[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240801121310.GA10274@breakpoint.cc>
Date: Thu, 1 Aug 2024 14:13:10 +0200
From: Florian Westphal <fw@...len.de>
To: Christian Hopps <chopps@...pps.org>
Cc: devel@...ux-ipsec.org, Steffen Klassert <steffen.klassert@...unet.com>,
netdev@...r.kernel.org, Christian Hopps <chopps@...n.net>
Subject: Re: [PATCH ipsec-next v7 07/16] xfrm: iptfs: add new iptfs xfrm mode
impl
Christian Hopps <chopps@...pps.org> wrote:
> +static int __iptfs_init_state(struct xfrm_state *x,
> + struct xfrm_iptfs_data *xtfs)
> +{
> + /* Modify type (esp) adjustment values */
> +
> + if (x->props.family == AF_INET)
> + x->props.header_len += sizeof(struct iphdr) + sizeof(struct ip_iptfs_hdr);
> + else if (x->props.family == AF_INET6)
> + x->props.header_len += sizeof(struct ipv6hdr) + sizeof(struct ip_iptfs_hdr);
> + x->props.enc_hdr_len = sizeof(struct ip_iptfs_hdr);
> +
> + /* Always have a module reference if x->mode_data is set */
> + if (!try_module_get(x->mode_cbs->owner))
> + return -EINVAL;
If the comment means that we already have a module owner ref taken
before this try_module_get, then this should use __module_get and
a mention where the first ref was taken.
If not, then this needs an explanation as to what prevents another cpu to
rmmod the owning module between the lookup in xfrm_init_state and the
module reference in __iptfs_init_state.
cpu0 cpu1
xfrm_init_state
-> xfrm_get_mode_cbs rmmod
-> __iptfs_init_state xfrm_iptfs_fini
<interrupt> xfrm_unregister_mode_cbs
release memory
<resume>
try_module_get -> UaF
Powered by blists - more mailing lists