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] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220128153548.GA1849166@nvidia.com>
Date:   Fri, 28 Jan 2022 11:35:48 -0400
From:   Jason Gunthorpe <jgg@...dia.com>
To:     Leon Romanovsky <leon@...nel.org>
Cc:     Maor Gottlieb <maorg@...dia.com>, linux-kernel@...r.kernel.org,
        linux-rdma@...r.kernel.org
Subject: Re: [PATCH rdma-next 1/3] RDMA/cma: Use correct address when leaving
 multicast group

On Tue, Jan 18, 2022 at 09:35:00AM +0200, Leon Romanovsky wrote:
> From: Maor Gottlieb <maorg@...dia.com>
> 
> In RoCE we should use cma_iboe_set_mgid and not cma_set_mgid to generate
> the mgid, otherwise we will try to remove incorrect address.
> 
> Fixes: b5de0c60cc30 ("RDMA/cma: Fix use after free race in roce multicast join")
> Signed-off-by: Maor Gottlieb <maorg@...dia.com>
> Signed-off-by: Leon Romanovsky <leonro@...dia.com>
> ---
>  drivers/infiniband/core/cma.c | 61 +++++++++++++++++------------------
>  1 file changed, 30 insertions(+), 31 deletions(-)

I didn't like this one so much and changed it into this:

@@ -67,8 +67,8 @@ static const char * const cma_events[] = {
        [RDMA_CM_EVENT_TIMEWAIT_EXIT]    = "timewait exit",
 };
 
-static void cma_set_mgid(struct rdma_id_private *id_priv, struct sockaddr *addr,
-                        union ib_gid *mgid);
+static void cma_iboe_set_mgid(struct sockaddr *addr, union ib_gid *mgid,
+                             enum ib_gid_type gid_type);
 
 const char *__attribute_const__ rdma_event_msg(enum rdma_cm_event_type event)
 {
@@ -1846,17 +1846,19 @@ static void destroy_mc(struct rdma_id_private *id_priv,
                if (dev_addr->bound_dev_if)
                        ndev = dev_get_by_index(dev_addr->net,
                                                dev_addr->bound_dev_if);
-               if (ndev) {
+               if (ndev && !send_only) {
+                       enum ib_gid_type gid_type;
                        union ib_gid mgid;
 
-                       cma_set_mgid(id_priv, (struct sockaddr *)&mc->addr,
-                                    &mgid);
-
-                       if (!send_only)
-                               cma_igmp_send(ndev, &mgid, false);
-
-                       dev_put(ndev);
+                       gid_type = id_priv->cma_dev->default_gid_type
+                                          [id_priv->id.port_num -
+                                           rdma_start_port(
+                                                   id_priv->cma_dev->device)];
+                       cma_iboe_set_mgid((struct sockaddr *)&mc->addr, &mgid,
+                                         gid_type);
+                       cma_igmp_send(ndev, &mgid, false);
                }
+               dev_put(ndev);
 
                cancel_work_sync(&mc->iboe_join.work);
        }

Thanks,
Jason

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ