PT-2026-152553
CVE-2026-99673
·
Published
2026-11-20
·
Updated
2026-11-20
Description by
Score
N/A
Not availableNo severity ratings or metrics are available. When they are, we'll update the corresponding info on the page.
In the Linux kernel, the following vulnerability has been resolved:
can: replace ARPHRD CAN type checks with can get ml priv()
Commit 4e096a18867a ("net: introduce CAN specific pointer in the struct
net device") introduced an explicit way to assign the midlayer private
pointer (dev->ml priv) to named users like ML PRIV CAN. A valid pointer
returned by can get ml priv() therefore identifies a real CAN device.
The referenced commit only converted the j1939 and proc code to this
check, while the other parts of the CAN subsystem still rely on
dev->type == ARPHRD CAN.
The TUNSETLINK ioctl allows a down TUN/TAP device to set its hardware
type to ARPHRD CAN while dev->ml priv remains NULL. Such a repurposed
netdevice passes the ARPHRD CAN checks.
Commit 8aa59e355949 ("can: af can: fix NULL pointer dereference in
can rx register()") already made can rx register() reject such devices.
But can rx unregister() still only checks dev->type and dereferences
the NULL ml priv when it walks the receive lists of the device.
CAN ISOTP reaches this path: isotp bind() ignores the return values of
can rx register() and binds the socket anyway. When the socket is
released or the device is unregistered, isotp release() and
isotp notify() remove the filters that have never been registered and
trigger the NULL pointer dereference in can rx unregister(). This only
requires owning a TUN/TAP device or CAP NET ADMIN in an unprivileged
user namespace.
Convert the remaining ARPHRD CAN checks in af can, bcm, gw, isotp and
raw to can get ml priv().
Found an issue in the description? Have something to add? Feel free to write us 👾