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, 179 insertions, 0 deletions

diff --git a/fs/coda/cnode.c b/fs/coda/cnode.c
new file mode 100644
index 000000000..70bb0579b
--- /dev/null
+++ b/fs/coda/cnode.c
@@ -0,0 +1,179 @@
+// SPDX-License-Identifier: GPL-2.0
+/* cnode related routines for the coda kernel code
+   (C) 1996 Peter Braam
+   */
+
+#include <linux/types.h>
+#include <linux/string.h>
+#include <linux/time.h>
+
+#include <linux/coda.h>
+#include <linux/pagemap.h>
+#include "coda_psdev.h"
+#include "coda_linux.h"
+
+static inline int coda_fideq(struct CodaFid *fid1, struct CodaFid *fid2)
+{
+	return memcmp(fid1, fid2, sizeof(*fid1)) == 0;
+}
+
+static const struct inode_operations coda_symlink_inode_operations = {
+	.get_link	= page_get_link,
+	.setattr	= coda_setattr,
+};
+
+/* cnode.c */
+static void coda_fill_inode(struct inode *inode, struct coda_vattr *attr)
+{
+        coda_vattr_to_iattr(inode, attr);
+
+        if (S_ISREG(inode->i_mode)) {
+                inode->i_op = &coda_file_inode_operations;
+                inode->i_fop = &coda_file_operations;
+        } else if (S_ISDIR(inode->i_mode)) {
+                inode->i_op = &coda_dir_inode_operations;
+                inode->i_fop = &coda_dir_operations;
+        } else if (S_ISLNK(inode->i_mode)) {
+		inode->i_op = &coda_symlink_inode_operations;
+		inode_nohighmem(inode);
+		inode->i_data.a_ops = &coda_symlink_aops;
+		inode->i_mapping = &inode->i_data;
+	} else
+                init_special_inode(inode, inode->i_mode, huge_decode_dev(attr->va_rdev));
+}
+
+static int coda_test_inode(struct inode *inode, void *data)
+{
+	struct CodaFid *fid = (struct CodaFid *)data;
+	struct coda_inode_info *cii = ITOC(inode);
+	return coda_fideq(&cii->c_fid, fid);
+}
+
+static int coda_set_inode(struct inode *inode, void *data)
+{
+	struct CodaFid *fid = (struct CodaFid *)data;
+	struct coda_inode_info *cii = ITOC(inode);
+	cii->c_fid = *fid;
+	return 0;
+}
+
+struct inode * coda_iget(struct super_block * sb, struct CodaFid * fid,
+			 struct coda_vattr * attr)
+{
+	struct inode *inode;
+	struct coda_inode_info *cii;
+	unsigned long hash = coda_f2i(fid);
+	umode_t inode_type = coda_inode_type(attr);
+
+retry:
+	inode = iget5_locked(sb, hash, coda_test_inode, coda_set_inode, fid);
+	if (!inode)
+		return ERR_PTR(-ENOMEM);
+
+	if (inode_state_read_once(inode) & I_NEW) {
+		cii = ITOC(inode);
+		/* we still need to set i_ino for things like stat(2) */
+		inode->i_ino = hash;
+		/* inode is locked and unique, no need to grab cii->c_lock */
+		cii->c_mapcount = 0;
+		coda_fill_inode(inode, attr);
+		unlock_new_inode(inode);
+	} else if ((inode->i_mode & S_IFMT) != inode_type) {
+		/* Inode has changed type, mark bad and grab a new one */
+		remove_inode_hash(inode);
+		coda_flag_inode(inode, C_PURGE);
+		iput(inode);
+		goto retry;
+	}
+	return inode;
+}
+
+/* this is effectively coda_iget:
+   - get attributes (might be cached)
+   - get the inode for the fid using vfs iget
+   - link the two up if this is needed
+   - fill in the attributes
+*/
+struct inode *coda_cnode_make(struct CodaFid *fid, struct super_block *sb)
+{
+        struct coda_vattr attr;
+	struct inode *inode;
+        int error;
+        
+	/* We get inode numbers from Venus -- see venus source */
+	error = venus_getattr(sb, fid, &attr);
+	if (error)
+		return ERR_PTR(error);
+
+	inode = coda_iget(sb, fid, &attr);
+	if (IS_ERR(inode))
+		pr_warn("%s: coda_iget failed\n", __func__);
+	return inode;
+}
+
+
+/* Although we treat Coda file identifiers as immutable, there is one
+ * special case for files created during a disconnection where they may
+ * not be globally unique. When an identifier collision is detected we
+ * first try to flush the cached inode from the kernel and finally
+ * resort to renaming/rehashing in-place. Userspace remembers both old
+ * and new values of the identifier to handle any in-flight upcalls.
+ * The real solution is to use globally unique UUIDs as identifiers, but
+ * retrofitting the existing userspace code for this is non-trivial. */
+void coda_replace_fid(struct inode *inode, struct CodaFid *oldfid, 
+		      struct CodaFid *newfid)
+{
+	struct coda_inode_info *cii = ITOC(inode);
+	unsigned long hash = coda_f2i(newfid);
+	
+	BUG_ON(!coda_fideq(&cii->c_fid, oldfid));
+
+	/* replace fid and rehash inode */
+	/* XXX we probably need to hold some lock here! */
+	remove_inode_hash(inode);
+	cii->c_fid = *newfid;
+	inode->i_ino = hash;
+	__insert_inode_hash(inode, hash);
+}
+
+/* convert a fid to an inode. */
+struct inode *coda_fid_to_inode(struct CodaFid *fid, struct super_block *sb) 
+{
+	struct inode *inode;
+	unsigned long hash = coda_f2i(fid);
+
+	inode = ilookup5(sb, hash, coda_test_inode, fid);
+	if ( !inode )
+		return NULL;
+
+	/* we should never see newly created inodes because we intentionally
+	 * fail in the initialization callback */
+	BUG_ON(inode_state_read_once(inode) & I_NEW);
+
+	return inode;
+}
+
+struct coda_file_info *coda_ftoc(struct file *file)
+{
+	struct coda_file_info *cfi = file->private_data;
+
+	BUG_ON(!cfi || cfi->cfi_magic != CODA_MAGIC);
+
+	return cfi;
+
+}
+
+/* the CONTROL inode is made without asking attributes from Venus */
+struct inode *coda_cnode_makectl(struct super_block *sb)
+{
+	struct inode *inode = new_inode(sb);
+	if (inode) {
+		inode->i_ino = CTL_INO;
+		inode->i_op = &coda_ioctl_inode_operations;
+		inode->i_fop = &coda_ioctl_operations;
+		inode->i_mode = 0444;
+		return inode;
+	}
+	return ERR_PTR(-ENOMEM);
+}
+