ACK: Re: [SRU][J][PATCH 0/1] CVE-2022-49390
Paolo Pisati
paolo.pisati at canonical.com
Tue Oct 14 14:43:31 UTC 2025
On Tue, Oct 14, 2025 at 04:20:46PM +0200, Massimiliano Pellizzer wrote:
> https://ubuntu.com/security/CVE-2022-49390
>
> [ Impact ]
>
> macsec: fix UAF bug for real_dev
>
> Creating a new macsec device without getting a reference to real_dev may
> trigger a use-after-free bug.
>
> [ Fix ]
>
> Backport commit 2bce1ebed17d (macsec: fix refcnt leak in module exit routine)
> from mainline.
>
> [ Test Plan ]
>
> Compile and boot tested.
> Tested basic macsec functionalities:
>
> $ unshare --map-root-user --net
> # ip link add dummy0 type dummy
> # ip link set dummy0 up
> # ip link add link dummy0 name macsec0 type macsec
> # ip link set macsec0 up
> # ip a
> 1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN group default qlen 1000
> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
> 2: dummy0: <BROADCAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN group default qlen 1000
> link/ether aa:ca:71:b5:0f:dd brd ff:ff:ff:ff:ff:ff
> inet6 fe80::a8ca:71ff:feb5:fdd/64 scope link
> valid_lft forever preferred_lft forever
> 3: macsec0 at dummy0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1468 qdisc noqueue state UP group default qlen 1000
> link/ether aa:ca:71:b5:0f:dd brd ff:ff:ff:ff:ff:ff
> inet6 fe80::a8ca:71ff:feb5:fdd/64 scope link tentative
> valid_lft forever preferred_lft forever
> # ip link del dummy0
> # ip a
> 1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN group default qlen 1000
> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
>
> [ Regression Potential ]
>
> The fix affects macsec's device handling of the lower (real) net_device
> lifetime. An issue with this patch may introduce refcount leaks that
> prevent lower devices from being freed, or incorrect release ordering
> that re-introduces use-after-free and breaks interface teardown.
>
> Ziyang Xuan (1):
> macsec: fix UAF bug for real_dev
>
> drivers/net/macsec.c | 5 +++++
> 1 file changed, 5 insertions(+)
Acked-by: Paolo Pisati <paolo.pisati at canonical.com>
--
bye,
p.
More information about the kernel-team
mailing list