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, 363 insertions, 0 deletions

diff --git a/fs/smb/server/misc.c b/fs/smb/server/misc.c
new file mode 100644
index 000000000..2dd91c9e9
--- /dev/null
+++ b/fs/smb/server/misc.c
@@ -0,0 +1,363 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *   Copyright (C) 2016 Namjae Jeon <linkinjeon@kernel.org>
+ *   Copyright (C) 2018 Samsung Electronics Co., Ltd.
+ */
+
+#include <linux/kernel.h>
+#include <linux/xattr.h>
+#include <linux/fs.h>
+#include <linux/unicode.h>
+
+#include "misc.h"
+#include "smb_common.h"
+#include "connection.h"
+#include "vfs.h"
+
+#include "mgmt/share_config.h"
+
+/**
+ * match_pattern() - compare a string with a pattern which might include
+ * wildcard '*' and '?'
+ * TODO : implement consideration about DOS_DOT, DOS_QM and DOS_STAR
+ *
+ * @str:	string to compare with a pattern
+ * @len:	string length
+ * @pattern:	pattern string which might include wildcard '*' and '?'
+ *
+ * Return:	0 if pattern matched with the string, otherwise non zero value
+ */
+int match_pattern(const char *str, size_t len, const char *pattern)
+{
+	const char *s = str;
+	const char *p = pattern;
+	bool star = false;
+
+	while (*s && len) {
+		switch (*p) {
+		case '?':
+			s++;
+			len--;
+			p++;
+			break;
+		case '*':
+			star = true;
+			str = s;
+			if (!*++p)
+				return true;
+			pattern = p;
+			break;
+		default:
+			if (tolower(*s) == tolower(*p)) {
+				s++;
+				len--;
+				p++;
+			} else {
+				if (!star)
+					return false;
+				str++;
+				s = str;
+				p = pattern;
+			}
+			break;
+		}
+	}
+
+	if (*p == '*')
+		++p;
+	return !*p;
+}
+
+/*
+ * is_char_allowed() - check for valid character
+ * @ch:		input character to be checked
+ *
+ * Return:	1 if char is allowed, otherwise 0
+ */
+static inline int is_char_allowed(char ch)
+{
+	/* check for control chars, wildcards etc. */
+	if (!(ch & 0x80) &&
+	    (ch <= 0x1f ||
+	     ch == '?' || ch == '"' || ch == '<' ||
+	     ch == '>' || ch == '|' || ch == '*'))
+		return 0;
+
+	return 1;
+}
+
+int ksmbd_validate_filename(char *filename)
+{
+	while (*filename) {
+		char c = *filename;
+
+		filename++;
+		if (!is_char_allowed(c)) {
+			ksmbd_debug(VFS, "File name validation failed: 0x%x\n", c);
+			return -ENOENT;
+		}
+	}
+
+	return 0;
+}
+
+static int ksmbd_validate_stream_name(char *stream_name)
+{
+	while (*stream_name) {
+		char c = *stream_name;
+
+		stream_name++;
+		if (c == '/' || c == ':' || c == '\\') {
+			pr_err("Stream name validation failed: %c\n", c);
+			return -ENOENT;
+		}
+	}
+
+	return 0;
+}
+
+int parse_stream_name(char *filename, char **stream_name, int *s_type)
+{
+	char *stream_type;
+	char *s_name;
+	int rc = 0;
+	bool has_stream_type = false;
+
+	*stream_name = NULL;
+	s_name = filename;
+	filename = strsep(&s_name, ":");
+	ksmbd_debug(SMB, "filename : %s, streams : %s\n", filename, s_name);
+	if (strchr(s_name, ':')) {
+		stream_type = s_name;
+		s_name = strsep(&stream_type, ":");
+
+		rc = ksmbd_validate_stream_name(s_name);
+		if (rc < 0) {
+			rc = -ENOENT;
+			goto out;
+		}
+
+		ksmbd_debug(SMB, "stream name : %s, stream type : %s\n", s_name,
+			    stream_type);
+		if (!strncasecmp("$data", stream_type, 5)) {
+			*s_type = DATA_STREAM;
+			has_stream_type = true;
+		} else if (!strncasecmp("$index_allocation", stream_type, 17)) {
+			*s_type = DIR_STREAM;
+			has_stream_type = true;
+		} else {
+			rc = -ENOENT;
+		}
+	}
+
+	if (has_stream_type && !s_name[0] && *s_type == DATA_STREAM)
+		goto out;
+
+	*stream_name = s_name;
+out:
+	return rc;
+}
+
+/**
+ * convert_to_nt_pathname() - extract and return windows path string
+ *      whose share directory prefix was removed from file path
+ * @share: ksmbd_share_config pointer
+ * @path: path to report
+ *
+ * Return : windows path string or error
+ */
+
+char *convert_to_nt_pathname(struct ksmbd_share_config *share,
+			     const struct path *path)
+{
+	char *pathname, *ab_pathname, *nt_pathname;
+	int share_path_len = share->path_sz;
+	size_t ab_pathname_len;
+	int prefix;
+
+	pathname = kmalloc(PATH_MAX, KSMBD_DEFAULT_GFP);
+	if (!pathname)
+		return ERR_PTR(-EACCES);
+
+	ab_pathname = d_path(path, pathname, PATH_MAX);
+	if (IS_ERR(ab_pathname)) {
+		nt_pathname = ERR_PTR(-EACCES);
+		goto free_pathname;
+	}
+
+	if (strncmp(ab_pathname, share->path, share_path_len)) {
+		nt_pathname = ERR_PTR(-EACCES);
+		goto free_pathname;
+	}
+
+	ab_pathname_len = strlen(&ab_pathname[share_path_len]);
+	prefix = ab_pathname[share_path_len] == '\0' ? 1 : 0;
+	nt_pathname = kmalloc(prefix + ab_pathname_len + 1, KSMBD_DEFAULT_GFP);
+	if (!nt_pathname) {
+		nt_pathname = ERR_PTR(-ENOMEM);
+		goto free_pathname;
+	}
+
+	if (prefix)
+		*nt_pathname = '/';
+	memcpy(nt_pathname + prefix, &ab_pathname[share_path_len],
+	       ab_pathname_len + 1);
+
+	ksmbd_conv_path_to_windows(nt_pathname);
+
+free_pathname:
+	kfree(pathname);
+	return nt_pathname;
+}
+
+int get_nlink(struct kstat *st)
+{
+	int nlink;
+
+	nlink = st->nlink;
+	if (S_ISDIR(st->mode))
+		nlink--;
+
+	return nlink;
+}
+
+void ksmbd_conv_path_to_unix(char *path)
+{
+	strreplace(path, '\\', '/');
+}
+
+void ksmbd_strip_last_slash(char *path)
+{
+	int len = strlen(path);
+
+	while (len && path[len - 1] == '/') {
+		path[len - 1] = '\0';
+		len--;
+	}
+}
+
+void ksmbd_conv_path_to_windows(char *path)
+{
+	strreplace(path, '/', '\\');
+}
+
+char *ksmbd_casefold_sharename(struct unicode_map *um, const char *name)
+{
+	char *cf_name;
+	int cf_len;
+
+	cf_name = kzalloc(KSMBD_REQ_MAX_SHARE_NAME, KSMBD_DEFAULT_GFP);
+	if (!cf_name)
+		return ERR_PTR(-ENOMEM);
+
+	if (IS_ENABLED(CONFIG_UNICODE) && um) {
+		const struct qstr q_name = {.name = name, .len = strlen(name)};
+
+		cf_len = utf8_casefold(um, &q_name, cf_name,
+				       KSMBD_REQ_MAX_SHARE_NAME);
+		if (cf_len < 0)
+			goto out_ascii;
+
+		return cf_name;
+	}
+
+out_ascii:
+	cf_len = strscpy(cf_name, name, KSMBD_REQ_MAX_SHARE_NAME);
+	if (cf_len < 0) {
+		kfree(cf_name);
+		return ERR_PTR(-E2BIG);
+	}
+
+	for (; *cf_name; ++cf_name)
+		*cf_name = isascii(*cf_name) ? tolower(*cf_name) : *cf_name;
+	return cf_name - cf_len;
+}
+
+/**
+ * ksmbd_extract_sharename() - get share name from tree connect request
+ * @um: pointer to a unicode_map structure for character encoding handling
+ * @treename:	buffer containing tree name and share name
+ *
+ * Return:      share name on success, otherwise error
+ */
+char *ksmbd_extract_sharename(struct unicode_map *um, const char *treename)
+{
+	const char *name = treename, *pos = strrchr(name, '\\');
+
+	if (pos)
+		name = (pos + 1);
+
+	/* caller has to free the memory */
+	return ksmbd_casefold_sharename(um, name);
+}
+
+char *ksmbd_convert_dir_info_name(struct ksmbd_dir_info *d_info,
+				  const struct nls_table *local_nls,
+				  int *conv_len)
+{
+	char *conv;
+	int  sz = min(4 * d_info->name_len, PATH_MAX);
+
+	if (!sz)
+		return NULL;
+
+	conv = kmalloc(sz, KSMBD_DEFAULT_GFP);
+	if (!conv)
+		return NULL;
+
+	/* XXX */
+	*conv_len = smbConvertToUTF16((__le16 *)conv, d_info->name,
+				      d_info->name_len, local_nls, 0);
+	*conv_len *= 2;
+
+	/* We allocate buffer twice bigger than needed. */
+	conv[*conv_len] = 0x00;
+	conv[*conv_len + 1] = 0x00;
+	return conv;
+}
+
+/*
+ * Convert the NT UTC (based 1601-01-01, in hundred nanosecond units)
+ * into Unix UTC (based 1970-01-01, in seconds).
+ */
+struct timespec64 ksmbd_NTtimeToUnix(__le64 ntutc)
+{
+	struct timespec64 ts;
+
+	/* Subtract the NTFS time offset, then convert to 1s intervals. */
+	s64 t = le64_to_cpu(ntutc) - NTFS_TIME_OFFSET;
+	u64 abs_t;
+
+	/*
+	 * Unfortunately can not use normal 64 bit division on 32 bit arch, but
+	 * the alternative, do_div, does not work with negative numbers so have
+	 * to special case them
+	 */
+	if (t < 0) {
+		abs_t = -t;
+		ts.tv_nsec = do_div(abs_t, 10000000) * 100;
+		ts.tv_nsec = -ts.tv_nsec;
+		ts.tv_sec = -abs_t;
+	} else {
+		abs_t = t;
+		ts.tv_nsec = do_div(abs_t, 10000000) * 100;
+		ts.tv_sec = abs_t;
+	}
+
+	return ts;
+}
+
+/* Convert the Unix UTC into NT UTC. */
+inline u64 ksmbd_UnixTimeToNT(struct timespec64 t)
+{
+	/* Convert to 100ns intervals and then add the NTFS time offset. */
+	return (u64)t.tv_sec * 10000000 + t.tv_nsec / 100 + NTFS_TIME_OFFSET;
+}
+
+inline long long ksmbd_systime(void)
+{
+	struct timespec64	ts;
+
+	ktime_get_real_ts64(&ts);
+	return ksmbd_UnixTimeToNT(ts);
+}