CVE-2026-97934: tracing: Fix memory corruption from a "STACKTRACE" histogram key
In the Linux kernel, the following vulnerability has been resolved:
tracing: Fix memory corruption from a "STACKTRACE" histogram key
"cpu", "CPU", "stacktrace" and "STACKTRACE" are generic fields, defined with an offset and a size of zero so that the filter code can match them by name. parsefield() maps them onto their common equivalents for backward compatibility, but unlike the common names it hands the placeholder back to the caller instead of NULL.
createhistfield() takes a non-NULL field as a promise that the record carries a stacktrace and picks HISTFIELDFNSTACK, so the dataloc word is read from offset 0, that is from commontype, and its low 16 bits are followed as an offset into the record. What is found there becomes the length of an unbounded memcpy. Pick an event whose id is small enough that the offset stays inside its own record and the length is a kernel text address:
# cd /sys/kernel/tracing # echo 'hist:keys=STACKTRACE' > events/ftrace/print/trigger # echo hello > tracemarker
Oops: general protection fault, probably for non-canonical address RIP: 0010:rbnext+0x23/0x60 </IRQ> RIP: 0010:memcpy+0xc/0x30 eventhisttrigger+0x2e7/0x12c0 Kernel panic - not syncing: Fatal exception in interrupt
Leave the field NULL, which is what the comment above the branch says the code does and what commonstacktrace already does. FILTERCPU and FILTERCOMM are left alone, their createhistfield() branches never look at the field.
Affected Software
Event History
Frequently Asked Questions
What access is needed to trigger this issue?
An attacker needs the ability to create or modify a tracing histogram trigger and supply STACKTRACE as a histogram key. The demonstrated trigger is written under /sys/kernel/tracing/events/ftrace/print/trigger, then activated by writing an event to trace_marker.
What is the likely impact of successful exploitation?
The malformed histogram key causes an unbounded memcpy using data derived from the trace record. The provided reproduction results in a general protection fault and kernel panic, so systems may be vulnerable to denial of service.
What can be done if an updated kernel cannot be deployed immediately?
Do not configure tracing histogram triggers with STACKTRACE as a key, and restrict access to the tracing interface so untrusted users cannot create or alter triggers. Review existing triggers, particularly the ftrace print event trigger, for histogram definitions using STACKTRACE.