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, 114 insertions, 0 deletions

diff --git a/fs/hpfs/name.c b/fs/hpfs/name.c
new file mode 100644
index 000000000..ef7ba77f3
--- /dev/null
+++ b/fs/hpfs/name.c
@@ -0,0 +1,114 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  linux/fs/hpfs/name.c
+ *
+ *  Mikulas Patocka (mikulas@artax.karlin.mff.cuni.cz), 1998-1999
+ *
+ *  operations with filenames
+ */
+
+#include "hpfs_fn.h"
+
+static inline int not_allowed_char(unsigned char c)
+{
+	return c<' ' || c=='"' || c=='*' || c=='/' || c==':' || c=='<' ||
+	      c=='>' || c=='?' || c=='\\' || c=='|';
+}
+
+static inline int no_dos_char(unsigned char c)
+{	/* Characters that are allowed in HPFS but not in DOS */
+	return c=='+' || c==',' || c==';' || c=='=' || c=='[' || c==']';
+}
+
+static inline unsigned char upcase(unsigned char *dir, unsigned char a)
+{
+	if (a<128 || a==255) return a>='a' && a<='z' ? a - 0x20 : a;
+	if (!dir) return a;
+	return dir[a-128];
+}
+
+unsigned char hpfs_upcase(unsigned char *dir, unsigned char a)
+{
+	return upcase(dir, a);
+}
+
+static inline unsigned char locase(unsigned char *dir, unsigned char a)
+{
+	if (a<128 || a==255) return a>='A' && a<='Z' ? a + 0x20 : a;
+	if (!dir) return a;
+	return dir[a];
+}
+
+int hpfs_chk_name(const unsigned char *name, unsigned *len)
+{
+	int i;
+	if (*len > 254) return -ENAMETOOLONG;
+	hpfs_adjust_length(name, len);
+	if (!*len) return -EINVAL;
+	for (i = 0; i < *len; i++) if (not_allowed_char(name[i])) return -EINVAL;
+	if (*len == 1) if (name[0] == '.') return -EINVAL;
+	if (*len == 2) if (name[0] == '.' && name[1] == '.') return -EINVAL;
+	return 0;
+}
+
+unsigned char *hpfs_translate_name(struct super_block *s, unsigned char *from,
+			  unsigned len, int lc, int lng)
+{
+	unsigned char *to;
+	int i;
+	if (hpfs_sb(s)->sb_chk >= 2) if (hpfs_is_name_long(from, len) != lng) {
+		pr_err("Long name flag mismatch - name ");
+		for (i = 0; i < len; i++)
+			pr_cont("%c", from[i]);
+		pr_cont(" misidentified as %s.\n", lng ? "short" : "long");
+		pr_err("It's nothing serious. It could happen because of bug in OS/2.\nSet checks=normal to disable this message.\n");
+	}
+	if (!lc) return from;
+	if (!(to = kmalloc(len, GFP_KERNEL))) {
+		pr_err("can't allocate memory for name conversion buffer\n");
+		return from;
+	}
+	for (i = 0; i < len; i++) to[i] = locase(hpfs_sb(s)->sb_cp_table,from[i]);
+	return to;
+}
+
+int hpfs_compare_names(struct super_block *s,
+		       const unsigned char *n1, unsigned l1,
+		       const unsigned char *n2, unsigned l2, int last)
+{
+	unsigned l = l1 < l2 ? l1 : l2;
+	unsigned i;
+	if (last) return -1;
+	for (i = 0; i < l; i++) {
+		unsigned char c1 = upcase(hpfs_sb(s)->sb_cp_table,n1[i]);
+		unsigned char c2 = upcase(hpfs_sb(s)->sb_cp_table,n2[i]);
+		if (c1 < c2) return -1;
+		if (c1 > c2) return 1;
+	}
+	if (l1 < l2) return -1;
+	if (l1 > l2) return 1;
+	return 0;
+}
+
+int hpfs_is_name_long(const unsigned char *name, unsigned len)
+{
+	int i,j;
+	for (i = 0; i < len && name[i] != '.'; i++)
+		if (no_dos_char(name[i])) return 1;
+	if (!i || i > 8) return 1;
+	if (i == len) return 0;
+	for (j = i + 1; j < len; j++)
+		if (name[j] == '.' || no_dos_char(name[i])) return 1;
+	return j - i > 4;
+}
+
+/* OS/2 clears dots and spaces at the end of file name, so we have to */
+
+void hpfs_adjust_length(const unsigned char *name, unsigned *len)
+{
+	if (!*len) return;
+	if (*len == 1 && name[0] == '.') return;
+	if (*len == 2 && name[0] == '.' && name[1] == '.') return;
+	while (*len && (name[*len - 1] == '.' || name[*len - 1] == ' '))
+		(*len)--;
+}