x86/kprobes: Fix crash when probing CS CALL instructionsgrafted

When using eBPF to probe CS CALL instructions within a function,
a crash can be triggered.

The eBPF tool probes offset 257 of the __hrtimer_run_queues()
function:

<__hrtimer_run_queues+249>:  nopl   0x0(%rax,%rax,1)
<__hrtimer_run_queues+254>:  mov    %r14,%rdi
<__hrtimer_run_queues+257>:  cs call <__x86_indirect_thunk_r12>
<__hrtimer_run_queues+263>:  mov    %eax,%r12d
<__hrtimer_run_queues+266>:  xchg   %ax,%ax
<__hrtimer_run_queues+268>:  mov    %r13,%rdi

Which triggers this crash:

  BUG: unable to handle page fault for address: 00000000000f41c9
  #PF: supervisor write access in kernel mode
  #PF: error_code(0x0002) - not-present page
  PGD 0 P4D 0
  Oops: 0002 [#1] SMP NOPTI
  CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P
  RIP: 0010:__hrtimer_run_queues+0x106/0x230

Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is
at the 6th byte of the above CS CALL instruction. Since the CS CALL
instruction occupies 6 bytes, the exception occurred in the middle of that
call instruction.

The root cause is that when using eBPF tools to probe in the middle of a
function, a kprobe with INT3 is used as the underlying implementation.

During single-step emulation of the original CALL instruction,
int3_emulate_call() assumes that the probed CALL instruction is 5 bytes
long. However, the actual CS-prefixed CALL instruction occupies 6 bytes,
so it constructs an incorrect exception return address. When the CPU
returns from the kprobe handler, the next instruction to be executed is at
the address of the last byte of that CS CALL instruction. Coincidentally,
starting from that address, the CPU fetches and decodes a completely
different instruction, which ultimately triggers a kernel crash.

Fix the issue by using the actual instruction length obtained from
the instruction decoder when constructing the exception return
address, rather than relying on the hardcoded CALL_INSN_SIZE macro.

[ mingo: Refined the changelog ]

Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step")
Suggested-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Jinke Han <jinkehan@didiglobal.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Borislav Petkov <bp@alien8.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000

1 files changed, 123 insertions, 0 deletions

diff --git a/include/trace/events/vmalloc.h b/include/trace/events/vmalloc.h
new file mode 100644
index 000000000..ad4e02191
--- /dev/null
+++ b/include/trace/events/vmalloc.h
@@ -0,0 +1,123 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM vmalloc
+
+#if !defined(_TRACE_VMALLOC_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_VMALLOC_H
+
+#include <linux/tracepoint.h>
+
+/**
+ * alloc_vmap_area - called when a new vmap allocation occurs
+ * @addr:	an allocated address
+ * @size:	a requested size
+ * @align:	a requested alignment
+ * @vstart:	a requested start range
+ * @vend:	a requested end range
+ * @failed:	an allocation failed or not
+ *
+ * This event is used for a debug purpose, it can give an extra
+ * information for a developer about how often it occurs and which
+ * parameters are passed for further validation.
+ */
+TRACE_EVENT(alloc_vmap_area,
+
+	TP_PROTO(unsigned long addr, unsigned long size, unsigned long align,
+		unsigned long vstart, unsigned long vend, int failed),
+
+	TP_ARGS(addr, size, align, vstart, vend, failed),
+
+	TP_STRUCT__entry(
+		__field(unsigned long, addr)
+		__field(unsigned long, size)
+		__field(unsigned long, align)
+		__field(unsigned long, vstart)
+		__field(unsigned long, vend)
+		__field(int, failed)
+	),
+
+	TP_fast_assign(
+		__entry->addr = addr;
+		__entry->size = size;
+		__entry->align = align;
+		__entry->vstart = vstart;
+		__entry->vend = vend;
+		__entry->failed = failed;
+	),
+
+	TP_printk("va_start: %lu size=%lu align=%lu vstart=0x%lx vend=0x%lx failed=%d",
+		__entry->addr, __entry->size, __entry->align,
+		__entry->vstart, __entry->vend, __entry->failed)
+);
+
+/**
+ * purge_vmap_area_lazy - called when vmap areas were lazily freed
+ * @start:		purging start address
+ * @end:		purging end address
+ * @npurged:	numbed of purged vmap areas
+ *
+ * This event is used for a debug purpose. It gives some
+ * indication about start:end range and how many objects
+ * are released.
+ */
+TRACE_EVENT(purge_vmap_area_lazy,
+
+	TP_PROTO(unsigned long start, unsigned long end,
+		unsigned int npurged),
+
+	TP_ARGS(start, end, npurged),
+
+	TP_STRUCT__entry(
+		__field(unsigned long, start)
+		__field(unsigned long, end)
+		__field(unsigned int, npurged)
+	),
+
+	TP_fast_assign(
+		__entry->start = start;
+		__entry->end = end;
+		__entry->npurged = npurged;
+	),
+
+	TP_printk("start=0x%lx end=0x%lx num_purged=%u",
+		__entry->start, __entry->end, __entry->npurged)
+);
+
+/**
+ * free_vmap_area_noflush - called when a vmap area is freed
+ * @va_start:		a start address of VA
+ * @nr_lazy:		number of current lazy pages
+ * @nr_lazy_max:	number of maximum lazy pages
+ *
+ * This event is used for a debug purpose. It gives some
+ * indication about a VA that is released, number of current
+ * outstanding areas and a maximum allowed threshold before
+ * dropping all of them.
+ */
+TRACE_EVENT(free_vmap_area_noflush,
+
+	TP_PROTO(unsigned long va_start, unsigned long nr_lazy,
+		unsigned long nr_lazy_max),
+
+	TP_ARGS(va_start, nr_lazy, nr_lazy_max),
+
+	TP_STRUCT__entry(
+		__field(unsigned long, va_start)
+		__field(unsigned long, nr_lazy)
+		__field(unsigned long, nr_lazy_max)
+	),
+
+	TP_fast_assign(
+		__entry->va_start = va_start;
+		__entry->nr_lazy = nr_lazy;
+		__entry->nr_lazy_max = nr_lazy_max;
+	),
+
+	TP_printk("va_start=0x%lx nr_lazy=%lu nr_lazy_max=%lu",
+		__entry->va_start, __entry->nr_lazy, __entry->nr_lazy_max)
+);
+
+#endif /*  _TRACE_VMALLOC_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>