On Thu, Oct 11, 2018 at 11:46:19AM +0000, Jason Gunthorpe wrote:
> On Thu, Oct 11, 2018 at 08:14:28AM +0300, Leon Romanovsky wrote:
> > On Wed, Oct 10, 2018 at 09:39:05PM +0000, Parav Pandit wrote:
> > >
> > >
> > > > From: linux-rdma-owner@xxxxxxxxxxxxxxx <linux-rdma-
> > > > owner@xxxxxxxxxxxxxxx> On Behalf Of Leon Romanovsky
> > > > Sent: Wednesday, October 10, 2018 1:19 AM
> > > > To: Doug Ledford <dledford@xxxxxxxxxx>; Jason Gunthorpe
> > > > <jgg@xxxxxxxxxxxx>
> > > > Cc: Leon Romanovsky <leonro@xxxxxxxxxxxx>; RDMA mailing list <linux-
> > > > rdma@xxxxxxxxxxxxxxx>
> > > > Subject: [PATCH rdma-next v3 2/2] RDMA/nldev: Allow IB device rename
> > > > through RDMA netlink
> > > >
> > > > From: Leon Romanovsky <leonro@xxxxxxxxxxxx>
> > > >
> > > > Provide an option to rename IB device name through RDMA netlink and limit
> > > > it to users with ADMIN capability only.
> > > >
> > > > Signed-off-by: Leon Romanovsky <leonro@xxxxxxxxxxxx>
> > > > drivers/infiniband/core/nldev.c | 34
> > > > ++++++++++++++++++++++++++++++++++
> > > > include/uapi/rdma/rdma_netlink.h | 3 ++-
> > > > 2 files changed, 36 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/infiniband/core/nldev.c b/drivers/infiniband/core/nldev.c
> > > > index ba5403fbcd88..573399e3ccc1 100644
> > > > +++ b/drivers/infiniband/core/nldev.c
> > > > @@ -646,6 +646,36 @@ static int nldev_get_doit(struct sk_buff *skb, struct
> > > > nlmsghdr *nlh,
> > > > return err;
> > > > }
> > > >
> > > > +static int nldev_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
> > > > + struct netlink_ext_ack *extack)
> > > > +{
> > > > + struct nlattr *tb[RDMA_NLDEV_ATTR_MAX];
> > > > + struct ib_device *device;
> > > > + u32 index;
> > > > + int err;
> > > > +
> > > > + err = nlmsg_parse(nlh, 0, tb, RDMA_NLDEV_ATTR_MAX - 1,
> > > > nldev_policy,
> > > > + extack);
> > > > + if (err || !tb[RDMA_NLDEV_ATTR_DEV_INDEX])
> > > > + return -EINVAL;
> > > > +
> > > > + index = nla_get_u32(tb[RDMA_NLDEV_ATTR_DEV_INDEX]);
> > > > + device = ib_device_get_by_index(index);
> > > > + if (!device)
> > > > + return -EINVAL;
> > > > +
> > > > + if (tb[RDMA_NLDEV_ATTR_DEV_NAME]) {
> > > > + char name[IB_DEVICE_NAME_MAX] = {};
> > > > +
> > > > + nla_strlcpy(name, tb[RDMA_NLDEV_ATTR_DEV_NAME],
> > > > + IB_DEVICE_NAME_MAX);
> > >
> > > I prefer sizeof(name) instead of IB_DEVICE_NAME_MAX in nla_strlcpy(). But its ok.
> >
> > IB_DEVICE_NAME_MAX and name are part of KAPI, it won't be changed
> > without introduction of new define and usage here emphasize it.
>
> strlcpy is the wrong thing here, if userspace provides a string too
> long for the kernel then it should fail not truncate.
It will fail, because it is protected by nla_policy used in nla_parse.
Thanks
>
> Jason
Attachment:
signature.asc
Description: PGP signature
