(ERR, RING, "Cannot free memory\n");
> > - return;
> > - }
> > -
> > ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
> > rte_mcfg_tailq_write_lock();
> >
> > @@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
> >
> > TAILQ_REMOVE(ring_list, te, next);
> >
> > + if (rte_memzone_free(r->memzone) != 0)
> > + RTE_LOG(ERR, RING, "Cannot free memory\n");
> > +
>
> I nit: I think it is a bit better to first release the lock and then free the
> memzone.
I think both of our suggestions are contradictory. Any reason why you want to free outside the locked region?
I thought, since it belongs to the ring being freed, it makes sense to free it while holding the lock to avoid any race conditions (though, I have not checked what those are).
> Apart from that, LGTM.
> Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
>
> > rte_mcfg_tailq_write_unlock();
> >
> > rte_free(te);
> > --
> > 2.33.0
>
01/05/2023 20:06, Honnappa Nagarahalli пишет:
>
>
>> -----Original Message-----
>> From: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
>> Sent: Monday, May 1, 2023 7:32 AM
>> To: wangyunjian@huawei.com
>> Cc: dev@dpdk.org; Honnappa Nagarahalli
>> <Honnappa.Nagarahalli@arm.com>; konstantin.v.ananyev@yandex.ru;
>> luyicai@huawei.com; stable@dpdk.org
>> Subject: Re: [dpdk-dev] [PATCH v2] ring: fix use after free in ring release
>>
>>
>>
>>> After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
>>> If rte_ring_lookup is called at this time, it will cause a
>>> use-after-free problem. This change prevents that from happening.
>>>
>>> Fixes: 4e32101f9b01 ("ring: support freeing")
>>> Cc: stable@dpdk.org
>>>
>>> Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
>>> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
>>> ---
>>> v2: update code suggested by Honnappa Nagarahalli
>>> ---
>>> lib/ring/rte_ring.c | 8 +++-----
>>> 1 file changed, 3 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c index
>>> 8ed455043d..2755323b8a 100644
>>> --- a/lib/ring/rte_ring.c
>>> +++ b/lib/ring/rte_ring.c
>>> @@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
>>> return;
>>> }
>>>
>>> - if (rte_memzone_free(r->memzone) != 0) {
>>> - RTE_LOG(ERR, RING, "Cannot free memory\n");
>>> - return;
>>> - }
>>> -
>>> ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
>>> rte_mcfg_tailq_write_lock();
>>>
>>> @@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
>>>
>>> TAILQ_REMOVE(ring_list, te, next);
>>>
>>> + if (rte_memzone_free(r->memzone) != 0)
>>> + RTE_LOG(ERR, RING, "Cannot free memory\n");
>>> +
>>
>> I nit: I think it is a bit better to first release the lock and then free the
>> memzone.
> I think both of our suggestions are contradictory. Any reason why you want to free outside the locked region?
Don't know what you mean by 'both suggestions' here.
I think I gave only one - move memzone_free() after tailq_write_unlock().
To be more precise:
1) rte_mcfg_tailq_write_lock();
...
2) TAILQ_REMOVE(...);
3) rte_mcfg_tailq_write_unlock();
4) rte_memzone_free(r->memzone);
As I remember, memzones are protected by their own lock (mlock), so
we don't need to hold qlock to free a memzone, after ring was already
removed from the ring_list.
>
> I thought, since it belongs to the ring being freed, it makes sense to free it while holding the lock to avoid any race conditions (though, I have not checked what those are).
As I understand, it is ok with current design to grab mlock while
holding qlock.
So, there is nothing wrong with current patch,
I just think that in that case it is excessive, and can be safely avoided.
>
>> Apart from that, LGTM.
>> Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
>>
>>> rte_mcfg_tailq_write_unlock();
>>>
>>> rte_free(te);
>>> --
>>> 2.33.0
>>
>
<snip>
> >>
> >>
> >>
> >>> After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
> >>> If rte_ring_lookup is called at this time, it will cause a
> >>> use-after-free problem. This change prevents that from happening.
> >>>
> >>> Fixes: 4e32101f9b01 ("ring: support freeing")
> >>> Cc: stable@dpdk.org
> >>>
> >>> Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> >>> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
> >>> ---
> >>> v2: update code suggested by Honnappa Nagarahalli
> >>> ---
> >>> lib/ring/rte_ring.c | 8 +++-----
> >>> 1 file changed, 3 insertions(+), 5 deletions(-)
> >>>
> >>> diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c index
> >>> 8ed455043d..2755323b8a 100644
> >>> --- a/lib/ring/rte_ring.c
> >>> +++ b/lib/ring/rte_ring.c
> >>> @@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
> >>> return;
> >>> }
> >>>
> >>> - if (rte_memzone_free(r->memzone) != 0) {
> >>> - RTE_LOG(ERR, RING, "Cannot free memory\n");
> >>> - return;
> >>> - }
> >>> -
> >>> ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
> >>> rte_mcfg_tailq_write_lock();
> >>>
> >>> @@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
> >>>
> >>> TAILQ_REMOVE(ring_list, te, next);
> >>>
> >>> + if (rte_memzone_free(r->memzone) != 0)
> >>> + RTE_LOG(ERR, RING, "Cannot free memory\n");
> >>> +
> >>
> >> I nit: I think it is a bit better to first release the lock and then
> >> free the memzone.
> > I think both of our suggestions are contradictory. Any reason why you want
> to free outside the locked region?
>
>
> Don't know what you mean by 'both suggestions' here.
I wrote 'both of our suggestions'. Essentially, in v1, freeing the memzone was outside of the lock. I suggested to move it inside and you are suggesting to move it inside.
> I think I gave only one - move memzone_free() after tailq_write_unlock().
> To be more precise:
> 1) rte_mcfg_tailq_write_lock();
> ...
> 2) TAILQ_REMOVE(...);
> 3) rte_mcfg_tailq_write_unlock();
> 4) rte_memzone_free(r->memzone);
>
> As I remember, memzones are protected by their own lock (mlock), so we
> don't need to hold qlock to free a memzone, after ring was already removed
> from the ring_list.
>
> >
> > I thought, since it belongs to the ring being freed, it makes sense to free it
> while holding the lock to avoid any race conditions (though, I have not
> checked what those are).
>
>
> As I understand, it is ok with current design to grab mlock while holding qlock.
> So, there is nothing wrong with current patch, I just think that in that case it is
> excessive, and can be safely avoided.
>
> >
> >> Apart from that, LGTM.
> >> Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
> >>
> >>> rte_mcfg_tailq_write_unlock();
> >>>
> >>> rte_free(te);
> >>> --
> >>> 2.33.0
> >>
> >
03/05/2023 06:44, Honnappa Nagarahalli пишет:
> <snip>
>
>>>>
>>>>
>>>>
>>>>> After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
>>>>> If rte_ring_lookup is called at this time, it will cause a
>>>>> use-after-free problem. This change prevents that from happening.
>>>>>
>>>>> Fixes: 4e32101f9b01 ("ring: support freeing")
>>>>> Cc: stable@dpdk.org
>>>>>
>>>>> Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
>>>>> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
>>>>> ---
>>>>> v2: update code suggested by Honnappa Nagarahalli
>>>>> ---
>>>>> lib/ring/rte_ring.c | 8 +++-----
>>>>> 1 file changed, 3 insertions(+), 5 deletions(-)
>>>>>
>>>>> diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c index
>>>>> 8ed455043d..2755323b8a 100644
>>>>> --- a/lib/ring/rte_ring.c
>>>>> +++ b/lib/ring/rte_ring.c
>>>>> @@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
>>>>> return;
>>>>> }
>>>>>
>>>>> - if (rte_memzone_free(r->memzone) != 0) {
>>>>> - RTE_LOG(ERR, RING, "Cannot free memory\n");
>>>>> - return;
>>>>> - }
>>>>> -
>>>>> ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
>>>>> rte_mcfg_tailq_write_lock();
>>>>>
>>>>> @@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
>>>>>
>>>>> TAILQ_REMOVE(ring_list, te, next);
>>>>>
>>>>> + if (rte_memzone_free(r->memzone) != 0)
>>>>> + RTE_LOG(ERR, RING, "Cannot free memory\n");
>>>>> +
>>>>
>>>> I nit: I think it is a bit better to first release the lock and then
>>>> free the memzone.
>>> I think both of our suggestions are contradictory. Any reason why you want
>> to free outside the locked region?
>>
>>
>> Don't know what you mean by 'both suggestions' here.
> I wrote 'both of our suggestions'. Essentially, in v1, freeing the memzone was outside of the lock. I suggested to move it inside and you are suggesting to move it inside.
Ah ok, I missed v1 and your comments for it.
As I said before, I don't think that we need to hold qlock here
while calling mmezone_free().
Though I don't see any harm with it either.
I'd personally would move memzone_free() after releasing qlock,
but if you guys prefer to keep it as it is - I wouldn't insist.
>
>> I think I gave only one - move memzone_free() after tailq_write_unlock().
>> To be more precise:
>> 1) rte_mcfg_tailq_write_lock();
>> ...
>> 2) TAILQ_REMOVE(...);
>> 3) rte_mcfg_tailq_write_unlock();
>> 4) rte_memzone_free(r->memzone);
>>
>> As I remember, memzones are protected by their own lock (mlock), so we
>> don't need to hold qlock to free a memzone, after ring was already removed
>> from the ring_list.
>>
>>>
>>> I thought, since it belongs to the ring being freed, it makes sense to free it
>> while holding the lock to avoid any race conditions (though, I have not
>> checked what those are).
>>
>>
>> As I understand, it is ok with current design to grab mlock while holding qlock.
>> So, there is nothing wrong with current patch, I just think that in that case it is
>> excessive, and can be safely avoided.
>>
>>>
>>>> Apart from that, LGTM.
>>>> Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
>>>>
>>>>> rte_mcfg_tailq_write_unlock();
>>>>>
>>>>> rte_free(te);
>>>>> --
>>>>> 2.33.0
>>>>
>>>
>
<snip>
> >
> >>>>
> >>>>
> >>>>
> >>>>> After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
> >>>>> If rte_ring_lookup is called at this time, it will cause a
> >>>>> use-after-free problem. This change prevents that from happening.
> >>>>>
> >>>>> Fixes: 4e32101f9b01 ("ring: support freeing")
> >>>>> Cc: stable@dpdk.org
> >>>>>
> >>>>> Suggested-by: Honnappa Nagarahalli
> <honnappa.nagarahalli@arm.com>
> >>>>> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
> >>>>> ---
> >>>>> v2: update code suggested by Honnappa Nagarahalli
> >>>>> ---
> >>>>> lib/ring/rte_ring.c | 8 +++-----
> >>>>> 1 file changed, 3 insertions(+), 5 deletions(-)
> >>>>>
> >>>>> diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c index
> >>>>> 8ed455043d..2755323b8a 100644
> >>>>> --- a/lib/ring/rte_ring.c
> >>>>> +++ b/lib/ring/rte_ring.c
> >>>>> @@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
> >>>>> return;
> >>>>> }
> >>>>>
> >>>>> - if (rte_memzone_free(r->memzone) != 0) {
> >>>>> - RTE_LOG(ERR, RING, "Cannot free memory\n");
> >>>>> - return;
> >>>>> - }
> >>>>> -
> >>>>> ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head,
> rte_ring_list);
> >>>>> rte_mcfg_tailq_write_lock();
> >>>>>
> >>>>> @@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
> >>>>>
> >>>>> TAILQ_REMOVE(ring_list, te, next);
> >>>>>
> >>>>> + if (rte_memzone_free(r->memzone) != 0)
> >>>>> + RTE_LOG(ERR, RING, "Cannot free memory\n");
> >>>>> +
> >>>>
> >>>> I nit: I think it is a bit better to first release the lock and
> >>>> then free the memzone.
> >>> I think both of our suggestions are contradictory. Any reason why
> >>> you want
> >> to free outside the locked region?
> >>
> >>
> >> Don't know what you mean by 'both suggestions' here.
> > I wrote 'both of our suggestions'. Essentially, in v1, freeing the memzone
> was outside of the lock. I suggested to move it inside and you are suggesting
> to move it inside.
>
>
> Ah ok, I missed v1 and your comments for it.
> As I said before, I don't think that we need to hold qlock here while calling
> mmezone_free().
> Though I don't see any harm with it either.
> I'd personally would move memzone_free() after releasing qlock, but if you
> guys prefer to keep it as it is - I wouldn't insist.
I looked at other libraries, stack library is the closest. Stack library frees the memzone outside the lock. I think we should keep it consistent.
I am fine to move the free outside the lock.
>
> >
> >> I think I gave only one - move memzone_free() after tailq_write_unlock().
> >> To be more precise:
> >> 1) rte_mcfg_tailq_write_lock();
> >> ...
> >> 2) TAILQ_REMOVE(...);
> >> 3) rte_mcfg_tailq_write_unlock();
> >> 4) rte_memzone_free(r->memzone);
> >>
> >> As I remember, memzones are protected by their own lock (mlock), so
> >> we don't need to hold qlock to free a memzone, after ring was already
> >> removed from the ring_list.
> >>
> >>>
> >>> I thought, since it belongs to the ring being freed, it makes sense
> >>> to free it
> >> while holding the lock to avoid any race conditions (though, I have
> >> not checked what those are).
> >>
> >>
> >> As I understand, it is ok with current design to grab mlock while holding
> qlock.
> >> So, there is nothing wrong with current patch, I just think that in
> >> that case it is excessive, and can be safely avoided.
> >>
> >>>
> >>>> Apart from that, LGTM.
> >>>> Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
> >>>>
> >>>>> rte_mcfg_tailq_write_unlock();
> >>>>>
> >>>>> rte_free(te);
> >>>>> --
> >>>>> 2.33.0
> >>>>
> >>>
> >
> -----Original Message-----
> From: Honnappa Nagarahalli [mailto:Honnappa.Nagarahalli@arm.com]
> Sent: Thursday, May 4, 2023 7:46 AM
> To: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>; wangyunjian
> <wangyunjian@huawei.com>
> Cc: dev@dpdk.org; luyicai <luyicai@huawei.com>; stable@dpdk.org; nd
> <nd@arm.com>; nd <nd@arm.com>
> Subject: RE: [dpdk-dev] [PATCH v2] ring: fix use after free in ring release
>
> <snip>
>
> > >
> > >>>>
> > >>>>
> > >>>>
> > >>>>> After the memzone is freed, it is not removed from the 'rte_ring_tailq'.
> > >>>>> If rte_ring_lookup is called at this time, it will cause a
> > >>>>> use-after-free problem. This change prevents that from happening.
> > >>>>>
> > >>>>> Fixes: 4e32101f9b01 ("ring: support freeing")
> > >>>>> Cc: stable@dpdk.org
> > >>>>>
> > >>>>> Suggested-by: Honnappa Nagarahalli
> > <honnappa.nagarahalli@arm.com>
> > >>>>> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
> > >>>>> ---
> > >>>>> v2: update code suggested by Honnappa Nagarahalli
> > >>>>> ---
> > >>>>> lib/ring/rte_ring.c | 8 +++-----
> > >>>>> 1 file changed, 3 insertions(+), 5 deletions(-)
> > >>>>>
> > >>>>> diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c index
> > >>>>> 8ed455043d..2755323b8a 100644
> > >>>>> --- a/lib/ring/rte_ring.c
> > >>>>> +++ b/lib/ring/rte_ring.c
> > >>>>> @@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
> > >>>>> return;
> > >>>>> }
> > >>>>>
> > >>>>> - if (rte_memzone_free(r->memzone) != 0) {
> > >>>>> - RTE_LOG(ERR, RING, "Cannot free memory\n");
> > >>>>> - return;
> > >>>>> - }
> > >>>>> -
> > >>>>> ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head,
> > rte_ring_list);
> > >>>>> rte_mcfg_tailq_write_lock();
> > >>>>>
> > >>>>> @@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
> > >>>>>
> > >>>>> TAILQ_REMOVE(ring_list, te, next);
> > >>>>>
> > >>>>> + if (rte_memzone_free(r->memzone) != 0)
> > >>>>> + RTE_LOG(ERR, RING, "Cannot free memory\n");
> > >>>>> +
> > >>>>
> > >>>> I nit: I think it is a bit better to first release the lock and
> > >>>> then free the memzone.
> > >>> I think both of our suggestions are contradictory. Any reason why
> > >>> you want
> > >> to free outside the locked region?
> > >>
> > >>
> > >> Don't know what you mean by 'both suggestions' here.
> > > I wrote 'both of our suggestions'. Essentially, in v1, freeing the
> > > memzone
> > was outside of the lock. I suggested to move it inside and you are
> > suggesting to move it inside.
> >
> >
> > Ah ok, I missed v1 and your comments for it.
> > As I said before, I don't think that we need to hold qlock here while
> > calling mmezone_free().
> > Though I don't see any harm with it either.
> > I'd personally would move memzone_free() after releasing qlock, but if
> > you guys prefer to keep it as it is - I wouldn't insist.
> I looked at other libraries, stack library is the closest. Stack library frees the
> memzone outside the lock. I think we should keep it consistent.
> I am fine to move the free outside the lock.
Thanks, Konstantin Ananyev and Honnappa Nagarahalli.
I will update the patch in v3 and move the free outside the lock.
>
> >
> > >
> > >> I think I gave only one - move memzone_free() after tailq_write_unlock().
> > >> To be more precise:
> > >> 1) rte_mcfg_tailq_write_lock();
> > >> ...
> > >> 2) TAILQ_REMOVE(...);
> > >> 3) rte_mcfg_tailq_write_unlock();
> > >> 4) rte_memzone_free(r->memzone);
> > >>
> > >> As I remember, memzones are protected by their own lock (mlock), so
> > >> we don't need to hold qlock to free a memzone, after ring was
> > >> already removed from the ring_list.
> > >>
> > >>>
> > >>> I thought, since it belongs to the ring being freed, it makes
> > >>> sense to free it
> > >> while holding the lock to avoid any race conditions (though, I have
> > >> not checked what those are).
> > >>
> > >>
> > >> As I understand, it is ok with current design to grab mlock while
> > >> holding
> > qlock.
> > >> So, there is nothing wrong with current patch, I just think that in
> > >> that case it is excessive, and can be safely avoided.
> > >>
> > >>>
> > >>>> Apart from that, LGTM.
> > >>>> Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
> > >>>>
> > >>>>> rte_mcfg_tailq_write_unlock();
> > >>>>>
> > >>>>> rte_free(te);
> > >>>>> --
> > >>>>> 2.33.0
> > >>>>
> > >>>
> > >
@@ -333,11 +333,6 @@ rte_ring_free(struct rte_ring *r)
return;
}
- if (rte_memzone_free(r->memzone) != 0) {
- RTE_LOG(ERR, RING, "Cannot free memory\n");
- return;
- }
-
ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
rte_mcfg_tailq_write_lock();
@@ -354,6 +349,9 @@ rte_ring_free(struct rte_ring *r)
TAILQ_REMOVE(ring_list, te, next);
+ if (rte_memzone_free(r->memzone) != 0)
+ RTE_LOG(ERR, RING, "Cannot free memory\n");
+
rte_mcfg_tailq_write_unlock();
rte_free(te);