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, 124 insertions, 0 deletions

diff --git a/arch/alpha/boot/head.S b/arch/alpha/boot/head.S
new file mode 100644
index 000000000..06a7c95fe
--- /dev/null
+++ b/arch/alpha/boot/head.S
@@ -0,0 +1,124 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * arch/alpha/boot/head.S
+ *
+ * initial bootloader stuff..
+ */
+
+#include <asm/pal.h>
+
+	.set noreorder
+	.globl	__start
+	.ent	__start
+__start:
+	br	$29,2f
+2:	ldgp	$29,0($29)
+	jsr	$26,start_kernel
+	call_pal PAL_halt
+	.end __start
+
+	.align 5
+	.globl	wrent
+	.ent	wrent
+wrent:
+	.prologue 0
+	call_pal PAL_wrent
+	ret ($26)
+	.end wrent
+
+	.align 5
+	.globl	wrkgp
+	.ent	wrkgp
+wrkgp:
+	.prologue 0
+	call_pal PAL_wrkgp
+	ret ($26)
+	.end wrkgp
+
+	.align 5
+	.globl	switch_to_osf_pal
+	.ent	switch_to_osf_pal
+switch_to_osf_pal:
+	subq	$30,128,$30
+	.frame	$30,128,$26
+	stq	$26,0($30)
+	stq	$1,8($30)
+	stq	$2,16($30)
+	stq	$3,24($30)
+	stq	$4,32($30)
+	stq	$5,40($30)
+	stq	$6,48($30)
+	stq	$7,56($30)
+	stq	$8,64($30)
+	stq	$9,72($30)
+	stq	$10,80($30)
+	stq	$11,88($30)
+	stq	$12,96($30)
+	stq	$13,104($30)
+	stq	$14,112($30)
+	stq	$15,120($30)
+	.prologue 0
+
+	stq	$30,0($17)	/* save KSP in PCB */
+
+	bis	$30,$30,$20	/* a4 = KSP */
+	br	$17,1f
+
+	ldq	$26,0($30)
+	ldq	$1,8($30)
+	ldq	$2,16($30)
+	ldq	$3,24($30)
+	ldq	$4,32($30)
+	ldq	$5,40($30)
+	ldq	$6,48($30)
+	ldq	$7,56($30)
+	ldq	$8,64($30)
+	ldq	$9,72($30)
+	ldq	$10,80($30)
+	ldq	$11,88($30)
+	ldq	$12,96($30)
+	ldq	$13,104($30)
+	ldq	$14,112($30)
+	ldq	$15,120($30)
+	addq	$30,128,$30
+	ret ($26)
+1:	call_pal PAL_swppal
+	.end	switch_to_osf_pal
+
+	.align 3
+	.globl	tbi
+	.ent	tbi
+tbi:
+	.prologue 0
+	call_pal PAL_tbi
+	ret	($26)
+	.end tbi
+
+	.align 3
+	.globl	halt
+	.ent	halt
+halt:
+	.prologue 0
+	call_pal PAL_halt
+	.end halt
+
+/* $16 - new stack page */
+	.align 3
+	.globl	move_stack
+	.ent	move_stack
+move_stack:
+	.prologue 0
+	lda	$0, 0x1fff($31)
+	and	$0, $30, $1			/* Stack offset */
+	or	$1, $16, $16			/* New stack pointer */
+	mov	$30, $1
+	mov	$16, $2
+1:	ldq	$3, 0($1)			/* Move the stack */
+	addq	$1, 8, $1
+	stq	$3, 0($2)
+	and	$0, $1, $4
+	addq	$2, 8, $2
+	bne	$4, 1b
+	mov	$16, $30
+	ret	($26)
+	.end move_stack