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, 422 insertions, 0 deletions

diff --git a/fs/jfs/inode.c b/fs/jfs/inode.c
new file mode 100644
index 000000000..c7914dbc9
--- /dev/null
+++ b/fs/jfs/inode.c
@@ -0,0 +1,422 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *   Copyright (C) International Business Machines Corp., 2000-2004
+ *   Portions Copyright (C) Christoph Hellwig, 2001-2002
+ */
+
+#include <linux/fs.h>
+#include <linux/mpage.h>
+#include <linux/buffer_head.h>
+#include <linux/pagemap.h>
+#include <linux/quotaops.h>
+#include <linux/uio.h>
+#include <linux/writeback.h>
+#include "jfs_incore.h"
+#include "jfs_inode.h"
+#include "jfs_filsys.h"
+#include "jfs_imap.h"
+#include "jfs_extent.h"
+#include "jfs_unicode.h"
+#include "jfs_debug.h"
+#include "jfs_dmap.h"
+
+
+struct inode *jfs_iget(struct super_block *sb, unsigned long ino)
+{
+	struct inode *inode;
+	int ret;
+
+	inode = iget_locked(sb, ino);
+	if (!inode)
+		return ERR_PTR(-ENOMEM);
+	if (!(inode_state_read_once(inode) & I_NEW))
+		return inode;
+
+	ret = diRead(inode);
+	if (ret < 0) {
+		iget_failed(inode);
+		return ERR_PTR(ret);
+	}
+
+	if (S_ISREG(inode->i_mode)) {
+		inode->i_op = &jfs_file_inode_operations;
+		inode->i_fop = &jfs_file_operations;
+		inode->i_mapping->a_ops = &jfs_aops;
+	} else if (S_ISDIR(inode->i_mode)) {
+		inode->i_op = &jfs_dir_inode_operations;
+		inode->i_fop = &jfs_dir_operations;
+	} else if (S_ISLNK(inode->i_mode)) {
+		if (inode->i_size >= IDATASIZE) {
+			inode->i_op = &page_symlink_inode_operations;
+			inode_nohighmem(inode);
+			inode->i_mapping->a_ops = &jfs_aops;
+		} else {
+			inode->i_op = &jfs_fast_symlink_inode_operations;
+			inode->i_link = JFS_IP(inode)->i_inline;
+			/*
+			 * The inline data should be null-terminated, but
+			 * don't let on-disk corruption crash the kernel
+			 */
+			inode->i_link[inode->i_size] = '\0';
+		}
+	} else if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode) ||
+		   S_ISFIFO(inode->i_mode) || S_ISSOCK(inode->i_mode)) {
+		inode->i_op = &jfs_file_inode_operations;
+		init_special_inode(inode, inode->i_mode, inode->i_rdev);
+	} else {
+		printk(KERN_DEBUG "JFS: Invalid file type 0%04o for inode %llu.\n",
+		       inode->i_mode, inode->i_ino);
+		iget_failed(inode);
+		return ERR_PTR(-EIO);
+	}
+	unlock_new_inode(inode);
+	return inode;
+}
+
+/*
+ * Workhorse of both fsync & write_inode
+ */
+int jfs_commit_inode(struct inode *inode, int wait)
+{
+	int rc = 0;
+	tid_t tid;
+	static int noisy = 5;
+
+	jfs_info("In jfs_commit_inode, inode = 0x%p", inode);
+
+	/*
+	 * Don't commit if inode has been committed since last being
+	 * marked dirty, or if it has been deleted.
+	 */
+	if (inode->i_nlink == 0 || !test_cflag(COMMIT_Dirty, inode))
+		return 0;
+
+	if (isReadOnly(inode)) {
+		/* kernel allows writes to devices on read-only
+		 * partitions and may think inode is dirty
+		 */
+		if (!special_file(inode->i_mode) && noisy) {
+			jfs_err("jfs_commit_inode(0x%p) called on read-only volume",
+				inode);
+			jfs_err("Is remount racy?");
+			noisy--;
+		}
+		return 0;
+	}
+
+	tid = txBegin(inode->i_sb, COMMIT_INODE);
+	mutex_lock(&JFS_IP(inode)->commit_mutex);
+
+	/*
+	 * Retest inode state after taking commit_mutex
+	 */
+	if (inode->i_nlink && test_cflag(COMMIT_Dirty, inode))
+		rc = txCommit(tid, 1, &inode, wait ? COMMIT_SYNC : 0);
+
+	txEnd(tid);
+	mutex_unlock(&JFS_IP(inode)->commit_mutex);
+	return rc;
+}
+
+int jfs_write_inode(struct inode *inode, struct writeback_control *wbc)
+{
+	int wait = wbc->sync_mode == WB_SYNC_ALL;
+
+	if (inode->i_nlink == 0)
+		return 0;
+	/*
+	 * If COMMIT_DIRTY is not set, the inode isn't really dirty.
+	 * It has been committed since the last change, but was still
+	 * on the dirty inode list.
+	 */
+	if (!test_cflag(COMMIT_Dirty, inode)) {
+		/* Make sure committed changes hit the disk */
+		jfs_flush_journal(JFS_SBI(inode->i_sb)->log, wait);
+		return 0;
+	}
+
+	if (jfs_commit_inode(inode, wait)) {
+		jfs_err("jfs_write_inode: jfs_commit_inode failed!");
+		return -EIO;
+	} else
+		return 0;
+}
+
+void jfs_evict_inode(struct inode *inode)
+{
+	struct jfs_inode_info *ji = JFS_IP(inode);
+
+	jfs_info("In jfs_evict_inode, inode = 0x%p", inode);
+
+	if (!inode->i_nlink && !is_bad_inode(inode)) {
+		dquot_initialize(inode);
+
+		truncate_inode_pages_final(&inode->i_data);
+		if (JFS_IP(inode)->fileset == FILESYSTEM_I) {
+			struct inode *ipimap = JFS_SBI(inode->i_sb)->ipimap;
+
+			if (test_cflag(COMMIT_Freewmap, inode))
+				jfs_free_zero_link(inode);
+
+			if (ipimap && JFS_IP(ipimap)->i_imap)
+				diFree(inode);
+
+			/*
+			 * Free the inode from the quota allocation.
+			 */
+			dquot_free_inode(inode);
+		}
+	} else {
+		truncate_inode_pages_final(&inode->i_data);
+	}
+	clear_inode(inode);
+	dquot_drop(inode);
+
+	BUG_ON(!list_empty(&ji->anon_inode_list));
+
+	spin_lock_irq(&ji->ag_lock);
+	if (ji->active_ag != -1) {
+		struct bmap *bmap = JFS_SBI(inode->i_sb)->bmap;
+		atomic_dec(&bmap->db_active[ji->active_ag]);
+		ji->active_ag = -1;
+	}
+	spin_unlock_irq(&ji->ag_lock);
+}
+
+void jfs_dirty_inode(struct inode *inode, int flags)
+{
+	static int noisy = 5;
+
+	if (isReadOnly(inode)) {
+		if (!special_file(inode->i_mode) && noisy) {
+			/* kernel allows writes to devices on read-only
+			 * partitions and may try to mark inode dirty
+			 */
+			jfs_err("jfs_dirty_inode called on read-only volume");
+			jfs_err("Is remount racy?");
+			noisy--;
+		}
+		return;
+	}
+
+	set_cflag(COMMIT_Dirty, inode);
+}
+
+int jfs_get_block(struct inode *ip, sector_t lblock,
+		  struct buffer_head *bh_result, int create)
+{
+	s64 lblock64 = lblock;
+	int rc = 0;
+	xad_t xad;
+	s64 xaddr;
+	int xflag;
+	s32 xlen = bh_result->b_size >> ip->i_blkbits;
+
+	/*
+	 * Take appropriate lock on inode
+	 */
+	if (create)
+		IWRITE_LOCK(ip, RDWRLOCK_NORMAL);
+	else
+		IREAD_LOCK(ip, RDWRLOCK_NORMAL);
+
+	if (((lblock64 << ip->i_sb->s_blocksize_bits) < ip->i_size) &&
+	    (!xtLookup(ip, lblock64, xlen, &xflag, &xaddr, &xlen, 0)) &&
+	    xaddr) {
+		if (xflag & XAD_NOTRECORDED) {
+			if (!create)
+				/*
+				 * Allocated but not recorded, read treats
+				 * this as a hole
+				 */
+				goto unlock;
+			XADoffset(&xad, lblock64);
+			XADlength(&xad, xlen);
+			XADaddress(&xad, xaddr);
+			rc = extRecord(ip, &xad);
+			if (rc)
+				goto unlock;
+			set_buffer_new(bh_result);
+		}
+
+		map_bh(bh_result, ip->i_sb, xaddr);
+		bh_result->b_size = xlen << ip->i_blkbits;
+		goto unlock;
+	}
+	if (!create)
+		goto unlock;
+
+	/*
+	 * Allocate a new block
+	 */
+	if ((rc = extHint(ip, lblock64 << ip->i_sb->s_blocksize_bits, &xad)))
+		goto unlock;
+	rc = extAlloc(ip, xlen, lblock64, &xad, false);
+	if (rc)
+		goto unlock;
+
+	set_buffer_new(bh_result);
+	map_bh(bh_result, ip->i_sb, addressXAD(&xad));
+	bh_result->b_size = lengthXAD(&xad) << ip->i_blkbits;
+
+      unlock:
+	/*
+	 * Release lock on inode
+	 */
+	if (create)
+		IWRITE_UNLOCK(ip);
+	else
+		IREAD_UNLOCK(ip);
+	return rc;
+}
+
+static int jfs_writepages(struct address_space *mapping,
+			struct writeback_control *wbc)
+{
+	return mpage_writepages(mapping, wbc, jfs_get_block);
+}
+
+static int jfs_read_folio(struct file *file, struct folio *folio)
+{
+	return mpage_read_folio(folio, jfs_get_block);
+}
+
+static void jfs_readahead(struct readahead_control *rac)
+{
+	mpage_readahead(rac, jfs_get_block);
+}
+
+static void jfs_write_failed(struct address_space *mapping, loff_t to)
+{
+	struct inode *inode = mapping->host;
+
+	if (to > inode->i_size) {
+		truncate_pagecache(inode, inode->i_size);
+		jfs_truncate(inode);
+	}
+}
+
+static int jfs_write_begin(const struct kiocb *iocb,
+			   struct address_space *mapping,
+			   loff_t pos, unsigned len,
+			   struct folio **foliop, void **fsdata)
+{
+	int ret;
+
+	ret = block_write_begin(mapping, pos, len, foliop, jfs_get_block);
+	if (unlikely(ret))
+		jfs_write_failed(mapping, pos + len);
+
+	return ret;
+}
+
+static int jfs_write_end(const struct kiocb *iocb,
+			 struct address_space *mapping,
+			 loff_t pos, unsigned len, unsigned copied,
+			 struct folio *folio, void *fsdata)
+{
+	int ret;
+
+	ret = generic_write_end(iocb, mapping, pos, len, copied, folio, fsdata);
+	if (ret < len)
+		jfs_write_failed(mapping, pos + len);
+	return ret;
+}
+
+static sector_t jfs_bmap(struct address_space *mapping, sector_t block)
+{
+	return generic_block_bmap(mapping, block, jfs_get_block);
+}
+
+static ssize_t jfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter)
+{
+	struct file *file = iocb->ki_filp;
+	struct address_space *mapping = file->f_mapping;
+	struct inode *inode = file->f_mapping->host;
+	size_t count = iov_iter_count(iter);
+	ssize_t ret;
+
+	ret = blockdev_direct_IO(iocb, inode, iter, jfs_get_block);
+
+	/*
+	 * In case of error extending write may have instantiated a few
+	 * blocks outside i_size. Trim these off again.
+	 */
+	if (unlikely(iov_iter_rw(iter) == WRITE && ret < 0)) {
+		loff_t isize = i_size_read(inode);
+		loff_t end = iocb->ki_pos + count;
+
+		if (end > isize)
+			jfs_write_failed(mapping, end);
+	}
+
+	return ret;
+}
+
+const struct address_space_operations jfs_aops = {
+	.dirty_folio	= block_dirty_folio,
+	.invalidate_folio = block_invalidate_folio,
+	.read_folio	= jfs_read_folio,
+	.readahead	= jfs_readahead,
+	.writepages	= jfs_writepages,
+	.write_begin	= jfs_write_begin,
+	.write_end	= jfs_write_end,
+	.bmap		= jfs_bmap,
+	.direct_IO	= jfs_direct_IO,
+	.migrate_folio	= buffer_migrate_folio,
+};
+
+/*
+ * Guts of jfs_truncate.  Called with locks already held.  Can be called
+ * with directory for truncating directory index table.
+ */
+void jfs_truncate_nolock(struct inode *ip, loff_t length)
+{
+	loff_t newsize;
+	tid_t tid;
+
+	ASSERT(length >= 0);
+
+	if (test_cflag(COMMIT_Nolink, ip) || isReadOnly(ip)) {
+		xtTruncate(0, ip, length, COMMIT_WMAP);
+		return;
+	}
+
+	do {
+		tid = txBegin(ip->i_sb, 0);
+
+		/*
+		 * The commit_mutex cannot be taken before txBegin.
+		 * txBegin may block and there is a chance the inode
+		 * could be marked dirty and need to be committed
+		 * before txBegin unblocks
+		 */
+		mutex_lock(&JFS_IP(ip)->commit_mutex);
+
+		newsize = xtTruncate(tid, ip, length,
+				     COMMIT_TRUNCATE | COMMIT_PWMAP);
+		if (newsize < 0) {
+			txEnd(tid);
+			mutex_unlock(&JFS_IP(ip)->commit_mutex);
+			break;
+		}
+
+		inode_set_mtime_to_ts(ip, inode_set_ctime_current(ip));
+		mark_inode_dirty(ip);
+
+		txCommit(tid, 1, &ip, 0);
+		txEnd(tid);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
+	} while (newsize > length);	/* Truncate isn't always atomic */
+}
+
+void jfs_truncate(struct inode *ip)
+{
+	jfs_info("jfs_truncate: size = 0x%lx", (ulong) ip->i_size);
+
+	block_truncate_page(ip->i_mapping, ip->i_size, jfs_get_block);
+
+	IWRITE_LOCK(ip, RDWRLOCK_NORMAL);
+	jfs_truncate_nolock(ip, ip->i_size);
+	IWRITE_UNLOCK(ip);
+}