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, 145 insertions, 0 deletions

diff --git a/fs/smb/client/reparse.h b/fs/smb/client/reparse.h
new file mode 100644
index 000000000..49efd85b1
--- /dev/null
+++ b/fs/smb/client/reparse.h
@@ -0,0 +1,145 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2024 Paulo Alcantara <pc@manguebit.com>
+ */
+
+#ifndef _CIFS_REPARSE_H
+#define _CIFS_REPARSE_H
+
+#include <linux/fs.h>
+#include <linux/stat.h>
+#include <linux/uidgid.h>
+#include "fs_context.h"
+#include "cifsglob.h"
+#include "../common/smbfsctl.h"
+
+#define REPARSE_SYM_PATH_MAX 4060
+
+/*
+ * Used only by cifs.ko to ignore reparse points from files when client or
+ * server doesn't support FSCTL_GET_REPARSE_POINT.
+ */
+#define IO_REPARSE_TAG_INTERNAL ((__u32)~0U)
+
+static inline dev_t reparse_mkdev(void *ptr)
+{
+	u64 v = le64_to_cpu(*(__le64 *)ptr);
+
+	return MKDEV(v & 0xffffffff, v >> 32);
+}
+
+static inline kuid_t wsl_make_kuid(struct cifs_sb_info *cifs_sb,
+				   void *ptr)
+{
+	u32 uid = le32_to_cpu(*(__le32 *)ptr);
+
+	if (cifs_sb_flags(cifs_sb) & CIFS_MOUNT_OVERR_UID)
+		return cifs_sb->ctx->linux_uid;
+	return make_kuid(current_user_ns(), uid);
+}
+
+static inline kgid_t wsl_make_kgid(struct cifs_sb_info *cifs_sb,
+				   void *ptr)
+{
+	u32 gid = le32_to_cpu(*(__le32 *)ptr);
+
+	if (cifs_sb_flags(cifs_sb) & CIFS_MOUNT_OVERR_GID)
+		return cifs_sb->ctx->linux_gid;
+	return make_kgid(current_user_ns(), gid);
+}
+
+static inline u64 reparse_mode_nfs_type(mode_t mode)
+{
+	switch (mode & S_IFMT) {
+	case S_IFLNK: return NFS_SPECFILE_LNK;
+	case S_IFBLK: return NFS_SPECFILE_BLK;
+	case S_IFCHR: return NFS_SPECFILE_CHR;
+	case S_IFIFO: return NFS_SPECFILE_FIFO;
+	case S_IFSOCK: return NFS_SPECFILE_SOCK;
+	}
+	return 0;
+}
+
+static inline u32 reparse_mode_wsl_tag(mode_t mode)
+{
+	switch (mode & S_IFMT) {
+	case S_IFLNK: return IO_REPARSE_TAG_LX_SYMLINK;
+	case S_IFBLK: return IO_REPARSE_TAG_LX_BLK;
+	case S_IFCHR: return IO_REPARSE_TAG_LX_CHR;
+	case S_IFIFO: return IO_REPARSE_TAG_LX_FIFO;
+	case S_IFSOCK: return IO_REPARSE_TAG_AF_UNIX;
+	}
+	return 0;
+}
+
+/*
+ * Match a reparse point inode if reparse tag and ctime haven't changed.
+ *
+ * Windows Server updates ctime of reparse points when their data have changed.
+ * The server doesn't allow changing reparse tags from existing reparse points,
+ * though it's worth checking.
+ */
+static inline bool reparse_inode_match(struct inode *inode,
+				       struct cifs_fattr *fattr)
+{
+	struct cifsInodeInfo *cinode = CIFS_I(inode);
+	struct timespec64 ctime = inode_get_ctime(inode);
+
+	/*
+	 * Do not match reparse tags when client or server doesn't support
+	 * FSCTL_GET_REPARSE_POINT.  @fattr->cf_cifstag should contain correct
+	 * reparse tag from query dir response but the client won't be able to
+	 * read the reparse point data anyway.  This spares us a revalidation.
+	 */
+	if (cinode->reparse_tag != IO_REPARSE_TAG_INTERNAL &&
+	    cinode->reparse_tag != fattr->cf_cifstag)
+		return false;
+	return (cinode->cifsAttrs & ATTR_REPARSE_POINT) &&
+		timespec64_equal(&ctime, &fattr->cf_ctime);
+}
+
+static inline u32 cifs_open_data_attrs(const struct cifs_open_info_data *data)
+{
+	if (data->contains_posix_file_info)
+		return le32_to_cpu(data->posix_fi.DosAttributes);
+
+	return le32_to_cpu(data->fi.Attributes);
+}
+
+static inline bool cifs_open_data_reparse(struct cifs_open_info_data *data)
+{
+	u32 attrs = cifs_open_data_attrs(data);
+
+	if (data->contains_posix_file_info) {
+		struct smb311_posix_qinfo *fi = &data->posix_fi;
+
+		if (data->reparse_point) {
+			attrs |= ATTR_REPARSE_POINT;
+			fi->DosAttributes = cpu_to_le32(attrs);
+		}
+
+	} else {
+		struct smb2_file_all_info *fi = &data->fi;
+
+		if (data->reparse_point) {
+			attrs |= ATTR_REPARSE_POINT;
+			fi->Attributes = cpu_to_le32(attrs);
+		}
+	}
+
+	return attrs & ATTR_REPARSE_POINT;
+}
+
+bool cifs_reparse_point_to_fattr(struct cifs_sb_info *cifs_sb,
+				 struct cifs_fattr *fattr,
+				 struct cifs_open_info_data *data);
+int create_reparse_symlink(const unsigned int xid, struct inode *inode,
+			   struct dentry *dentry, struct cifs_tcon *tcon,
+			   const char *full_path, const char *symname);
+int mknod_reparse(unsigned int xid, struct inode *inode, struct dentry *dentry,
+		  struct cifs_tcon *tcon, const char *full_path, umode_t mode,
+		  dev_t dev);
+struct reparse_data_buffer *smb2_get_reparse_point_buffer(const struct kvec *rsp_iov,
+							  u32 *plen);
+
+#endif /* _CIFS_REPARSE_H */