linux-rdma.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Zhu Yanjun <zyjzyj2000@gmail.com>
To: Kamal Heib <kheib@redhat.com>
Cc: Leon Romanovsky <leon@kernel.org>,
	Zhu Yanjun <yanjun.zhu@intel.com>, Yi Zhang <yi.zhang@redhat.com>,
	Doug Ledford <dledford@redhat.com>,
	RDMA mailing list <linux-rdma@vger.kernel.org>,
	Jason Gunthorpe <jgg@ziepe.ca>
Subject: Re: [PATCHv5 for-next 1/1] RDMA/rxe: Disable ipv6 features when ipv6.disable in cmdline
Date: Thu, 15 Apr 2021 22:53:55 +0800	[thread overview]
Message-ID: <CAD=hENfmxwpCJPpbmWoCkk+WLk+AmZeNdSzg_LnaHZ=DHD-Wvg@mail.gmail.com> (raw)
In-Reply-To: <9d410fde-ade3-c7fe-e73e-ca0103ec67c5@redhat.com>

On Wed, Apr 14, 2021 at 3:34 AM Kamal Heib <kheib@redhat.com> wrote:
>
>
>
> On 4/13/21 5:39 PM, Leon Romanovsky wrote:
> > On Tue, Apr 13, 2021 at 04:56:05PM +0300, Kamal Heib wrote:
> >>
> >>
> >> On 4/14/21 2:42 AM, Zhu Yanjun wrote:
> >>> From: Zhu Yanjun <zyjzyj2000@gmail.com>
> >>>
> >>> When ipv6.disable=1 is set in cmdline, ipv6 is actually disabled
> >>> in the stack. As such, the operations of ipv6 in RXE will fail.
> >>> So ipv6 features in RXE should also be disabled in RXE.
> >>>
> >>> Link: https://lore.kernel.org/linux-rdma/880d7b59-4b17-a44f-1a91-88257bfc3aaa@redhat.com/T/#t
> >>> Fixes: 8700e3e7c4857 ("Soft RoCE driver")
> >>> Reported-by: Yi Zhang <yi.zhang@redhat.com>
> >>> Signed-off-by: Zhu Yanjun <zyjzyj2000@gmail.com>
> >>> ---
> >>> V4->V5: Clean up signature block and remove error message
> >>> V3->V4: Check the returned value instead of ipv6 module
> >>> V2->V3: Remove print message
> >>> V1->V2: Modify the pr_info messages
> >>> ---
> >>>  drivers/infiniband/sw/rxe/rxe_net.c | 13 ++++++++++++-
> >>>  1 file changed, 12 insertions(+), 1 deletion(-)
> >>>
> >>> diff --git a/drivers/infiniband/sw/rxe/rxe_net.c b/drivers/infiniband/sw/rxe/rxe_net.c
> >>> index 01662727dca0..984c3ac449bd 100644
> >>> --- a/drivers/infiniband/sw/rxe/rxe_net.c
> >>> +++ b/drivers/infiniband/sw/rxe/rxe_net.c
> >>> @@ -208,7 +208,13 @@ static struct socket *rxe_setup_udp_tunnel(struct net *net, __be16 port,
> >>>     /* Create UDP socket */
> >>>     err = udp_sock_create(net, &udp_cfg, &sock);
> >>>     if (err < 0) {
> >>> -           pr_err("failed to create udp socket. err = %d\n", err);
> >>> +           /* If UDP tunnel over ipv6 fails with -EAFNOSUPPORT, the tunnel
> >>> +            * over ipv4 still works. This error message will not pop out.
> >>> +            * If UDP tunnle over ipv4 fails or other errors with ipv6
> >>> +            * tunnel, this error should pop out.
> >>> +            */
> >>> +           if (!((err == -EAFNOSUPPORT) && (ipv6)))
> >>> +                   pr_err("failed to create udp socket. err = %d\n", err);
> >>>             return ERR_PTR(err);
> >>>     }
> >>>
> >>> @@ -620,6 +626,11 @@ static int rxe_net_ipv6_init(void)
> >>>     recv_sockets.sk6 = rxe_setup_udp_tunnel(&init_net,
> >>>                                             htons(ROCE_V2_UDP_DPORT), true);
> >>>     if (IS_ERR(recv_sockets.sk6)) {
> >>> +           /* Though IPv6 is not supported, IPv4 still needs to continue
> >>> +            */
> >>> +           if (PTR_ERR(recv_sockets.sk6) == -EAFNOSUPPORT)
> >>> +                   return 0;
> >>> +
> >>>             recv_sockets.sk6 = NULL;
> >>>             pr_err("Failed to create IPv6 UDP tunnel\n");
> >>>             return -1;
> >>>
> >>
> >> I think the following change is much simpler than changing the udp_sock_create()
> >> helper function?
> >>
> >>
> >> diff --git a/drivers/infiniband/sw/rxe/rxe_net.c
> >> b/drivers/infiniband/sw/rxe/rxe_net.c
> >> index 01662727dca0..b56d6f76ab31 100644
> >> --- a/drivers/infiniband/sw/rxe/rxe_net.c
> >> +++ b/drivers/infiniband/sw/rxe/rxe_net.c
> >> @@ -621,6 +621,11 @@ static int rxe_net_ipv6_init(void)
> >>                                                 htons(ROCE_V2_UDP_DPORT), true);
> >>         if (IS_ERR(recv_sockets.sk6)) {
> >>                 recv_sockets.sk6 = NULL;
> >> +               if (PTR_ERR(recv_sockets.sk6) == -EAFNOSUPPORT) {
> >
> > You have "recv_sockets.sk6 = NULL;" in the line above.
> >
>
> Sorry, my bad...
>
> The idea is to handle this issue in the error path of rxe_net_ipv6_init()
> instead of changing the udp_sock_create(), also to make sure that

We do not change udp_sock_create.

> "recv_sockets.sk6" is set to NULL.

Is it necessary to set recv_sockets_sk6 to NULL?

Zhu Yanjun
>
> diff --git a/drivers/infiniband/sw/rxe/rxe_net.c
> b/drivers/infiniband/sw/rxe/rxe_net.c
> index 01662727dca0..445a47f82f42 100644
> --- a/drivers/infiniband/sw/rxe/rxe_net.c
> +++ b/drivers/infiniband/sw/rxe/rxe_net.c
> @@ -615,17 +615,25 @@ static int rxe_net_ipv4_init(void)
>
>  static int rxe_net_ipv6_init(void)
>  {
> +       int err = 0;
>  #if IS_ENABLED(CONFIG_IPV6)
>
>         recv_sockets.sk6 = rxe_setup_udp_tunnel(&init_net,
>                                                 htons(ROCE_V2_UDP_DPORT), true);
>         if (IS_ERR(recv_sockets.sk6)) {
> +
> +               if (PTR_ERR(recv_sockets.sk6) == -EAFNOSUPPORT) {
> +                       pr_warn("Create IPv6 UDP tunnel is not supported\n");
> +                       err = 0;
> +               } else {
> +                       pr_err("Failed to create IPv6 UDP tunnel\n");
> +                       err = -1;
> +               }
> +
>                 recv_sockets.sk6 = NULL;
> -               pr_err("Failed to create IPv6 UDP tunnel\n");
> -               return -1;
>         }
>  #endif
> -       return 0;
> +       return err;
>  }
>
> >> +                       pr_warn("Create IPv6 UDP tunnel is not supported\n");
> >> +                       return 0;
> >> +               }
> >> +
> >>                 pr_err("Failed to create IPv6 UDP tunnel\n");
> >>                 return -1;
> >>         }
> >> --
> >> 2.26.3
> >>
> >>
> >> Thanks,
> >> Kamal
> >>
> >
>

      reply	other threads:[~2021-04-15 14:55 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-13 23:42 [PATCHv5 for-next 1/1] RDMA/rxe: Disable ipv6 features when ipv6.disable in cmdline Zhu Yanjun
2021-04-13 10:48 ` Leon Romanovsky
2021-04-14  3:46   ` Zhu Yanjun
2021-04-13 13:56 ` Kamal Heib
2021-04-13 14:39   ` Leon Romanovsky
2021-04-13 19:34     ` Kamal Heib
2021-04-15 14:53       ` Zhu Yanjun [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAD=hENfmxwpCJPpbmWoCkk+WLk+AmZeNdSzg_LnaHZ=DHD-Wvg@mail.gmail.com' \
    --to=zyjzyj2000@gmail.com \
    --cc=dledford@redhat.com \
    --cc=jgg@ziepe.ca \
    --cc=kheib@redhat.com \
    --cc=leon@kernel.org \
    --cc=linux-rdma@vger.kernel.org \
    --cc=yanjun.zhu@intel.com \
    --cc=yi.zhang@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).