[dpdk-dev] [PATCH] ring: fix use after free in ring release

[wangyunjian]

> -----Original Message-----
> From: Honnappa Nagarahalli [mailto:Honnappa.Nagarahalli at arm.com]
> Sent: Thursday, April 20, 2023 5:44 AM
> To: wangyunjian <wangyunjian at huawei.com>; dev at dpdk.org
> Cc: konstantin.v.ananyev at yandex.ru; luyicai <luyicai at huawei.com>;
> stable at dpdk.org; nd <nd at arm.com>; nd <nd at arm.com>
> Subject: RE: [dpdk-dev] [PATCH] ring: fix use after free in ring release
> 
> <snip>
> 
> > >
> > > > -----Original Message-----
> > > > From: Yunjian Wang <wangyunjian at huawei.com>
> > > > Sent: Monday, April 17, 2023 8:12 AM
> > > > To: dev at dpdk.org
> > > > Cc: Honnappa Nagarahalli <Honnappa.Nagarahalli at arm.com>;
> > > > konstantin.v.ananyev at yandex.ru; luyicai at huawei.com; Yunjian Wang
> > > > <wangyunjian at huawei.com>; stable at dpdk.org
> > > > Subject: [dpdk-dev] [PATCH] ring: fix use after free in ring
> > > > release
> > > >
> > > > When using the ring to find out tailq entry, however it had been
> > > > freed by rte_memzone_free function. This change prevents that from
> > happening.
> > > I am unable to follow the problem you are describing.
> > > After the memzone for the ring is released, the contents of the
> > > memzone are not being used. I understand that the variable 'r' is
> > > being used, but that should not cause any issues.
> > >
> > > >
> > > > Fixes: 4e32101f9b01 ("ring: support freeing")
> > > > Cc: stable at dpdk.org
> > > >
> > > > Signed-off-by: Yunjian Wang <wangyunjian at huawei.com>
> > > > ---
> > > >  lib/ring/rte_ring.c | 11 +++++------
> > > >  1 file changed, 5 insertions(+), 6 deletions(-)
> > > >
> > > > diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c index
> > > > 8ed455043d..17d2d7f8a8 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;
> > > > -	}
> > > Why do we need to free the memzone later?
> >
> > 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.
> Thanks, understood
> 
> >
> > Thanks,
> > Yunjian
> > >
> > > > -
> > > >  	ring_list = RTE_TAILQ_CAST(rte_ring_tailq.head, rte_ring_list);
> > > >  	rte_mcfg_tailq_write_lock();
> > > >
> > > > @@ -349,7 +344,7 @@ rte_ring_free(struct rte_ring *r)
> > > >
> > > >  	if (te == NULL) {
> > > >  		rte_mcfg_tailq_write_unlock();
> > > > -		return;
> > > > +		goto free_memzone;
> We do not need this. If 'te == NULL' is true, then the ring was not found or
> possibly already freed.

OK

> 
> > > >  	}
> > > >
> > > >  	TAILQ_REMOVE(ring_list, te, next); @@ -357,6 +352,10 @@
> > > > rte_ring_free(struct rte_ring *r)
> We should free the memzone here while holding the lock

OK, You are right. I fix it on your suggestions.

https://patchwork.dpdk.org/project/dpdk/patch/c23b1135e1b0676ef7d82969b39a21df992d418f.1681972694.git.wangyunjian@huawei.com/

Thanks,
Yunjian

> 
> > > >  	rte_mcfg_tailq_write_unlock();
> > > >
> > > >  	rte_free(te);
> > > > +
> > > > +free_memzone:
> > > > +	if (rte_memzone_free(r->memzone) != 0)
> > > > +		RTE_LOG(ERR, RING, "Cannot free memory\n");
> > > >  }
> Should be moved up as mentioned above
> 
> > > >
> > > >  /* dump the status of the ring on the console */
> > > > --
> > > > 2.33.0