Re: [Patch net] cls_flower: call nla_ok() before nla_next()

On Tue, Jan 12, 2021 at 5:38 PM Jakub Kicinski [off-list ref] wrote:

On Mon, 11 Jan 2021 18:55:48 -0800 Cong Wang wrote:

quoted

From: Cong Wang <redacted>

fl_set_enc_opt() simply checks if there are still bytes left to parse,
but this is not sufficent as syzbot seems to be able to generate
malformatted netlink messages. nla_ok() is more strict so should be
used to validate the next nlattr here.

And nla_validate_nested_deprecated() has less strict check too, it is
probably too late to switch to the strict version, but we can just
call nla_ok() too after it.

Reported-and-tested-by: syzbot+2624e3778b18fc497c92@syzkaller.appspotmail.com
Fixes: 0a6e77784f49 ("net/sched: allow flower to match tunnel options")
Fixes: 79b1011cb33d ("net: sched: allow flower to match erspan options")
Cc: Pieter Jansen van Vuuren <redacted>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Xin Long <lucien.xin@gmail.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Cong Wang <redacted>

Thanks for keeping up with the syzbot bugs!

quoted

diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 1319986693fc..e265c443536e 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c

@@ -1272,6 +1272,8 @@ static int fl_set_enc_opt(struct nlattr **tb, struct fl_flow_key *key,

              nla_opt_msk = nla_data(tb[TCA_FLOWER_KEY_ENC_OPTS_MASK]);
              msk_depth = nla_len(tb[TCA_FLOWER_KEY_ENC_OPTS_MASK]);
+             if (!nla_ok(nla_opt_msk, msk_depth))
+                     return -EINVAL;

Can we just add another call to nla_validate_nested_deprecated()
here instead of having to worry about each attr individually?

No, we can not parse the nested attr here because different key types
have different attributes.

See below..

quoted

      }

      nla_for_each_attr(nla_opt_key, nla_enc_key,

@@ -1308,7 +1310,7 @@ static int fl_set_enc_opt(struct nlattr **tb, struct fl_flow_key *key,
                              return -EINVAL;
                      }

-                     if (msk_depth)
+                     if (nla_ok(nla_opt_msk, msk_depth))
                              nla_opt_msk = nla_next(nla_opt_msk, &msk_depth);

Should we not error otherwise? if msk_depth && !nla_ok() then the
message is clearly misformatted. If we don't error out we'll keep
reusing the same mask over and over, while the intention of this
code was to have mask per key AFAICT.

Yeah, erroring out sounds better. Will change this in v2.

Thanks!