[PATCH net v2] net: usb: catc: bound the RX packet length in catc_rx_done()

[PATCH net v2] net: usb: catc: bound the RX packet length in catc_rx_done()

[Aamir Ahmed]

catc_rx_done() walks a multi-packet URB, reading a two-byte length from
each packet header. Its bound, pkt_len > urb->actual_length, ignores the
header offset and compares against the whole transfer rather than the
bytes left from pkt_start, so a crafted packet header makes
skb_copy_to_linear_data() read past the buffer.

A length below ETH_HLEN is also accepted, including zero, and
eth_type_trans() then reads a MAC header from the uninitialised tailroom
of a shorter skb. The is_f5u011 branch takes its length straight from
the transfer, so a zero-length URB reaches the same path.

Track the bytes remaining from the current packet, and reject a header
that does not fit, a length past what is left, and a length below an
Ethernet header.

A transfer shorter than an Ethernet header, including a zero-length one,
previously became a runt skb passed to netif_rx() and counted as
received; it is now counted in rx_length_errors and ends the walk.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Assisted-by: LLM
Signed-off-by: Aamir Ahmed <elb12345@hotmail.co.uk>
---
v2:
  - reject pkt_len below ETH_HLEN, covering the is_f5u011 branch too (Sashiko)
  - hoist the remaining-bytes calculation so one check bounds both ends
  - drop Cc: stable; no conforming device reaches this
  - add Assisted-by: LLM
v1: https://lore.kernel.org/netdev/AS8P251MB00013A0DCA600A79DC58B0FEC8B22@AS8P251MB0001.EURP251.PROD.OUTLOOK.COM/

Sashiko also suggested a sanity maximum on pkt_len. I left it out: PKT_SZ
would reject the 1537-1600 byte transfers the is_f5u011 path allows, since
RX_PKT_SZ is 1600, and dev->mtu excludes the MAC header that pkt_len
includes.

Built with W=1 (catc.o) on x86_64; no warnings. I have no CATC hardware,
so this is not runtime-tested.

 drivers/net/usb/catc.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/drivers/net/usb/catc.c b/drivers/net/usb/catc.c
index 96e82f94edcf..39b678f175dd 100644
--- a/drivers/net/usb/catc.c
+++ b/drivers/net/usb/catc.c
@@ -233,17 +233,26 @@ static void catc_rx_done(struct urb *urb)
 	}
 
 	do {
-		if(!catc->is_f5u011) {
-			pkt_len = le16_to_cpup((__le16*)pkt_start);
-			if (pkt_len > urb->actual_length) {
+		int remaining = urb->actual_length -
+				(pkt_start - (u8 *)urb->transfer_buffer);
+
+		if (!catc->is_f5u011) {
+			if (remaining < pkt_offset) {
 				catc->netdev->stats.rx_length_errors++;
 				catc->netdev->stats.rx_errors++;
 				break;
 			}
+			pkt_len = le16_to_cpup((__le16 *)pkt_start);
 		} else {
 			pkt_len = urb->actual_length;
 		}
 
+		if (pkt_len < ETH_HLEN || pkt_len + pkt_offset > remaining) {
+			catc->netdev->stats.rx_length_errors++;
+			catc->netdev->stats.rx_errors++;
+			break;
+		}
+
 		if (!(skb = dev_alloc_skb(pkt_len)))
 			return;
 

base-commit: c297ed90fbba72d32b7759aae362b36d15b2db1f
-- 
2.43.0

Re: [PATCH net v2] net: usb: catc: bound the RX packet length in catc_rx_done()

[Simon Horman]

On Tue, Sep 15, 2026 at 12:06:58AM +0100, Aamir Ahmed wrote:
> catc_rx_done() walks a multi-packet URB, reading a two-byte length from
> each packet header. Its bound, pkt_len > urb->actual_length, ignores the
> header offset and compares against the whole transfer rather than the
> bytes left from pkt_start, so a crafted packet header makes
> skb_copy_to_linear_data() read past the buffer.
> 
> A length below ETH_HLEN is also accepted, including zero, and
> eth_type_trans() then reads a MAC header from the uninitialised tailroom
> of a shorter skb. The is_f5u011 branch takes its length straight from
> the transfer, so a zero-length URB reaches the same path.
> 
> Track the bytes remaining from the current packet, and reject a header
> that does not fit, a length past what is left, and a length below an
> Ethernet header.
> 
> A transfer shorter than an Ethernet header, including a zero-length one,
> previously became a runt skb passed to netif_rx() and counted as
> received; it is now counted in rx_length_errors and ends the walk.
> 
> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> Assisted-by: LLM
> Signed-off-by: Aamir Ahmed <elb12345@hotmail.co.uk>
> ---
> v2:
>   - reject pkt_len below ETH_HLEN, covering the is_f5u011 branch too (Sashiko)
>   - hoist the remaining-bytes calculation so one check bounds both ends
>   - drop Cc: stable; no conforming device reaches this
>   - add Assisted-by: LLM
> v1: https://lore.kernel.org/netdev/AS8P251MB00013A0DCA600A79DC58B0FEC8B22@AS8P251MB0001.EURP251.PROD.OUTLOOK.COM/
> 
> Sashiko also suggested a sanity maximum on pkt_len. I left it out: PKT_SZ
> would reject the 1537-1600 byte transfers the is_f5u011 path allows, since
> RX_PKT_SZ is 1600, and dev->mtu excludes the MAC header that pkt_len
> includes.
> 
> Built with W=1 (catc.o) on x86_64; no warnings. I have no CATC hardware,
> so this is not runtime-tested.

Reviewed-by: Simon Horman <horms@kernel.org>

Re: [PATCH net v2] net: usb: catc: bound the RX packet length in catc_rx_done()

[Paolo Abeni]

On 9/15/26 01:06, Aamir Ahmed wrote:
> catc_rx_done() walks a multi-packet URB, reading a two-byte length from
> each packet header. Its bound, pkt_len > urb->actual_length, ignores the
> header offset and compares against the whole transfer rather than the
> bytes left from pkt_start, so a crafted packet header makes
> skb_copy_to_linear_data() read past the buffer.
> 
> A length below ETH_HLEN is also accepted, including zero, and
> eth_type_trans() then reads a MAC header from the uninitialised tailroom
> of a shorter skb. The is_f5u011 branch takes its length straight from
> the transfer, so a zero-length URB reaches the same path.
> 
> Track the bytes remaining from the current packet, and reject a header
> that does not fit, a length past what is left, and a length below an
> Ethernet header.
> 
> A transfer shorter than an Ethernet header, including a zero-length one,
> previously became a runt skb passed to netif_rx() and counted as
> received; it is now counted in rx_length_errors and ends the walk.
> 
> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> Assisted-by: LLM
> Signed-off-by: Aamir Ahmed <elb12345@hotmail.co.uk>
I'm sorry, but no minor fixes for legacy drivers.

/P