Re: [PATCH] net: ipv4: fix memory leak in netlbl_cipsov4_add_std

[Paul Moore]

On Mon, Jun 7, 2021 at 10:31 PM Dongliang Mu <mudongliangabcd@xxxxxxxxx> wrote:
> On Tue, Jun 8, 2021 at 9:57 AM Paul Moore <paul@xxxxxxxxxxxxxx> wrote:
> > On Mon, Jun 7, 2021 at 9:19 PM Nanyong Sun <sunnanyong@xxxxxxxxxx> wrote:
> > >
> > > Reported by syzkaller:
> > > BUG: memory leak
> > > unreferenced object 0xffff888105df7000 (size 64):
> > > comm "syz-executor842", pid 360, jiffies 4294824824 (age 22.546s)
> > > hex dump (first 32 bytes):
> > > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> > > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> > > backtrace:
> > > [<00000000e67ed558>] kmalloc include/linux/slab.h:590 [inline]
> > > [<00000000e67ed558>] kzalloc include/linux/slab.h:720 [inline]
> > > [<00000000e67ed558>] netlbl_cipsov4_add_std net/netlabel/netlabel_cipso_v4.c:145 [inline]
> > > [<00000000e67ed558>] netlbl_cipsov4_add+0x390/0x2340 net/netlabel/netlabel_cipso_v4.c:416
> > > [<0000000006040154>] genl_family_rcv_msg_doit.isra.0+0x20e/0x320 net/netlink/genetlink.c:739
> > > [<00000000204d7a1c>] genl_family_rcv_msg net/netlink/genetlink.c:783 [inline]
> > > [<00000000204d7a1c>] genl_rcv_msg+0x2bf/0x4f0 net/netlink/genetlink.c:800
> > > [<00000000c0d6a995>] netlink_rcv_skb+0x134/0x3d0 net/netlink/af_netlink.c:2504
> > > [<00000000d78b9d2c>] genl_rcv+0x24/0x40 net/netlink/genetlink.c:811
> > > [<000000009733081b>] netlink_unicast_kernel net/netlink/af_netlink.c:1314 [inline]
> > > [<000000009733081b>] netlink_unicast+0x4a0/0x6a0 net/netlink/af_netlink.c:1340
> > > [<00000000d5fd43b8>] netlink_sendmsg+0x789/0xc70 net/netlink/af_netlink.c:1929
> > > [<000000000a2d1e40>] sock_sendmsg_nosec net/socket.c:654 [inline]
> > > [<000000000a2d1e40>] sock_sendmsg+0x139/0x170 net/socket.c:674
> > > [<00000000321d1969>] ____sys_sendmsg+0x658/0x7d0 net/socket.c:2350
> > > [<00000000964e16bc>] ___sys_sendmsg+0xf8/0x170 net/socket.c:2404
> > > [<000000001615e288>] __sys_sendmsg+0xd3/0x190 net/socket.c:2433
> > > [<000000004ee8b6a5>] do_syscall_64+0x37/0x90 arch/x86/entry/common.c:47
> > > [<00000000171c7cee>] entry_SYSCALL_64_after_hwframe+0x44/0xae
> > >
> > > The memory of doi_def->map.std pointing is allocated in
> > > netlbl_cipsov4_add_std, but no place has freed it. It should be
> > > freed in cipso_v4_doi_free which frees the cipso DOI resource.
> > >
> > > Fixes: 96cb8e3313c7a ("[NetLabel]: CIPSOv4 and Unlabeled packet integration")
> > > Reported-by: Hulk Robot <hulkci@xxxxxxxxxx>
> > > Signed-off-by: Nanyong Sun <sunnanyong@xxxxxxxxxx>
> > > ---
> > >  net/ipv4/cipso_ipv4.c | 1 +
> > >  1 file changed, 1 insertion(+)
> >
> > Nice catch, thanks for fixing this.
> >
> > Acked-by: Paul Moore <paul@xxxxxxxxxxxxxx>
> >
> > > diff --git a/net/ipv4/cipso_ipv4.c b/net/ipv4/cipso_ipv4.c
> > > index d6e3a92841e3..099259fc826a 100644
> > > --- a/net/ipv4/cipso_ipv4.c
> > > +++ b/net/ipv4/cipso_ipv4.c
> > > @@ -471,6 +471,7 @@ void cipso_v4_doi_free(struct cipso_v4_doi *doi_def)
> > >                 kfree(doi_def->map.std->lvl.local);
> > >                 kfree(doi_def->map.std->cat.cipso);
> > >                 kfree(doi_def->map.std->cat.local);
> > > +               kfree(doi_def->map.std);
> > >                 break;
> > >         }
> > >         kfree(doi_def);
>
> Hi kernel developers,
>
> I doubt this patch may cause invalid free in other functions where
> map.std is not allocated or initialized, such as
> netlbl_cipsov4_add_local, netlbl_cipsov4_add_pass.

It isn't perfectly clear to me if you are implying there is a problem
with the proposed patch or not, so I thought it might help to try and
add some clarity.

The patch above frees the cipso_v4_doi->map.std field, which is only
valid when the cipso_v4_doi->type field is equal to
CIPSO_V4_MAP_TRANS.  This is why the cipso_v4_doi_free() function
checks the type field before freeing the cipso_v4_doi->map related
fields, and why the proposed patch places the new kfree() inside that
conditional code block.

If we look at netlalbel_cipsov4_add_pass() we see that the first thing
it does after allocating a cipso_v4_doi struct is to set the type
field to CIPSO_V4_MAP_PASS.  Any calls to cipso_v4_doi_free after this
point will not end up calling the proposed kfree() addition due to the
cipso_v4_doi->type check.

We see something very similar with netlbl_cipsov4_add_local(),
although in this case the type field is set to CIPSO_V4_MAP_LOCAL.
This type value will also not trigger the proposed kfree().

If you are aware of any other potential issues with this patch please
do let us know, but from what I can see, the two concerns you
presented here are not problems with the current or proposed code.

-- 
paul moore
www.paul-moore.com