Re: [PATCH net] net: fix use-after-free when UDP GRO with shared fraglist

On 2021/4/19 8:35, Dongseok Yi wrote:

On Sat, Apr 17, 2021 at 11:44:35AM +0800, Yunsheng Lin wrote:

quoted

On 2021/1/6 11:32, Dongseok Yi wrote:

quoted

On 2021-01-06 12:07, Willem de Bruijn wrote:

quoted

On Tue, Jan 5, 2021 at 8:29 PM Dongseok Yi [off-list ref] wrote:

quoted

On 2021-01-05 06:03, Willem de Bruijn wrote:

quoted

On Mon, Jan 4, 2021 at 4:00 AM Dongseok Yi [off-list ref] wrote:

quoted

skbs in frag_list could be shared by pskb_expand_head() from BPF.

Can you elaborate on the BPF connection?

With the following registered ptypes,

/proc/net # cat ptype
Type Device      Function
ALL           tpacket_rcv
0800          ip_rcv.cfi_jt
0011          llc_rcv.cfi_jt
0004          llc_rcv.cfi_jt
0806          arp_rcv
86dd          ipv6_rcv.cfi_jt

BPF checks skb_ensure_writable between tpacket_rcv and ip_rcv
(or ipv6_rcv). And it calls pskb_expand_head.

[  132.051228] pskb_expand_head+0x360/0x378
[  132.051237] skb_ensure_writable+0xa0/0xc4
[  132.051249] bpf_skb_pull_data+0x28/0x60
[  132.051262] bpf_prog_331d69c77ea5e964_schedcls_ingres+0x5f4/0x1000
[  132.051273] cls_bpf_classify+0x254/0x348
[  132.051284] tcf_classify+0xa4/0x180

Ah, you have a BPF program loaded at TC. That was not entirely obvious.

This program gets called after packet sockets with ptype_all, before
those with a specific protocol.

Tcpdump will have inserted a program with ptype_all, which cloned the
skb. This triggers skb_ensure_writable -> pskb_expand_head ->
skb_clone_fraglist -> skb_get.

quoted

[  132.051294] __netif_receive_skb_core+0x590/0xd28
[  132.051303] __netif_receive_skb+0x50/0x17c
[  132.051312] process_backlog+0x15c/0x1b8

quoted

quoted

While tcpdump, sk_receive_queue of PF_PACKET has the original frag_list.
But the same frag_list is queued to PF_INET (or PF_INET6) as the fraglist
chain made by skb_segment_list().

If the new skb (not frag_list) is queued to one of the sk_receive_queue,
multiple ptypes can see this. The skb could be released by ptypes and
it causes use-after-free.

If I understand correctly, a udp-gro-list skb makes it up the receive
path with one or more active packet sockets.

The packet socket will call skb_clone after accepting the filter. This
replaces the head_skb, but shares the skb_shinfo and thus frag_list.

udp_rcv_segment later converts the udp-gro-list skb to a list of
regular packets to pass these one-by-one to udp_queue_rcv_one_skb.
Now all the frags are fully fledged packets, with headers pushed
before the payload. This does not change their refcount anymore than
the skb_clone in pf_packet did. This should be 1.

Eventually udp_recvmsg will call skb_consume_udp on each packet.

The packet socket eventually also frees its cloned head_skb, which triggers

  kfree_skb_list(shinfo->frag_list)
    kfree_skb
      skb_unref
        refcount_dec_and_test(&skb->users)

Every your understanding is right, but

quoted

quoted

[ 4443.426215] ------------[ cut here ]------------
[ 4443.426222] refcount_t: underflow; use-after-free.
[ 4443.426291] WARNING: CPU: 7 PID: 28161 at lib/refcount.c:190
refcount_dec_and_test_checked+0xa4/0xc8
[ 4443.426726] pstate: 60400005 (nZCv daif +PAN -UAO)
[ 4443.426732] pc : refcount_dec_and_test_checked+0xa4/0xc8
[ 4443.426737] lr : refcount_dec_and_test_checked+0xa0/0xc8
[ 4443.426808] Call trace:
[ 4443.426813]  refcount_dec_and_test_checked+0xa4/0xc8
[ 4443.426823]  skb_release_data+0x144/0x264
[ 4443.426828]  kfree_skb+0x58/0xc4
[ 4443.426832]  skb_queue_purge+0x64/0x9c
[ 4443.426844]  packet_set_ring+0x5f0/0x820
[ 4443.426849]  packet_setsockopt+0x5a4/0xcd0
[ 4443.426853]  __sys_setsockopt+0x188/0x278
[ 4443.426858]  __arm64_sys_setsockopt+0x28/0x38
[ 4443.426869]  el0_svc_common+0xf0/0x1d0
[ 4443.426873]  el0_svc_handler+0x74/0x98
[ 4443.426880]  el0_svc+0x8/0xc

Fixes: 3a1296a38d0c (net: Support GRO/GSO fraglist chaining.)
Signed-off-by: Dongseok Yi <redacted>
---
 net/core/skbuff.c | 20 +++++++++++++++++++-
 1 file changed, 19 insertions(+), 1 deletion(-)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index f62cae3..1dcbda8 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c

@@ -3655,7 +3655,8 @@ struct sk_buff *skb_segment_list(struct sk_buff *skb,
        unsigned int delta_truesize = 0;
        unsigned int delta_len = 0;
        struct sk_buff *tail = NULL;
-       struct sk_buff *nskb;
+       struct sk_buff *nskb, *tmp;
+       int err;

        skb_push(skb, -skb_network_offset(skb) + offset);

@@ -3665,11 +3666,28 @@ struct sk_buff *skb_segment_list(struct sk_buff *skb,
                nskb = list_skb;
                list_skb = list_skb->next;

+               err = 0;
+               if (skb_shared(nskb)) {

I must be missing something still. This does not square with my
understanding that the two sockets are operating on clones, with each
frag_list skb having skb->users == 1.

Unless the packet socket patch previously also triggered an
skb_unclone/pskb_expand_head, as that call skb_clone_fraglist, which
calls skb_get on each frag_list skb.

A cloned skb after tpacket_rcv cannot go through skb_ensure_writable
with the original shinfo. pskb_expand_head reallocates the shinfo of
the skb and call skb_clone_fraglist. skb_release_data in
pskb_expand_head could not reduce skb->users of the each frag_list skb
if skb_shinfo(skb)->dataref == 2.

After the reallocation, skb_shinfo(skb)->dataref == 1 but each frag_list
skb could have skb->users == 2.

Hi, Dongseok
   I understand there is liner head data shared between the frag_list skb in the
cloned skb(cloned by pf_packet?) and original skb, which should not be shared
when skb_segment_list() converts the frag_list skb into regular packet.

   But both skb->users of original and cloned skb is one(skb_shinfo(skb)->dataref
is one for both skb too), and skb->users of each fraglist skb is two because both
original and cloned skb is linking to the same fraglist pointer, and there is
"skb_shinfo(skb)->frag_list = NULL" for original skb in the begin of skb_segment_list(),
if kfree_skb() is called with original skb, the fraglist skb will not be freed.
If kfree_skb is called with original skb,cloned skb and each fraglist skb here, the
reference counter for three of them seem right here, so why is there a refcount_t
warning in the commit log? am I missing something obvious here?

Sorry for bringing up this thread again.

A skb which detects use-after-free was not a part of frag_list. Please
check the commit msg once again.

I checked the commit msg again, but still have not figured it out yet:)

So I tried to see if I understand the skb'reference counting correctly:

skb->user is used to reference counting the "struct sk_buff", and
skb_shinfo(skb)->dataref is used to reference counting head data.

skb_clone(): allocate a sperate "struct sk_buff" but share the head data
             with the original skb, so skb_shinfo()->dataref need
             incrmenting.

pskb_expand_head(): allocate a sperate head data(which includes the space
                    for skb_shinfo(skb)), since the original head data
		    and the new head data' skb_shinfo()->frag_list both
                    point to the same fraglist skb, so each fraglist_skb's
		    skb->users need incrmenting, and original head data's
		    skb_shinfo() need decrmenting.


So after pf_packet called skb_clone() and pskb_expand_head(), we have:

    old skb              new skb
      |                     |
      |                     |
old head data         new head data
        \                   /
          \                /
           \              /
             \           /
              \         /
             fraglist_skb1 -> fraglist_skb2 -> fraglist_skb3 .....

So both old and new skb' skb->user is one, both old and new head data's
skb_shinfo()->dataref is one, and both old and new head data'
skb_shinfo()->frag_list points to fraglist_skb1, and each fraglist_skb's
skb->user is two.

Each fraglist_skb points to a head data, and its skb_shinfo()->dataref
is one too.

Suppose old skb is called with skb_segment_list(), without this patch,
we have:

                         new skb
                            |
                            |
                     new head data
                            /
                           /
                          /
                         /
                        /
       old skb -> fraglist_skb1 -> fraglist_skb2 -> fraglist_skb3 .....
          |
          |
    old head data

And old skb and each fraglist_skb become a regular packet, so freeing
the old skb, new skb and each fraglist_skb here do not seems to have
any reference counting problem, because each fraglist_skb's skb->user
is two, right?

Both sk_receive_queue of PF_PACKET and PF_INET (or PF_INET6) can have
a link for the same frag_skbs chain. 

Does "frag_skbs chain" means fraglist_skb1? It seems only new head data's
skb_shinfo()->frag_list points to fraglist_skb1


If a new skb (*not frags*) is

queued to one of the sk_receive_queue, multiple ptypes can see and
release this. It causes use-after-free.

Does "a new skb" mean each fraglist_skb after skb_segment_list()? Or other
new incoming skb?

I am not so familiar with the PF_PACKET and PF_INET, so still have hard
time figuring how the reference counting goes wrong here:)

quoted

quoted

quoted

Yes, that makes sense. skb_clone_fraglist just increments the
frag_list skb's refcounts.

skb_segment_list must create an unshared struct sk_buff before it
changes skb data to insert the protocol headers.

.