netdev - Re: [PATCH net-next v2] net/packet: Add getsockopt support for PACKET_COPY

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAL+tcoAwPv7bquhmaUr_NckEfQbp2xuNNOSswtmxRHHVCiY9+A@mail.gmail.com>
Date: Fri, 8 Mar 2024 21:27:24 +0800
From: Jason Xing <kerneljasonxing@...il.com>
To: Juntong Deng <juntong.deng@...look.com>
Cc: Eric Dumazet <edumazet@...gle.com>, willemdebruijn.kernel@...il.com, 
	davem@...emloft.net, kuba@...nel.org, pabeni@...hat.com, 
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next v2] net/packet: Add getsockopt support for PACKET_COPY_THRESH

On Fri, Mar 8, 2024 at 9:19 PM Juntong Deng <juntong.deng@...look.com> wrote:
>
> On 2024/3/8 21:11, Jason Xing wrote:
> > On Fri, Mar 8, 2024 at 8:56 PM Eric Dumazet <edumazet@...gle.com> wrote:
> >>
> >> On Fri, Mar 8, 2024 at 1:43 PM Juntong Deng <juntong.deng@...look.com> wrote:
> >>>
> >>> Currently getsockopt does not support PACKET_COPY_THRESH,
> >>> and we are unable to get the value of PACKET_COPY_THRESH
> >>> socket option through getsockopt.
> >>>
> >>> This patch adds getsockopt support for PACKET_COPY_THRESH.
> >>>
> >>> In addition, this patch converts access to copy_thresh to
> >>> READ_ONCE/WRITE_ONCE.
> >>>
> >>> Signed-off-by: Juntong Deng <juntong.deng@...look.com>
> >>> ---
> >>> V1 -> V2: Use READ_ONCE/WRITE_ONCE.
> >>>
> >>>   net/packet/af_packet.c | 7 +++++--
> >>>   1 file changed, 5 insertions(+), 2 deletions(-)
> >>>
> >>> diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
> >>> index 0db31ca4982d..61270826b9ac 100644
> >>> --- a/net/packet/af_packet.c
> >>> +++ b/net/packet/af_packet.c
> >>> @@ -2318,7 +2318,7 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev,
> >>>          }
> >>>          if (po->tp_version <= TPACKET_V2) {
> >>>                  if (macoff + snaplen > po->rx_ring.frame_size) {
> >>> -                       if (po->copy_thresh &&
> >>> +                       if (READ_ONCE(po->copy_thresh) &&
> >>>                              atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
> >>>                                  if (skb_shared(skb)) {
> >>>                                          copy_skb = skb_clone(skb, GFP_ATOMIC);
> >>> @@ -3836,7 +3836,7 @@ packet_setsockopt(struct socket *sock, int level, int optname, sockptr_t optval,
> >>>                  if (copy_from_sockptr(&val, optval, sizeof(val)))
> >>>                          return -EFAULT;
> >>>
> >>> -               pkt_sk(sk)->copy_thresh = val;
> >>> +               WRITE_ONCE(pkt_sk(sk)->copy_thresh, val);
> >>>                  return 0;
> >>>          }
> >>>          case PACKET_VERSION:
> >>> @@ -4090,6 +4090,9 @@ static int packet_getsockopt(struct socket *sock, int level, int optname,
> >>>          case PACKET_VNET_HDR_SZ:
> >>>                  val = READ_ONCE(po->vnet_hdr_sz);
> >>>                  break;
> >>> +       case PACKET_COPY_THRESH:
> >>> +               val = READ_ONCE(pkt_sk(sk)->copy_thresh);
> >>> +               break;
> >>>          case PACKET_VERSION:
> >>>                  val = po->tp_version;
> >>>                  break;
> >>> --
> >>> 2.39.2
> >>>
> >>
> >> I think you forgot to change net/packet/diag.c pdiag_put_info()
> >>
> >
> > Ah, he updated his patch so soon even without taking a break.
> >
> > I just replied to the v1 thread with three changes made.
> >
> > Juntong, you could check your v1 patch thread and you will see the
> > missing point.
> >
> > Thanks,
> > Jason
>
> Hi Jason,
>
> Thanks, I have fixed this in the v3 patch.

The patch itself looks good to me.

But next time please do not post the patch too often, you have to wait
around 24 hour at least. We need to obey the rule. Then you have
plenty of time to collect/rethink about those suggestions...

Thanks,
Jason