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, 298 insertions, 0 deletions

diff --git a/fs/smb/client/namespace.c b/fs/smb/client/namespace.c
new file mode 100644
index 000000000..52a51b032
--- /dev/null
+++ b/fs/smb/client/namespace.c
@@ -0,0 +1,298 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *   Contains mounting routines used for handling traversal via SMB junctions.
+ *
+ *   Copyright (c) 2007 Igor Mammedov
+ *   Copyright (C) International Business Machines  Corp., 2008
+ *   Author(s): Igor Mammedov (niallain@gmail.com)
+ *		Steve French (sfrench@us.ibm.com)
+ *   Copyright (c) 2023 Paulo Alcantara <palcantara@suse.de>
+ */
+
+#include <linux/dcache.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+#include <linux/slab.h>
+#include <linux/vfs.h>
+#include <linux/fs.h>
+#include <linux/inet.h>
+#include "cifsglob.h"
+#include "cifsproto.h"
+#include "cifsfs.h"
+#include "cifs_debug.h"
+#include "fs_context.h"
+
+static LIST_HEAD(cifs_automount_list);
+
+static void cifs_expire_automounts(struct work_struct *work);
+static DECLARE_DELAYED_WORK(cifs_automount_task,
+			    cifs_expire_automounts);
+static int cifs_mountpoint_expiry_timeout = 500 * HZ;
+
+static void cifs_expire_automounts(struct work_struct *work)
+{
+	struct list_head *list = &cifs_automount_list;
+
+	mark_mounts_for_expiry(list);
+	if (!list_empty(list))
+		schedule_delayed_work(&cifs_automount_task,
+				      cifs_mountpoint_expiry_timeout);
+}
+
+void cifs_release_automount_timer(void)
+{
+	if (WARN_ON(!list_empty(&cifs_automount_list)))
+		return;
+	cancel_delayed_work_sync(&cifs_automount_task);
+}
+
+/**
+ * cifs_build_devname - build a devicename from a UNC and optional prepath
+ * @nodename:	pointer to UNC string
+ * @prepath:	pointer to prefixpath (or NULL if there isn't one)
+ *
+ * Build a new cifs devicename after chasing a DFS referral. Allocate a buffer
+ * big enough to hold the final thing. Copy the UNC from the nodename, and
+ * concatenate the prepath onto the end of it if there is one.
+ *
+ * Returns pointer to the built string, or a ERR_PTR. Caller is responsible
+ * for freeing the returned string.
+ */
+char *
+cifs_build_devname(char *nodename, const char *prepath)
+{
+	size_t pplen;
+	size_t unclen;
+	char *dev;
+	char *pos;
+
+	/* skip over any preceding delimiters */
+	nodename += strspn(nodename, "\\");
+	if (!*nodename)
+		return ERR_PTR(-EINVAL);
+
+	/* get length of UNC and set pos to last char */
+	unclen = strlen(nodename);
+	pos = nodename + unclen - 1;
+
+	/* trim off any trailing delimiters */
+	while (*pos == '\\') {
+		--pos;
+		--unclen;
+	}
+
+	/* allocate a buffer:
+	 * +2 for preceding "//"
+	 * +1 for delimiter between UNC and prepath
+	 * +1 for trailing NULL
+	 */
+	pplen = prepath ? strlen(prepath) : 0;
+	dev = kmalloc(2 + unclen + 1 + pplen + 1, GFP_KERNEL);
+	if (!dev)
+		return ERR_PTR(-ENOMEM);
+
+	pos = dev;
+	/* add the initial "//" */
+	*pos = '/';
+	++pos;
+	*pos = '/';
+	++pos;
+
+	/* copy in the UNC portion from referral */
+	memcpy(pos, nodename, unclen);
+	pos += unclen;
+
+	/* copy the prefixpath remainder (if there is one) */
+	if (pplen) {
+		*pos = '/';
+		++pos;
+		memcpy(pos, prepath, pplen);
+		pos += pplen;
+	}
+
+	/* NULL terminator */
+	*pos = '\0';
+
+	convert_delimiter(dev, '/');
+	return dev;
+}
+
+static bool is_dfs_mount(struct dentry *dentry)
+{
+	struct cifs_sb_info *cifs_sb = CIFS_SB(dentry->d_sb);
+	struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb);
+	bool ret;
+
+	spin_lock(&tcon->tc_lock);
+	ret = !!tcon->origin_fullpath;
+	spin_unlock(&tcon->tc_lock);
+	return ret;
+}
+
+/* Return full path out of a dentry set for automount */
+static char *automount_fullpath(struct dentry *dentry, void *page)
+{
+	struct cifs_sb_info *cifs_sb = CIFS_SB(dentry->d_sb);
+	struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb);
+	size_t len;
+	char *s;
+
+	spin_lock(&tcon->tc_lock);
+	if (!tcon->origin_fullpath) {
+		spin_unlock(&tcon->tc_lock);
+		return build_path_from_dentry_optional_prefix(dentry,
+							      page,
+							      true);
+	}
+	spin_unlock(&tcon->tc_lock);
+
+	if (unlikely(!page))
+		return ERR_PTR(-ENOMEM);
+
+	s = dentry_path_raw(dentry, page, PATH_MAX);
+	if (IS_ERR(s))
+		return s;
+	/* for root, we want "" */
+	if (!s[1])
+		s++;
+
+	spin_lock(&tcon->tc_lock);
+	len = strlen(tcon->origin_fullpath);
+	if (s < (char *)page + len) {
+		spin_unlock(&tcon->tc_lock);
+		return ERR_PTR(-ENAMETOOLONG);
+	}
+
+	s -= len;
+	memcpy(s, tcon->origin_fullpath, len);
+	spin_unlock(&tcon->tc_lock);
+	convert_delimiter(s, '/');
+
+	return s;
+}
+
+static void fs_context_set_ids(struct smb3_fs_context *ctx)
+{
+	kuid_t uid = current_fsuid();
+	kgid_t gid = current_fsgid();
+
+	if (ctx->multiuser) {
+		if (!ctx->uid_specified)
+			ctx->linux_uid = uid;
+		if (!ctx->gid_specified)
+			ctx->linux_gid = gid;
+	}
+	if (!ctx->cruid_specified)
+		ctx->cred_uid = uid;
+}
+
+/*
+ * Create a vfsmount that we can automount
+ */
+static struct vfsmount *cifs_do_automount(struct path *path)
+{
+	int rc;
+	struct dentry *mntpt = path->dentry;
+	struct fs_context *fc;
+	void *page = NULL;
+	struct smb3_fs_context *ctx, *cur_ctx;
+	struct smb3_fs_context tmp;
+	char *full_path;
+	struct vfsmount *mnt;
+	struct cifs_sb_info *mntpt_sb;
+	struct cifs_ses *ses;
+
+	if (IS_ROOT(mntpt))
+		return ERR_PTR(-ESTALE);
+
+	mntpt_sb = CIFS_SB(mntpt->d_sb);
+	ses = cifs_sb_master_tcon(mntpt_sb)->ses;
+	cur_ctx = mntpt_sb->ctx;
+
+	/*
+	 * At this point, the root session should be in the mntpt sb. We should
+	 * bring the sb context passwords in sync with the root session's
+	 * passwords. This would help prevent unnecessary retries and password
+	 * swaps for automounts.
+	 */
+	mutex_lock(&ses->session_mutex);
+	rc = smb3_sync_session_ctx_passwords(mntpt_sb, ses);
+	mutex_unlock(&ses->session_mutex);
+
+	if (rc)
+		return ERR_PTR(rc);
+
+	fc = fs_context_for_submount(path->mnt->mnt_sb->s_type, mntpt);
+	if (IS_ERR(fc))
+		return ERR_CAST(fc);
+
+	ctx = smb3_fc2context(fc);
+
+	page = alloc_dentry_path();
+	full_path = automount_fullpath(mntpt, page);
+	if (IS_ERR(full_path)) {
+		mnt = ERR_CAST(full_path);
+		goto out;
+	}
+
+	tmp = *cur_ctx;
+	tmp.source = NULL;
+	tmp.leaf_fullpath = NULL;
+	tmp.UNC = tmp.prepath = NULL;
+	tmp.dfs_root_ses = NULL;
+	fs_context_set_ids(&tmp);
+
+	rc = smb3_fs_context_dup(ctx, &tmp);
+	if (rc) {
+		mnt = ERR_PTR(rc);
+		goto out;
+	}
+
+	rc = smb3_parse_devname(full_path, ctx);
+	if (rc) {
+		mnt = ERR_PTR(rc);
+		goto out;
+	}
+
+	ctx->source = smb3_fs_context_fullpath(ctx, '/');
+	if (IS_ERR(ctx->source)) {
+		mnt = ERR_CAST(ctx->source);
+		ctx->source = NULL;
+		goto out;
+	}
+	ctx->dfs_automount = ctx->dfs_conn = is_dfs_mount(mntpt);
+	cifs_dbg(FYI, "%s: ctx: source=%s UNC=%s prepath=%s dfs_automount=%d\n",
+		 __func__, ctx->source, ctx->UNC, ctx->prepath, ctx->dfs_automount);
+
+	mnt = fc_mount(fc);
+out:
+	put_fs_context(fc);
+	free_dentry_path(page);
+	return mnt;
+}
+
+/*
+ * Attempt to automount the referral
+ */
+struct vfsmount *cifs_d_automount(struct path *path)
+{
+	struct vfsmount *newmnt;
+
+	cifs_dbg(FYI, "%s: %pd\n", __func__, path->dentry);
+
+	newmnt = cifs_do_automount(path);
+	if (IS_ERR(newmnt)) {
+		cifs_dbg(FYI, "leaving %s [automount failed]\n" , __func__);
+		return newmnt;
+	}
+
+	mnt_set_expiry(newmnt, &cifs_automount_list);
+	schedule_delayed_work(&cifs_automount_task,
+			      cifs_mountpoint_expiry_timeout);
+	cifs_dbg(FYI, "leaving %s [ok]\n" , __func__);
+	return newmnt;
+}
+
+const struct inode_operations cifs_namespace_inode_operations = {
+	.fileattr_get	= cifs_fileattr_get,
+};