WARNING in nla_get_range_unsigned

[Wei Chen]

Dear Linux Developers,

Recently, when using our tool to fuzz kernel, the following crash was triggered. Although this crash has been reported by syzbot https://syzkaller.appspot.com/bug?id=32e20c07949c6d6006f26466022469e33ae69108 and fixed in commit netlink: policy: correct validation type check, it still happens in the latest kernel version.

HEAD commit:  76dcd734eca
git tree: linux-next
compiler: clang 12.0.0
console output: https://drive.google.com/file/d/1reeOfFkfJp4-GUz_uMTh-uWXPLBJDcA6/view?usp=share_link
kernel config: https://drive.google.com/file/d/1jH4qV5XblPADvMDUlvS7DwtW0FroMoVB/view?usp=share_link
syz repro: https://drive.google.com/file/d/1Ong8vQn675RFU7R1O5HfiwWxp4UhnaIF/view?usp=share_link

IMPORTANT: if you fix the bug, please add the following tag to the commit:
Reported-by: Wei Chen <harperc...@gmail.com>

------------[ cut here ]------------
WARNING: CPU: 0 PID: 17743 at lib/nlattr.c:118 nla_get_range_unsigned+0x1d8/0x1e0 lib/nlattr.c:117
Modules linked in:
CPU: 0 PID: 17743 Comm: syz-executor.0 Not tainted 6.1.0-rc8 #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.13.0-48-gd9c812dda519-prebuilt.qemu.org 04/01/2014
RIP: 0010:nla_get_range_unsigned+0x1d8/0x1e0 lib/nlattr.c:117
Code: 8d ff 49 8b 75 08 ba 10 00 00 00 4c 89 f7 e8 0f d8 f8 02 5b 41 5c 41 5d 41 5e 41 5f 5d c3 e8 0f 57 7a ff eb 05 e8 08 57 7a ff <0f> 0b e9 a9 fe ff ff 90 55 41 57 41 56 41 54 53 49 89 f6 49 89 fc
RSP: 0018:ffffc90002df39b8 EFLAGS: 00010287
RAX: ffffffff81ad2f51 RBX: ffffffff85364d28 RCX: 0000000000040000
RDX: ffffc90000add000 RSI: 0000000000000268 RDI: 0000000000000269
RBP: 000000000000f940 R08: ffffffff81ad2dd8 R09: 0000000000000000
R10: 0001ffffffffffff R11: ffff888045136780 R12: ffff88803e174000
R13: ffffffff85364d20 R14: ffffc90002df3a30 R15: ffffffff85364d21
FS:  00007fab1e5c8700(0000) GS:ffff88807dc00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000000073f8d0 CR3: 000000004a789000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000600
Call Trace:
 <TASK>
 __netlink_policy_dump_write_attr+0x23d/0x990 net/netlink/policy.c:310
 netlink_policy_dump_write_attr+0x22/0x30 net/netlink/policy.c:411
 netlink_ack_tlv_fill net/netlink/af_netlink.c:2454 [inline]
 netlink_ack+0x546/0x760 net/netlink/af_netlink.c:2506
 netlink_rcv_skb+0x1b7/0x240 net/netlink/af_netlink.c:2546
 rtnetlink_rcv+0x18/0x20 net/core/rtnetlink.c:6109
 netlink_unicast_kernel net/netlink/af_netlink.c:1319 [inline]
 netlink_unicast+0x5e9/0x6b0 net/netlink/af_netlink.c:1345
 netlink_sendmsg+0x739/0x860 net/netlink/af_netlink.c:1921
 sock_sendmsg_nosec net/socket.c:714 [inline]
 sock_sendmsg net/socket.c:734 [inline]
 ____sys_sendmsg+0x38f/0x500 net/socket.c:2482
 ___sys_sendmsg net/socket.c:2536 [inline]
 __sys_sendmsg+0x197/0x230 net/socket.c:2565
 __do_sys_sendmsg net/socket.c:2574 [inline]
 __se_sys_sendmsg net/socket.c:2572 [inline]
 __x64_sys_sendmsg+0x42/0x50 net/socket.c:2572
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x2b/0x70 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x4697f9
Code: f7 d8 64 89 02 b8 ff ff ff ff c3 66 0f 1f 44 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 bc ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fab1e5c7c48 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 000000000077bf80 RCX: 00000000004697f9
RDX: 0000000000000000 RSI: 0000000020000000 RDI: 0000000000000003
RBP: 00000000004d29e9 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 000000000077bf80
R13: 0000000000000000 R14: 000000000077bf80 R15: 00007ffd7c0e6920
 </TASK>
---[ end trace 0000000000000000 ]---

Best,
Wei

[Ido Schimmel]

On Sat, Dec 17, 2022 at 05:21:24PM +0800, Wei Chen wrote:
> Dear Linux Developers,
>
> Recently, when using our tool to fuzz kernel, the following crash was
> triggered. Although this crash has been reported by syzbot
> https://syzkaller.appspot.com/bug?id=32e20c07949c6d6006f26466022469e33ae69108
> and fixed in commit netlink: policy: correct validation type check

> <https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=c30a3c957c885e618ddffc065f888be4f8d5a9bd>,

> it still happens in the latest kernel version.
>
> HEAD commit: 76dcd734eca
> git tree: linux-next
> compiler: clang 12.0.0
> console output:
> https://drive.google.com/file/d/1reeOfFkfJp4-GUz_uMTh-uWXPLBJDcA6/view?usp=share_link
> kernel config:
> https://drive.google.com/file/d/1jH4qV5XblPADvMDUlvS7DwtW0FroMoVB/view?usp=share_link
> syz repro:
> https://drive.google.com/file/d/1Ong8vQn675RFU7R1O5HfiwWxp4UhnaIF/view?usp=share_link

Can be reproduced with:

# tc action add mpls push label 3

Assuming you patch iproute2 to encode a wrong label length. For example:

diff --git a/tc/m_mpls.c b/tc/m_mpls.c
index 9b39d8533c21..2a43ca6c4dd3 100644
--- a/tc/m_mpls.c
+++ b/tc/m_mpls.c
@@ -191,7 +191,7 @@ static int parse_mpls(struct action_util *a, int *argc_p, char ***argv_p,
tail = addattr_nest(n, MAX_MSG, tca_id | NLA_F_NESTED);
addattr_l(n, MAX_MSG, TCA_MPLS_PARMS, &parm, sizeof(parm));
if (label != 0xffffffff)
- addattr_l(n, MAX_MSG, TCA_MPLS_LABEL, &label, sizeof(label));
+ addattr_l(n, MAX_MSG, TCA_MPLS_LABEL, &label, 8);
if (proto)
addattr_l(n, MAX_MSG, TCA_MPLS_PROTO, &proto, sizeof(proto));
if (tc != 0xff)

It does not seem valid to use NLA_POLICY_VALIDATE_FN() without
NLA_BINARY. Fixed for me by:

diff --git a/net/sched/act_mpls.c b/net/sched/act_mpls.c
index ff47ce4d3968..6b26bdb999d7 100644
--- a/net/sched/act_mpls.c
+++ b/net/sched/act_mpls.c
@@ -134,6 +134,11 @@ static int valid_label(const struct nlattr *attr,
{
const u32 *label = nla_data(attr);

+ if (nla_len(attr) != sizeof(*label)) {
+ NL_SET_ERR_MSG_MOD(extack, "Invalid MPLS label length");
+ return -EINVAL;
+ }
+
if (*label & ~MPLS_LABEL_MASK || *label == MPLS_LABEL_IMPLNULL) {
NL_SET_ERR_MSG_MOD(extack, "MPLS label out of range");
return -EINVAL;
@@ -145,7 +150,8 @@ static int valid_label(const struct nlattr *attr,
static const struct nla_policy mpls_policy[TCA_MPLS_MAX + 1] = {
[TCA_MPLS_PARMS] = NLA_POLICY_EXACT_LEN(sizeof(struct tc_mpls)),
[TCA_MPLS_PROTO] = { .type = NLA_U16 },
- [TCA_MPLS_LABEL] = NLA_POLICY_VALIDATE_FN(NLA_U32, valid_label),
+ [TCA_MPLS_LABEL] = NLA_POLICY_VALIDATE_FN(NLA_BINARY,
+ valid_label),
[TCA_MPLS_TC] = NLA_POLICY_RANGE(NLA_U8, 0, 7),
[TCA_MPLS_TTL] = NLA_POLICY_MIN(NLA_U8, 1),
[TCA_MPLS_BOS] = NLA_POLICY_RANGE(NLA_U8, 0, 1),

But please test with your reproducer as well.

For net-next we can try to remove the first argument from
NLA_POLICY_VALIDATE_FN() and set NLA_BINARY which is what everyone is
passing anyway.

Adding Johannes in case he has a better idea.

> --
> You received this message because you are subscribed to the Google Groups "syzkaller" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to syzkaller+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/syzkaller/CAO4mrffa_3PhjfA9hxTq_U9GjC%2B%2B0suGnme9oNcKE%3DGn%2Bg1iRg%40mail.gmail.com.