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, 298 insertions, 0 deletions

diff --git a/net/mac802154/trace.h b/net/mac802154/trace.h
new file mode 100644
index 000000000..1574ecc48
--- /dev/null
+++ b/net/mac802154/trace.h
@@ -0,0 +1,298 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Based on net/mac80211/trace.h */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM mac802154
+
+#if !defined(__MAC802154_DRIVER_TRACE) || defined(TRACE_HEADER_MULTI_READ)
+#define __MAC802154_DRIVER_TRACE
+
+#include <linux/tracepoint.h>
+
+#include <net/mac802154.h>
+#include "ieee802154_i.h"
+
+#define MAXNAME		32
+#define LOCAL_ENTRY	__array(char, wpan_phy_name, MAXNAME)
+#define LOCAL_ASSIGN	strscpy(__entry->wpan_phy_name, \
+				wpan_phy_name(local->hw.phy), MAXNAME)
+#define LOCAL_PR_FMT	"%s"
+#define LOCAL_PR_ARG	__entry->wpan_phy_name
+
+#define CCA_ENTRY __field(enum nl802154_cca_modes, cca_mode) \
+		  __field(enum nl802154_cca_opts, cca_opt)
+#define CCA_ASSIGN \
+	do {                                     \
+		(__entry->cca_mode) = cca->mode; \
+		(__entry->cca_opt) = cca->opt;   \
+	} while (0)
+#define CCA_PR_FMT "cca_mode: %d, cca_opt: %d"
+#define CCA_PR_ARG __entry->cca_mode, __entry->cca_opt
+
+#define BOOL_TO_STR(bo) (bo) ? "true" : "false"
+
+/* Tracing for driver callbacks */
+
+DECLARE_EVENT_CLASS(local_only_evt4,
+	TP_PROTO(struct ieee802154_local *local),
+	TP_ARGS(local),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+	),
+	TP_printk(LOCAL_PR_FMT, LOCAL_PR_ARG)
+);
+
+DEFINE_EVENT(local_only_evt4, 802154_drv_return_void,
+	TP_PROTO(struct ieee802154_local *local),
+	TP_ARGS(local)
+);
+
+TRACE_EVENT(802154_drv_return_int,
+	TP_PROTO(struct ieee802154_local *local, int ret),
+	TP_ARGS(local, ret),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(int, ret)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->ret = ret;
+	),
+	TP_printk(LOCAL_PR_FMT ", returned: %d", LOCAL_PR_ARG,
+		  __entry->ret)
+);
+
+DEFINE_EVENT(local_only_evt4, 802154_drv_start,
+	TP_PROTO(struct ieee802154_local *local),
+	TP_ARGS(local)
+);
+
+DEFINE_EVENT(local_only_evt4, 802154_drv_stop,
+	TP_PROTO(struct ieee802154_local *local),
+	TP_ARGS(local)
+);
+
+TRACE_EVENT(802154_drv_set_channel,
+	TP_PROTO(struct ieee802154_local *local, u8 page, u8 channel),
+	TP_ARGS(local, page, channel),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(u8, page)
+		__field(u8, channel)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->page = page;
+		__entry->channel = channel;
+	),
+	TP_printk(LOCAL_PR_FMT ", page: %d, channel: %d", LOCAL_PR_ARG,
+		  __entry->page, __entry->channel)
+);
+
+TRACE_EVENT(802154_drv_set_cca_mode,
+	TP_PROTO(struct ieee802154_local *local,
+		 const struct wpan_phy_cca *cca),
+	TP_ARGS(local, cca),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		CCA_ENTRY
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		CCA_ASSIGN;
+	),
+	TP_printk(LOCAL_PR_FMT ", " CCA_PR_FMT, LOCAL_PR_ARG,
+		  CCA_PR_ARG)
+);
+
+TRACE_EVENT(802154_drv_set_cca_ed_level,
+	TP_PROTO(struct ieee802154_local *local, s32 mbm),
+	TP_ARGS(local, mbm),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(s32, mbm)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->mbm = mbm;
+	),
+	TP_printk(LOCAL_PR_FMT ", ed level: %d", LOCAL_PR_ARG,
+		  __entry->mbm)
+);
+
+TRACE_EVENT(802154_drv_set_tx_power,
+	TP_PROTO(struct ieee802154_local *local, s32 power),
+	TP_ARGS(local, power),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(s32, power)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->power = power;
+	),
+	TP_printk(LOCAL_PR_FMT ", mbm: %d", LOCAL_PR_ARG,
+		 __entry->power)
+);
+
+TRACE_EVENT(802154_drv_set_lbt_mode,
+	TP_PROTO(struct ieee802154_local *local, bool mode),
+	TP_ARGS(local, mode),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(bool, mode)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->mode = mode;
+	),
+	TP_printk(LOCAL_PR_FMT ", lbt mode: %s", LOCAL_PR_ARG,
+		  BOOL_TO_STR(__entry->mode))
+);
+
+TRACE_EVENT(802154_drv_set_short_addr,
+	TP_PROTO(struct ieee802154_local *local, __le16 short_addr),
+	TP_ARGS(local, short_addr),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(__le16, short_addr)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->short_addr = short_addr;
+	),
+	TP_printk(LOCAL_PR_FMT ", short addr: 0x%04x", LOCAL_PR_ARG,
+		  le16_to_cpu(__entry->short_addr))
+);
+
+TRACE_EVENT(802154_drv_set_pan_id,
+	TP_PROTO(struct ieee802154_local *local, __le16 pan_id),
+	TP_ARGS(local, pan_id),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(__le16, pan_id)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->pan_id = pan_id;
+	),
+	TP_printk(LOCAL_PR_FMT ", pan id: 0x%04x", LOCAL_PR_ARG,
+		  le16_to_cpu(__entry->pan_id))
+);
+
+TRACE_EVENT(802154_drv_set_extended_addr,
+	TP_PROTO(struct ieee802154_local *local, __le64 extended_addr),
+	TP_ARGS(local, extended_addr),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(__le64, extended_addr)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->extended_addr = extended_addr;
+	),
+	TP_printk(LOCAL_PR_FMT ", extended addr: 0x%llx", LOCAL_PR_ARG,
+		  le64_to_cpu(__entry->extended_addr))
+);
+
+TRACE_EVENT(802154_drv_set_pan_coord,
+	TP_PROTO(struct ieee802154_local *local, bool is_coord),
+	TP_ARGS(local, is_coord),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(bool, is_coord)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->is_coord = is_coord;
+	),
+	TP_printk(LOCAL_PR_FMT ", is_coord: %s", LOCAL_PR_ARG,
+		  BOOL_TO_STR(__entry->is_coord))
+);
+
+TRACE_EVENT(802154_drv_set_csma_params,
+	TP_PROTO(struct ieee802154_local *local, u8 min_be, u8 max_be,
+		 u8 max_csma_backoffs),
+	TP_ARGS(local, min_be, max_be, max_csma_backoffs),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(u8, min_be)
+		__field(u8, max_be)
+		__field(u8, max_csma_backoffs)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN,
+		__entry->min_be = min_be;
+		__entry->max_be = max_be;
+		__entry->max_csma_backoffs = max_csma_backoffs;
+	),
+	TP_printk(LOCAL_PR_FMT ", min be: %d, max be: %d, max csma backoffs: %d",
+		  LOCAL_PR_ARG, __entry->min_be, __entry->max_be,
+		  __entry->max_csma_backoffs)
+);
+
+TRACE_EVENT(802154_drv_set_max_frame_retries,
+	TP_PROTO(struct ieee802154_local *local, s8 max_frame_retries),
+	TP_ARGS(local, max_frame_retries),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(s8, max_frame_retries)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->max_frame_retries = max_frame_retries;
+	),
+	TP_printk(LOCAL_PR_FMT ", max frame retries: %d", LOCAL_PR_ARG,
+		  __entry->max_frame_retries)
+);
+
+TRACE_EVENT(802154_drv_set_promiscuous_mode,
+	TP_PROTO(struct ieee802154_local *local, bool on),
+	TP_ARGS(local, on),
+	TP_STRUCT__entry(
+		LOCAL_ENTRY
+		__field(bool, on)
+	),
+	TP_fast_assign(
+		LOCAL_ASSIGN;
+		__entry->on = on;
+	),
+	TP_printk(LOCAL_PR_FMT ", promiscuous mode: %s", LOCAL_PR_ARG,
+		  BOOL_TO_STR(__entry->on))
+);
+
+TRACE_EVENT(802154_new_scan_event,
+	TP_PROTO(struct ieee802154_coord_desc *desc),
+	TP_ARGS(desc),
+	TP_STRUCT__entry(
+		__field(__le16, pan_id)
+		__field(__le64, addr)
+		__field(u8, channel)
+		__field(u8, page)
+	),
+	TP_fast_assign(
+		__entry->page = desc->page;
+		__entry->channel = desc->channel;
+		__entry->pan_id = desc->addr.pan_id;
+		__entry->addr = desc->addr.extended_addr;
+	),
+	TP_printk("panid: %u, coord_addr: 0x%llx, page: %u, channel: %u",
+		  __le16_to_cpu(__entry->pan_id), __le64_to_cpu(__entry->addr),
+		  __entry->page, __entry->channel)
+);
+
+DEFINE_EVENT(802154_new_scan_event, 802154_scan_event,
+	TP_PROTO(struct ieee802154_coord_desc *desc),
+	TP_ARGS(desc)
+);
+
+#endif /* !__MAC802154_DRIVER_TRACE || TRACE_HEADER_MULTI_READ */
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE trace
+#include <trace/define_trace.h>