Hi Maxime, > -----Original Message----- > From: Maxime Coquelin <[email protected]> > Sent: Tuesday, October 20, 2020 1:34 AM > To: [email protected]; Xia, Chenbo <[email protected]>; [email protected] > Cc: Maxime Coquelin <[email protected]>; [email protected] > Subject: [PATCH 2/7] vhost: validate index in available entries API > > This patch validates the queue index parameter, in order > to ensure neither out-of-bound accesses nor NULL pointer > dereferencing happen. > > Fixes: a67f286a6596 ("vhost: export queue free entries") > Cc: [email protected] > > Signed-off-by: Maxime Coquelin <[email protected]> > --- > lib/librte_vhost/vhost.c | 5 +++++ > 1 file changed, 5 insertions(+) > > diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c > index 0c9ba3b3af..193dafc369 100644 > --- a/lib/librte_vhost/vhost.c > +++ b/lib/librte_vhost/vhost.c > @@ -1260,7 +1260,12 @@ rte_vhost_avail_entries(int vid, uint16_t queue_id) > if (!dev) > return 0; > > + if (queue_id >= VHOST_MAX_VRING) > + return 0; > + > vq = dev->virtqueue[queue_id]; > + if (!vq) > + return 0; > > rte_spinlock_lock(&vq->access_lock); > > -- > 2.26.2
Looking at the API again, I don't know if it is good to return 0 when there are no available entries or other errors. For this patch: Reviewed-by: Chenbo Xia <[email protected]>
