PT-2026-117031 · Linux · Linux
CVE-2026-98812
·
Published
2026-10-18
·
Updated
2026-10-18
None
No 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:
net/mctp: publish the mctp dev only after it is initialised
mctp add dev() links the new mctp dev into dev->mctp ptr before mdev->dev
is assigned, so a reader that picks the device up in that window
dereferences a net device pointer that is still NULL:
KASAN: null-ptr-deref in range [0x00000000000000b0-0x00000000000000b7]
Call Trace:
? pfx mctp sendmsg (./include/linux/sockptr.h:49)
? pfx mctp dst output (net/mctp/route.c:41)
? selinux socket sendmsg (security/selinux/hooks.c:5278)
sys sendmsg (net/socket.c:775 (discriminator 1) net/socket.c:790
(discriminator 1) net/socket.c:2684 (discriminator 1))
? pfx sys sendmsg (net/socket.c:1131)
? pfx copy msghdr from user (net/socket.c:2590)
? update cfs rq load avg (kernel/sched/fair.c:5478)
sys sendmsg (net/socket.c:2738)
? pfx sys sendmsg (net/socket.c:2625)
? perf event task tick (./include/linux/rcupdate.h:838)
? sched tick (kernel/sched/core.c:5818)
? clockevents program event (kernel/time/clockevents.c:372)
? fdget (./include/linux/rcupdate.h:873 fs/file.c:1100)
sys sendmsg (net/socket.c:2770)
? pfx sys sendmsg (net/socket.c:2751)
do syscall 64 (arch/x86/entry/syscall 64.c:63)
entry SYSCALL 64 after hwframe (arch/x86/entry/entry 64.S:121)
Fix by publishing only once the object is fully initialised:
mctp dev get() runs under rcu read lock() only and hands
this object to readers as soon as mctp ptr is assigned, so
they must never observe mdev->dev == NULL.
Found an issue in the description? Have something to add? Feel free to write us 👾
Related Identifiers
Affected Products
Linux