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, 152 insertions, 0 deletions

diff --git a/arch/csky/abiv2/memcmp.S b/arch/csky/abiv2/memcmp.S
new file mode 100644
index 000000000..bf0d809f0
--- /dev/null
+++ b/arch/csky/abiv2/memcmp.S
@@ -0,0 +1,152 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include "sysdep.h"
+
+ENTRY(memcmp)
+	/* Test if len less than 4 bytes.  */
+	mov	r3, r0
+	movi	r0, 0
+	mov	r12, r4
+	cmplti	r2, 4
+	bt	.L_compare_by_byte
+
+	andi	r13, r0, 3
+	movi	r19, 4
+
+	/* Test if s1 is not 4 bytes aligned.  */
+	bnez	r13, .L_s1_not_aligned
+
+	LABLE_ALIGN
+.L_s1_aligned:
+	/* If dest is aligned, then copy.  */
+	zext	r18, r2, 31, 4
+	/* Test if len less than 16 bytes.  */
+	bez	r18, .L_compare_by_word
+
+.L_compare_by_4word:
+	/* If aligned, load word each time.  */
+	ldw	r20, (r3, 0)
+	ldw	r21, (r1, 0)
+	/* If s1[i] != s2[i], goto .L_byte_check.  */
+	cmpne	r20, r21
+	bt	.L_byte_check
+
+	ldw	r20, (r3, 4)
+	ldw	r21, (r1, 4)
+	cmpne	r20, r21
+	bt	.L_byte_check
+
+	ldw	r20, (r3, 8)
+	ldw	r21, (r1, 8)
+	cmpne	r20, r21
+	bt	.L_byte_check
+
+	ldw	r20, (r3, 12)
+	ldw	r21, (r1, 12)
+	cmpne	r20, r21
+	bt	.L_byte_check
+
+	PRE_BNEZAD (r18)
+	addi	a3, 16
+	addi	a1, 16
+
+	BNEZAD (r18, .L_compare_by_4word)
+
+.L_compare_by_word:
+	zext	r18, r2, 3, 2
+	bez	r18, .L_compare_by_byte
+.L_compare_by_word_loop:
+	ldw	r20, (r3, 0)
+	ldw	r21, (r1, 0)
+	addi	r3, 4
+	PRE_BNEZAD (r18)
+	cmpne	r20, r21
+	addi    r1, 4
+	bt	.L_byte_check
+	BNEZAD (r18, .L_compare_by_word_loop)
+
+.L_compare_by_byte:
+        zext    r18, r2, 1, 0
+        bez     r18, .L_return
+.L_compare_by_byte_loop:
+        ldb     r0, (r3, 0)
+        ldb     r4, (r1, 0)
+        addi    r3, 1
+        subu    r0, r4
+        PRE_BNEZAD (r18)
+        addi    r1, 1
+        bnez    r0, .L_return
+        BNEZAD (r18, .L_compare_by_byte_loop)
+
+.L_return:
+        mov     r4, r12
+        rts
+
+# ifdef __CSKYBE__
+/* d[i] != s[i] in word, so we check byte 0.  */
+.L_byte_check:
+        xtrb0   r0, r20
+        xtrb0   r2, r21
+        subu    r0, r2
+        bnez    r0, .L_return
+
+        /* check byte 1 */
+        xtrb1   r0, r20
+        xtrb1   r2, r21
+        subu    r0, r2
+        bnez    r0, .L_return
+
+        /* check byte 2 */
+        xtrb2   r0, r20
+        xtrb2   r2, r21
+        subu    r0, r2
+        bnez    r0, .L_return
+
+        /* check byte 3 */
+        xtrb3   r0, r20
+        xtrb3   r2, r21
+        subu    r0, r2
+# else
+/* s1[i] != s2[i] in word, so we check byte 3.  */
+.L_byte_check:
+	xtrb3	r0, r20
+	xtrb3	r2, r21
+        subu    r0, r2
+        bnez    r0, .L_return
+
+	/* check byte 2 */
+	xtrb2	r0, r20
+	xtrb2	r2, r21
+        subu    r0, r2
+        bnez    r0, .L_return
+
+	/* check byte 1 */
+	xtrb1	r0, r20
+	xtrb1	r2, r21
+	subu	r0, r2
+	bnez    r0, .L_return
+
+	/* check byte 0 */
+	xtrb0	r0, r20
+	xtrb0	r2, r21
+	subu	r0, r2
+	br	.L_return
+# endif /* !__CSKYBE__ */
+
+/* Compare when s1 is not aligned.  */
+.L_s1_not_aligned:
+	sub	r13, r19, r13
+	sub	r2, r13
+.L_s1_not_aligned_loop:
+	ldb	r0, (r3, 0)
+	ldb	r4, (r1, 0)
+	addi	r3, 1
+	subu	r0, r4
+	PRE_BNEZAD (r13)
+	addi	r1, 1
+	bnez	r0, .L_return
+	BNEZAD (r13, .L_s1_not_aligned_loop)
+	br	.L_s1_aligned
+ENDPROC(memcmp)