[PATCH net 0/3] Fix out of bounds when parsing TCP options

This series fixes out-of-bounds access in various places in the kernel
where parsing of TCP options takes place. Fortunately, many more
occurrences don't have this bug.

Maxim Mikityanskiy (3):
  netfilter: synproxy: Fix out of bounds when parsing TCP options
  mptcp: Fix out of bounds when parsing TCP options
  sch_cake: Fix out of bounds when parsing TCP options

 net/mptcp/options.c              | 2 ++
 net/netfilter/nf_synproxy_core.c | 2 ++
 net/sched/sch_cake.c             | 4 ++++
 3 files changed, 8 insertions(+)

-- 
2.25.1

The TCP option parser in synproxy (synproxy_parse_options) could read
one byte out of bounds. When the length is 1, the execution flow gets
into the loop, reads one byte of the opcode, and if the opcode is
neither TCPOPT_EOL nor TCPOPT_NOP, it reads one more byte, which exceeds
the length of 1.

This fix is inspired by commit 9609dad263f8 ("ipv4: tcp_input: fix stack
out of bounds when parsing TCP options.").

Cc: Young Xiao <redacted>
Fixes: 48b1de4c110a ("netfilter: add SYNPROXY core/target")
Signed-off-by: Maxim Mikityanskiy <redacted>
---
 net/netfilter/nf_synproxy_core.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/netfilter/nf_synproxy_core.c b/net/netfilter/nf_synproxy_core.c
index b100c04a0e43..621eb5ef9727 100644
--- a/net/netfilter/nf_synproxy_core.c
+++ b/net/netfilter/nf_synproxy_core.c

@@ -47,6 +47,8 @@ synproxy_parse_options(const struct sk_buff *skb, unsigned int doff,
 			length--;
 			continue;
 		default:
+			if (length < 2)
+				return true;
 			opsize = *ptr++;
 			if (opsize < 2)
 				return true;

-- 
2.25.1

The TCP option parser in mptcp (mptcp_get_options) could read one byte
out of bounds. When the length is 1, the execution flow gets into the
loop, reads one byte of the opcode, and if the opcode is neither
TCPOPT_EOL nor TCPOPT_NOP, it reads one more byte, which exceeds the
length of 1.

This fix is inspired by commit 9609dad263f8 ("ipv4: tcp_input: fix stack
out of bounds when parsing TCP options.").

Cc: Young Xiao <redacted>
Fixes: cec37a6e41aa ("mptcp: Handle MP_CAPABLE options for outgoing connections")
Signed-off-by: Maxim Mikityanskiy <redacted>
---
 net/mptcp/options.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/mptcp/options.c b/net/mptcp/options.c
index 6b825fb3fa83..9b263f27ce9b 100644
--- a/net/mptcp/options.c
+++ b/net/mptcp/options.c

@@ -356,6 +356,8 @@ void mptcp_get_options(const struct sk_buff *skb,
 			length--;
 			continue;
 		default:
+			if (length < 2)
+				return;
 			opsize = *ptr++;
 			if (opsize < 2) /* "silly options" */
 				return;

-- 
2.25.1

The TCP option parser in cake qdisc (cake_get_tcpopt and
cake_tcph_may_drop) could read one byte out of bounds. When the length
is 1, the execution flow gets into the loop, reads one byte of the
opcode, and if the opcode is neither TCPOPT_EOL nor TCPOPT_NOP, it reads
one more byte, which exceeds the length of 1.

This fix is inspired by commit 9609dad263f8 ("ipv4: tcp_input: fix stack
out of bounds when parsing TCP options.").

Cc: Young Xiao <redacted>
Fixes: 8b7138814f29 ("sch_cake: Add optional ACK filter")
Signed-off-by: Maxim Mikityanskiy <redacted>
---
 net/sched/sch_cake.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/sched/sch_cake.c b/net/sched/sch_cake.c
index 7d37638ee1c7..6b03eebf0a78 100644
--- a/net/sched/sch_cake.c
+++ b/net/sched/sch_cake.c

@@ -967,6 +967,8 @@ static const void *cake_get_tcpopt(const struct tcphdr *tcph,
 			length--;
 			continue;
 		}
+		if (length < 2)
+			break;
 		opsize = *ptr++;
 		if (opsize < 2 || opsize > length)
 			break;

@@ -1104,6 +1106,8 @@ static bool cake_tcph_may_drop(const struct tcphdr *tcph,
 			length--;
 			continue;
 		}
+		if (length < 2)
+			break;
 		opsize = *ptr++;
 		if (opsize < 2 || opsize > length)
 			break;

-- 
2.25.1

Maxim Mikityanskiy [off-list ref] wrote:

Would you mind a v2 that also rejects bogus th->doff value when
computing the length?

Thanks.

Maxim Mikityanskiy [off-list ref] writes:

The TCP option parser in cake qdisc (cake_get_tcpopt and
cake_tcph_may_drop) could read one byte out of bounds. When the length
is 1, the execution flow gets into the loop, reads one byte of the
opcode, and if the opcode is neither TCPOPT_EOL nor TCPOPT_NOP, it reads
one more byte, which exceeds the length of 1.

This fix is inspired by commit 9609dad263f8 ("ipv4: tcp_input: fix stack
out of bounds when parsing TCP options.").

Cc: Young Xiao <redacted>
Fixes: 8b7138814f29 ("sch_cake: Add optional ACK filter")
Signed-off-by: Maxim Mikityanskiy <redacted>

Thanks for fixing this!

Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>

On Wed, 9 Jun 2021, Maxim Mikityanskiy wrote:

Florian's comment on patch 1 prompted me to double-check th->doff 
validation, and for MPTCP we're covered by the check in tcp_v4_rcv(). So 
this patch looks good:

Reviewed-by: Mat Martineau <redacted>

If you send a v2 series, please also cc: mptcp@lists.linux.dev

Thanks!

--
Mat Martineau
Intel

On 2021-06-09 17:51, Florian Westphal wrote:

Maxim Mikityanskiy [off-list ref] wrote:

quoted

The TCP option parser in synproxy (synproxy_parse_options) could read
one byte out of bounds. When the length is 1, the execution flow gets
into the loop, reads one byte of the opcode, and if the opcode is
neither TCPOPT_EOL nor TCPOPT_NOP, it reads one more byte, which exceeds
the length of 1.

This fix is inspired by commit 9609dad263f8 ("ipv4: tcp_input: fix stack
out of bounds when parsing TCP options.").

Cc: Young Xiao <redacted>
Fixes: 48b1de4c110a ("netfilter: add SYNPROXY core/target")
Signed-off-by: Maxim Mikityanskiy <redacted>
---
  net/netfilter/nf_synproxy_core.c | 2 ++
  1 file changed, 2 insertions(+)

diff --git a/net/netfilter/nf_synproxy_core.c b/net/netfilter/nf_synproxy_core.c
index b100c04a0e43..621eb5ef9727 100644
--- a/net/netfilter/nf_synproxy_core.c
+++ b/net/netfilter/nf_synproxy_core.c

@@ -47,6 +47,8 @@ synproxy_parse_options(const struct sk_buff *skb, unsigned int doff,
  			length--;
  			continue;
  		default:
+			if (length < 2)
+				return true;

Would you mind a v2 that also rejects bogus th->doff value when
computing the length?

Could you elaborate? The length is a signed int calculated as `(th->doff 
* 4) - sizeof(*th)`. Invalid doff values (0..4) lead to negative length, 
so we never enter the loop. Or are you concerned of passing a negative 
length to skb_header_pointer?

Thanks.

Maxim Mikityanskiy [off-list ref] wrote:

On 2021-06-09 17:51, Florian Westphal wrote:

quoted

Maxim Mikityanskiy [off-list ref] wrote:

quoted

The TCP option parser in synproxy (synproxy_parse_options) could read
one byte out of bounds. When the length is 1, the execution flow gets
into the loop, reads one byte of the opcode, and if the opcode is
neither TCPOPT_EOL nor TCPOPT_NOP, it reads one more byte, which exceeds
the length of 1.

This fix is inspired by commit 9609dad263f8 ("ipv4: tcp_input: fix stack
out of bounds when parsing TCP options.").

Cc: Young Xiao <redacted>
Fixes: 48b1de4c110a ("netfilter: add SYNPROXY core/target")
Signed-off-by: Maxim Mikityanskiy <redacted>
---
  net/netfilter/nf_synproxy_core.c | 2 ++
  1 file changed, 2 insertions(+)

diff --git a/net/netfilter/nf_synproxy_core.c b/net/netfilter/nf_synproxy_core.c
index b100c04a0e43..621eb5ef9727 100644
--- a/net/netfilter/nf_synproxy_core.c
+++ b/net/netfilter/nf_synproxy_core.c

@@ -47,6 +47,8 @@ synproxy_parse_options(const struct sk_buff *skb, unsigned int doff,
  			length--;
  			continue;
  		default:
+			if (length < 2)
+				return true;

Would you mind a v2 that also rejects bogus th->doff value when
computing the length?

Could you elaborate? The length is a signed int calculated as `(th->doff *
4) - sizeof(*th)`. Invalid doff values (0..4) lead to negative length, so we
never enter the loop. Or are you concerned of passing a negative length to
skb_header_pointer?

Yes, negative length to skb_header_pointer.  For other usage (mptcp for
example) tcp stack validated th->doff already, but thats not the case for synproxy.

On 2021-06-10 00:51, Toke Høiland-Jørgensen wrote:

Maxim Mikityanskiy [off-list ref] writes:

quoted

The TCP option parser in cake qdisc (cake_get_tcpopt and
cake_tcph_may_drop) could read one byte out of bounds. When the length
is 1, the execution flow gets into the loop, reads one byte of the
opcode, and if the opcode is neither TCPOPT_EOL nor TCPOPT_NOP, it reads
one more byte, which exceeds the length of 1.

This fix is inspired by commit 9609dad263f8 ("ipv4: tcp_input: fix stack
out of bounds when parsing TCP options.").

Cc: Young Xiao <redacted>
Fixes: 8b7138814f29 ("sch_cake: Add optional ACK filter")
Signed-off-by: Maxim Mikityanskiy <redacted>

Thanks for fixing this!

Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>

Could you also review whether Florian's comment on patch 1 is relevant 
to this patch too? I have concerns about cake_get_tcphdr, which returns 
`skb_header_pointer(skb, offset, min(__tcp_hdrlen(tcph), bufsize), 
buf)`. Although I don't see a way for it to get out of bounds (it will 
read garbage instead of TCP header in the worst case), such code doesn't 
look robust.

It's not possible for it to get out of bounds, because there is a call 
to skb_header_pointer above with sizeof(_tcph), which ensures that the 
SKB has at least 20 bytes after the beginning of the TCP header, which 
means that the second skb_header_pointer will either point to SKB (where 
we have at least 20 bytes) or to buf (which is allocated by the caller, 
so the caller shouldn't overflow its own buffer).

On the other hand, parsing garbage doesn't look like a valid behavior 
compared to dropping/ignoring/whatever-cake-does-with-bad-packets, so we 
may want to handle it, for example:

          return skb_header_pointer(skb, offset,
-                                  min(__tcp_hdrlen(tcph), bufsize), buf);
+                                  min(max(sizeof(struct tcphdr), 
__tcp_hdrlen(tcph)), bufsize), buf);

What do you think? Or did I just miss some early check for doff?

(I realize it's egress path and the packets produced by the system 
itself are unlikely to have bad doff, but it's not impossible, for 
example, with AF_PACKET, BPF hooks in tc, etc.)

Maxim Mikityanskiy [off-list ref] writes:

On 2021-06-10 00:51, Toke Høiland-Jørgensen wrote:

quoted

Maxim Mikityanskiy [off-list ref] writes:

quoted

The TCP option parser in cake qdisc (cake_get_tcpopt and
cake_tcph_may_drop) could read one byte out of bounds. When the length
is 1, the execution flow gets into the loop, reads one byte of the
opcode, and if the opcode is neither TCPOPT_EOL nor TCPOPT_NOP, it reads
one more byte, which exceeds the length of 1.

This fix is inspired by commit 9609dad263f8 ("ipv4: tcp_input: fix stack
out of bounds when parsing TCP options.").

Cc: Young Xiao <redacted>
Fixes: 8b7138814f29 ("sch_cake: Add optional ACK filter")
Signed-off-by: Maxim Mikityanskiy <redacted>

Thanks for fixing this!

Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>

Could you also review whether Florian's comment on patch 1 is relevant 
to this patch too? I have concerns about cake_get_tcphdr, which returns 
`skb_header_pointer(skb, offset, min(__tcp_hdrlen(tcph), bufsize), 
buf)`. Although I don't see a way for it to get out of bounds (it will 
read garbage instead of TCP header in the worst case), such code doesn't 
look robust.

It's not possible for it to get out of bounds, because there is a call 
to skb_header_pointer above with sizeof(_tcph), which ensures that the 
SKB has at least 20 bytes after the beginning of the TCP header, which 
means that the second skb_header_pointer will either point to SKB (where 
we have at least 20 bytes) or to buf (which is allocated by the caller, 
so the caller shouldn't overflow its own buffer).

On the other hand, parsing garbage doesn't look like a valid behavior 
compared to dropping/ignoring/whatever-cake-does-with-bad-packets, so we 
may want to handle it, for example:

          return skb_header_pointer(skb, offset,
-                                  min(__tcp_hdrlen(tcph), bufsize), buf);
+                                  min(max(sizeof(struct tcphdr), 
__tcp_hdrlen(tcph)), bufsize), buf);

What do you think? Or did I just miss some early check for doff?

No, I think your analysis is correct: It won't lead to any out-of-bounds
reads, but I suppose we could end up trying to parse garbage. However,
if we do get a packet that sets doff to an invalid value, and we try to
parse it, we're essentially parsing garbage anyway. So I think the fix
should rather be something like:

diff --git a/net/sched/sch_cake.c b/net/sched/sch_cake.c
index 7d37638ee1c7..d312d75ab698 100644
--- a/net/sched/sch_cake.c
+++ b/net/sched/sch_cake.c

@@ -943,7 +943,7 @@ static struct tcphdr *cake_get_tcphdr(const struct sk_buff *skb,
        }
 
        tcph = skb_header_pointer(skb, offset, sizeof(_tcph), &_tcph);
-       if (!tcph)
+       if (!tcph || tcph->doff < 5)
                return NULL;
 
        return skb_header_pointer(skb, offset,

(I realize it's egress path and the packets produced by the system 
itself are unlikely to have bad doff, but it's not impossible, for 
example, with AF_PACKET, BPF hooks in tc, etc.)

Most CAKE deployments primarily handles forwarded packets, and I suppose
malformed TCP packets could make it through the forwarding path as
well...

-Toke