In the Linux kernel, the following vulnerability has been resolved:
dmaengine: fsl-edma: tracing: no ptr dereference during log output
The fsl edma events store a pointer to a struct fsl_edma_engine in the
ringbuffer and dereference it when a log entry is printed. At this time,
the pointer may no longer be valid.
Event injection can be used to trigger a crash:
$ cd /sys/kernel/tracing
$ echo 'value = 0' > events/fsl_edma/edma_writeb/inject
$ cat trace
The log output needs only edma->membase. Add a membase field at the end
of the event and use the new field for log output. Keep the existing
fields for backward compatibility.
CVSS Vector: CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
CVSS Score: 7.8
AV:L - The stale dereference runs only when a local actor reads formatted fsl_edma events from tracefs (/sys/kernel/tracing/trace or trace_pipe). DMA MMIO tracing and event injection only enqueue ring-buffer records; there is no network, adjacent-radio, or physical-device path into TP_printk().
AC:L - An attacker with tracefs access controls the full sequence: enable edma_log_io events, either inject a record via events/fsl_edma/*/inject or generate eDMA I/O then unbind/remove fsl-edma so the stored engine pointer is freed, then read trace. No race or uninfluenced victim state is required.
PR:L - tracing_check_open_get_tr() enforces only LOCKDOWN_TRACEFS and DAC, with no capable() check. Event enable and trace files are mode 0640 and are routinely delegated via tracefs gid= to tracing-group members on Android Automotive/NXP i.MX, ChromeOS, and developer kernels, matching prior TP_printk UAF CVEs.
UI:N - No victim interaction is required. The attacker enables fsl_edma events, injects or records a stale edma pointer, and reads the formatted trace themselves to invoke the vulnerable TP_printk() path.
S:U - Impact stays inside the host kernel's trace formatting path (stale fsl_edma_engine dereference). This is standard local kernel memory-safety impact with no VM escape, IOMMU/DMA boundary bypass, or other cross-authority effect; membase is used only for log offset arithmetic.
C:H - TP_printk() dereferences __entry->edma->membase when formatting output. That is a use-after-free read of a freed fsl_edma_engine, and with event injection the edma field is attacker-chosen, yielding an arbitrary kernel pointer read whose result is printed as the register offset.
I:H - The bug is a use-after-free of the kernel heap object fsl_edma_engine, which embeds dma_device function pointers. Freed memory can be reclaimed before the trace read, enabling heap grooming and write/control-flow primitives; kernel CNA guidance rates UAF integrity impact as High.
A:H - Dereferencing a NULL (inject with a zeroed edma pointer) or freed fsl_edma_engine pointer during trace formatting causes a kernel oops or panic. The fix commit demonstrates a deterministic crash via inject followed by cat trace.
| Attack Vector |
Local |
Scope |
Unchanged |
| Attack Complexity |
Low |
Confidentiality Impact |
High |
| Privileges Required |
Low |
Integrity Impact |
High |
| User Interaction |
None |
Availability Impact |
High |
AV:L - The stale dereference runs only when a local actor reads formatted fsl_edma events from tracefs (/sys/kernel/tracing/trace or trace_pipe). DMA MMIO tracing and event injection only enqueue ring-buffer records; there is no network, adjacent-radio, or physical-device path into TP_printk().
AC:L - An attacker with tracefs access controls the full sequence: enable edma_log_io events, either inject a record via events/fsl_edma/*/inject or generate eDMA I/O then unbind/remove fsl-edma so the stored engine pointer is freed, then read trace. No race or uninfluenced victim state is required.
PR:L - tracing_check_open_get_tr() enforces only LOCKDOWN_TRACEFS and DAC, with no capable() check. Event enable and trace files are mode 0640 and are routinely delegated via tracefs gid= to tracing-group members on Android Automotive/NXP i.MX, ChromeOS, and developer kernels, matching prior TP_printk UAF CVEs.
UI:N - No victim interaction is required. The attacker enables fsl_edma events, injects or records a stale edma pointer, and reads the formatted trace themselves to invoke the vulnerable TP_printk() path.
S:U - Impact stays inside the host kernel's trace formatting path (stale fsl_edma_engine dereference). This is standard local kernel memory-safety impact with no VM escape, IOMMU/DMA boundary bypass, or other cross-authority effect; membase is used only for log offset arithmetic.
C:H - TP_printk() dereferences __entry->edma->membase when formatting output. That is a use-after-free read of a freed fsl_edma_engine, and with event injection the edma field is attacker-chosen, yielding an arbitrary kernel pointer read whose result is printed as the register offset.
I:H - The bug is a use-after-free of the kernel heap object fsl_edma_engine, which embeds dma_device function pointers. Freed memory can be reclaimed before the trace read, enabling heap grooming and write/control-flow primitives; kernel CNA guidance rates UAF integrity impact as High.
A:H - Dereferencing a NULL (inject with a zeroed edma pointer) or freed fsl_edma_engine pointer during trace formatting causes a kernel oops or panic. The fix commit demonstrates a deterministic crash via inject followed by cat trace.
CVSS 3.1