Re: [PATCH] target: iscsi: use GFP_NOIO with loopback connections

From: Maurizio Lombardi <mlombard@redhat.com>
To: martin.petersen@oracle.com
Cc: michael.christie@oracle.com, target-devel@vger.kernel.org
Subject: Re: [PATCH] target: iscsi: use GFP_NOIO with loopback connections
Date: Wed, 22 Mar 2023 14:59:37 +0100	[thread overview]
Message-ID: <CAFL455n3w6MoWTC6sj2T_oSg38oF-zSpQFaiJshknd91OJShTg@mail.gmail.com> (raw)
In-Reply-To: <20230227075735.8695-1-mlombard@redhat.com>

po 27. 2. 2023 v 9:00 odesílatel Maurizio Lombardi <mlombard@redhat.com> napsal:
>
> Fix the issue by using memalloc_noio_*() to enable implicit GFP_NOIO
> in the vulnerable code paths, when the connection is in loopback.
>
> @@ -1289,6 +1291,12 @@ static int __iscsi_target_login_thread(struct iscsi_np *np)
>                 iscsit_free_conn(conn);
>                 return 1;
>         }
> +
> +       dst = sk_dst_get(conn->sock->sk);
> +       if (dst && dst->dev && dst->dev->flags & IFF_LOOPBACK)
> +               conn->loopback = true;
> +       dst_release(dst);
> +

As discussed in the other thread, conn->sock is not set when
isert/cxgbit etc are used.
So we need to check it here otherwise the kernel would crash.

I think that "conn->loopback" should be set by the iscsit_accept_np() callback.

Maurizio