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, 484 insertions, 0 deletions

diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
new file mode 100644
index 000000000..159295c63
--- /dev/null
+++ b/fs/xfs/xfs_itable.c
@@ -0,0 +1,484 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ */
+#include "xfs_platform.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_inode.h"
+#include "xfs_btree.h"
+#include "xfs_ialloc.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_iwalk.h"
+#include "xfs_itable.h"
+#include "xfs_error.h"
+#include "xfs_icache.h"
+#include "xfs_health.h"
+#include "xfs_trans.h"
+
+/*
+ * Bulk Stat
+ * =========
+ *
+ * Use the inode walking functions to fill out struct xfs_bulkstat for every
+ * allocated inode, then pass the stat information to some externally provided
+ * iteration function.
+ */
+
+struct xfs_bstat_chunk {
+	bulkstat_one_fmt_pf	formatter;
+	struct xfs_ibulk	*breq;
+	struct xfs_bulkstat	*buf;
+};
+
+static inline bool
+want_metadir_file(
+	struct xfs_inode	*ip,
+	struct xfs_ibulk	*breq)
+{
+	return xfs_is_metadir_inode(ip) && (breq->flags & XFS_IBULK_METADIR);
+}
+
+/*
+ * Fill out the bulkstat info for a single inode and report it somewhere.
+ *
+ * bc->breq->lastino is effectively the inode cursor as we walk through the
+ * filesystem.  Therefore, we update it any time we need to move the cursor
+ * forward, regardless of whether or not we're sending any bstat information
+ * back to userspace.  If the inode is internal metadata or, has been freed
+ * out from under us, we just simply keep going.
+ *
+ * However, if any other type of error happens we want to stop right where we
+ * are so that userspace will call back with exact number of the bad inode and
+ * we can send back an error code.
+ *
+ * Note that if the formatter tells us there's no space left in the buffer we
+ * move the cursor forward and abort the walk.
+ */
+STATIC int
+xfs_bulkstat_one_int(
+	struct xfs_mount	*mp,
+	struct mnt_idmap	*idmap,
+	struct xfs_trans	*tp,
+	xfs_ino_t		ino,
+	struct xfs_bstat_chunk	*bc)
+{
+	struct user_namespace	*sb_userns = mp->m_super->s_user_ns;
+	struct xfs_inode	*ip;		/* incore inode pointer */
+	struct inode		*inode;
+	struct xfs_bulkstat	*buf = bc->buf;
+	xfs_extnum_t		nextents;
+	int			error = -EINVAL;
+	vfsuid_t		vfsuid;
+	vfsgid_t		vfsgid;
+
+	error = xfs_iget(mp, tp, ino,
+			 (XFS_IGET_DONTCACHE | XFS_IGET_UNTRUSTED),
+			 XFS_ILOCK_SHARED, &ip);
+	if (error == -ENOENT || error == -EINVAL)
+		goto out_advance;
+	if (error)
+		goto out;
+
+	/* Reload the incore unlinked list to avoid failure in inodegc. */
+	if (xfs_inode_unlinked_incomplete(ip)) {
+		error = xfs_inode_reload_unlinked_bucket(tp, ip);
+		if (error) {
+			xfs_iunlock(ip, XFS_ILOCK_SHARED);
+			xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+			xfs_irele(ip);
+			return error;
+		}
+	}
+
+	ASSERT(ip != NULL);
+	ASSERT(ip->i_imap.im_agbno != 0);
+	inode = VFS_I(ip);
+	vfsuid = i_uid_into_vfsuid(idmap, inode);
+	vfsgid = i_gid_into_vfsgid(idmap, inode);
+
+	/*
+	 * If caller wants files from the metadata directories, push out the
+	 * bare minimum information for enabling scrub.
+	 */
+	if (want_metadir_file(ip, bc->breq)) {
+		memset(buf, 0, sizeof(*buf));
+		buf->bs_ino = ino;
+		buf->bs_gen = inode->i_generation;
+		buf->bs_mode = inode->i_mode & S_IFMT;
+		xfs_bulkstat_health(ip, buf);
+		buf->bs_version = XFS_BULKSTAT_VERSION_V5;
+		xfs_iunlock(ip, XFS_ILOCK_SHARED);
+		xfs_irele(ip);
+
+		error = bc->formatter(bc->breq, buf);
+		if (!error || error == -ECANCELED)
+			goto out_advance;
+		goto out;
+	}
+
+	/* If this is a private inode, don't leak its details to userspace. */
+	if (IS_PRIVATE(inode) || xfs_is_sb_inum(mp, ino)) {
+		xfs_iunlock(ip, XFS_ILOCK_SHARED);
+		xfs_irele(ip);
+		error = -EINVAL;
+		goto out_advance;
+	}
+
+	/* xfs_iget returns the following without needing
+	 * further change.
+	 */
+	buf->bs_projectid = ip->i_projid;
+	buf->bs_ino = ino;
+	buf->bs_uid = from_kuid(sb_userns, vfsuid_into_kuid(vfsuid));
+	buf->bs_gid = from_kgid(sb_userns, vfsgid_into_kgid(vfsgid));
+	buf->bs_size = ip->i_disk_size;
+
+	buf->bs_nlink = inode->i_nlink;
+	buf->bs_atime = inode_get_atime_sec(inode);
+	buf->bs_atime_nsec = inode_get_atime_nsec(inode);
+	buf->bs_mtime = inode_get_mtime_sec(inode);
+	buf->bs_mtime_nsec = inode_get_mtime_nsec(inode);
+	buf->bs_ctime = inode_get_ctime_sec(inode);
+	buf->bs_ctime_nsec = inode_get_ctime_nsec(inode);
+	buf->bs_gen = inode->i_generation;
+	buf->bs_mode = inode->i_mode;
+
+	buf->bs_xflags = xfs_ip2xflags(ip);
+	buf->bs_extsize_blks = ip->i_extsize;
+
+	nextents = xfs_ifork_nextents(&ip->i_df);
+	if (!(bc->breq->flags & XFS_IBULK_NREXT64))
+		buf->bs_extents = min(nextents, XFS_MAX_EXTCNT_DATA_FORK_SMALL);
+	else
+		buf->bs_extents64 = nextents;
+
+	xfs_bulkstat_health(ip, buf);
+	buf->bs_aextents = xfs_ifork_nextents(&ip->i_af);
+	buf->bs_forkoff = xfs_inode_fork_boff(ip);
+	buf->bs_version = XFS_BULKSTAT_VERSION_V5;
+
+	if (xfs_has_v3inodes(mp)) {
+		buf->bs_btime = ip->i_crtime.tv_sec;
+		buf->bs_btime_nsec = ip->i_crtime.tv_nsec;
+		if (ip->i_diflags2 & XFS_DIFLAG2_COWEXTSIZE)
+			buf->bs_cowextsize_blks = ip->i_cowextsize;
+	}
+
+	switch (ip->i_df.if_format) {
+	case XFS_DINODE_FMT_DEV:
+		buf->bs_rdev = sysv_encode_dev(inode->i_rdev);
+		buf->bs_blksize = BLKDEV_IOSIZE;
+		buf->bs_blocks = 0;
+		break;
+	case XFS_DINODE_FMT_LOCAL:
+		buf->bs_rdev = 0;
+		buf->bs_blksize = mp->m_sb.sb_blocksize;
+		buf->bs_blocks = 0;
+		break;
+	case XFS_DINODE_FMT_EXTENTS:
+	case XFS_DINODE_FMT_BTREE:
+		buf->bs_rdev = 0;
+		buf->bs_blksize = mp->m_sb.sb_blocksize;
+		buf->bs_blocks = ip->i_nblocks + ip->i_delayed_blks;
+		break;
+	}
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+	xfs_irele(ip);
+
+	error = bc->formatter(bc->breq, buf);
+	if (error == -ECANCELED)
+		goto out_advance;
+	if (error)
+		goto out;
+
+out_advance:
+	/*
+	 * Advance the cursor to the inode that comes after the one we just
+	 * looked at.  We want the caller to move along if the bulkstat
+	 * information was copied successfully; if we tried to grab the inode
+	 * but it's no longer allocated; or if it's internal metadata.
+	 */
+	bc->breq->startino = ino + 1;
+out:
+	return error;
+}
+
+/* Bulkstat a single inode. */
+int
+xfs_bulkstat_one(
+	struct xfs_ibulk	*breq,
+	bulkstat_one_fmt_pf	formatter)
+{
+	struct xfs_bstat_chunk	bc = {
+		.formatter	= formatter,
+		.breq		= breq,
+	};
+	struct xfs_trans	*tp;
+	int			error;
+
+	if (breq->idmap != &nop_mnt_idmap) {
+		xfs_warn_ratelimited(breq->mp,
+			"bulkstat not supported inside of idmapped mounts.");
+		return -EINVAL;
+	}
+
+	ASSERT(breq->icount == 1);
+
+	bc.buf = kzalloc_obj(struct xfs_bulkstat,
+			     GFP_KERNEL | __GFP_RETRY_MAYFAIL);
+	if (!bc.buf)
+		return -ENOMEM;
+
+	/*
+	 * Grab an empty transaction so that we can use its recursive buffer
+	 * locking abilities to detect cycles in the inobt without deadlocking.
+	 */
+	tp = xfs_trans_alloc_empty(breq->mp);
+	error = xfs_bulkstat_one_int(breq->mp, breq->idmap, tp,
+			breq->startino, &bc);
+	xfs_trans_cancel(tp);
+	kfree(bc.buf);
+
+	/*
+	 * If we reported one inode to userspace then we abort because we hit
+	 * the end of the buffer.  Don't leak that back to userspace.
+	 */
+	if (error == -ECANCELED)
+		error = 0;
+
+	return error;
+}
+
+static int
+xfs_bulkstat_iwalk(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_ino_t		ino,
+	void			*data)
+{
+	struct xfs_bstat_chunk	*bc = data;
+	int			error;
+
+	error = xfs_bulkstat_one_int(mp, bc->breq->idmap, tp, ino, data);
+	/* bulkstat just skips over missing inodes */
+	if (error == -ENOENT || error == -EINVAL)
+		return 0;
+	return error;
+}
+
+/*
+ * Check the incoming lastino parameter.
+ *
+ * We allow any inode value that could map to physical space inside the
+ * filesystem because if there are no inodes there, bulkstat moves on to the
+ * next chunk.  In other words, the magic agino value of zero takes us to the
+ * first chunk in the AG, and an agino value past the end of the AG takes us to
+ * the first chunk in the next AG.
+ *
+ * Therefore we can end early if the requested inode is beyond the end of the
+ * filesystem or doesn't map properly.
+ */
+static inline bool
+xfs_bulkstat_already_done(
+	struct xfs_mount	*mp,
+	xfs_ino_t		startino)
+{
+	xfs_agnumber_t		agno = XFS_INO_TO_AGNO(mp, startino);
+	xfs_agino_t		agino = XFS_INO_TO_AGINO(mp, startino);
+
+	return agno >= mp->m_sb.sb_agcount ||
+	       startino != XFS_AGINO_TO_INO(mp, agno, agino);
+}
+
+/* Return stat information in bulk (by-inode) for the filesystem. */
+int
+xfs_bulkstat(
+	struct xfs_ibulk	*breq,
+	bulkstat_one_fmt_pf	formatter)
+{
+	struct xfs_bstat_chunk	bc = {
+		.formatter	= formatter,
+		.breq		= breq,
+	};
+	struct xfs_trans	*tp;
+	int			error;
+
+	if (breq->idmap != &nop_mnt_idmap) {
+		xfs_warn_ratelimited(breq->mp,
+			"bulkstat not supported inside of idmapped mounts.");
+		return -EINVAL;
+	}
+	if (xfs_bulkstat_already_done(breq->mp, breq->startino))
+		return 0;
+
+	bc.buf = kzalloc_obj(struct xfs_bulkstat,
+			     GFP_KERNEL | __GFP_RETRY_MAYFAIL);
+	if (!bc.buf)
+		return -ENOMEM;
+
+	/*
+	 * Grab an empty transaction so that we can use its recursive buffer
+	 * locking abilities to detect cycles in the inobt without deadlocking.
+	 */
+	tp = xfs_trans_alloc_empty(breq->mp);
+	error = xfs_iwalk(breq->mp, tp, breq->startino, breq->iwalk_flags,
+			xfs_bulkstat_iwalk, breq->icount, &bc);
+	xfs_trans_cancel(tp);
+	kfree(bc.buf);
+
+	/*
+	 * We found some inodes, so clear the error status and return them.
+	 * The lastino pointer will point directly at the inode that triggered
+	 * any error that occurred, so on the next call the error will be
+	 * triggered again and propagated to userspace as there will be no
+	 * formatted inodes in the buffer.
+	 */
+	if (breq->ocount > 0)
+		error = 0;
+
+	return error;
+}
+
+/* Convert bulkstat (v5) to bstat (v1). */
+void
+xfs_bulkstat_to_bstat(
+	struct xfs_mount		*mp,
+	struct xfs_bstat		*bs1,
+	const struct xfs_bulkstat	*bstat)
+{
+	/* memset is needed here because of padding holes in the structure. */
+	memset(bs1, 0, sizeof(struct xfs_bstat));
+	bs1->bs_ino = bstat->bs_ino;
+	bs1->bs_mode = bstat->bs_mode;
+	bs1->bs_nlink = bstat->bs_nlink;
+	bs1->bs_uid = bstat->bs_uid;
+	bs1->bs_gid = bstat->bs_gid;
+	bs1->bs_rdev = bstat->bs_rdev;
+	bs1->bs_blksize = bstat->bs_blksize;
+	bs1->bs_size = bstat->bs_size;
+	bs1->bs_atime.tv_sec = bstat->bs_atime;
+	bs1->bs_mtime.tv_sec = bstat->bs_mtime;
+	bs1->bs_ctime.tv_sec = bstat->bs_ctime;
+	bs1->bs_atime.tv_nsec = bstat->bs_atime_nsec;
+	bs1->bs_mtime.tv_nsec = bstat->bs_mtime_nsec;
+	bs1->bs_ctime.tv_nsec = bstat->bs_ctime_nsec;
+	bs1->bs_blocks = bstat->bs_blocks;
+	bs1->bs_xflags = bstat->bs_xflags;
+	bs1->bs_extsize = XFS_FSB_TO_B(mp, bstat->bs_extsize_blks);
+	bs1->bs_extents = bstat->bs_extents;
+	bs1->bs_gen = bstat->bs_gen;
+	bs1->bs_projid_lo = bstat->bs_projectid & 0xFFFF;
+	bs1->bs_forkoff = bstat->bs_forkoff;
+	bs1->bs_projid_hi = bstat->bs_projectid >> 16;
+	bs1->bs_sick = bstat->bs_sick;
+	bs1->bs_checked = bstat->bs_checked;
+	bs1->bs_cowextsize = XFS_FSB_TO_B(mp, bstat->bs_cowextsize_blks);
+	bs1->bs_dmevmask = 0;
+	bs1->bs_dmstate = 0;
+	bs1->bs_aextents = bstat->bs_aextents;
+}
+
+struct xfs_inumbers_chunk {
+	inumbers_fmt_pf		formatter;
+	struct xfs_ibulk	*breq;
+};
+
+/*
+ * INUMBERS
+ * ========
+ * This is how we export inode btree records to userspace, so that XFS tools
+ * can figure out where inodes are allocated.
+ */
+
+/*
+ * Format the inode group structure and report it somewhere.
+ *
+ * Similar to xfs_bulkstat_one_int, lastino is the inode cursor as we walk
+ * through the filesystem so we move it forward unless there was a runtime
+ * error.  If the formatter tells us the buffer is now full we also move the
+ * cursor forward and abort the walk.
+ */
+STATIC int
+xfs_inumbers_walk(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		agno,
+	const struct xfs_inobt_rec_incore *irec,
+	void			*data)
+{
+	struct xfs_inumbers	inogrp = {
+		.xi_startino	= XFS_AGINO_TO_INO(mp, agno, irec->ir_startino),
+		.xi_alloccount	= irec->ir_count - irec->ir_freecount,
+		.xi_allocmask	= ~irec->ir_free,
+		.xi_version	= XFS_INUMBERS_VERSION_V5,
+	};
+	struct xfs_inumbers_chunk *ic = data;
+	int			error;
+
+	error = ic->formatter(ic->breq, &inogrp);
+	if (error && error != -ECANCELED)
+		return error;
+
+	ic->breq->startino = XFS_AGINO_TO_INO(mp, agno, irec->ir_startino) +
+			XFS_INODES_PER_CHUNK;
+	return error;
+}
+
+/*
+ * Return inode number table for the filesystem.
+ */
+int
+xfs_inumbers(
+	struct xfs_ibulk	*breq,
+	inumbers_fmt_pf		formatter)
+{
+	struct xfs_inumbers_chunk ic = {
+		.formatter	= formatter,
+		.breq		= breq,
+	};
+	struct xfs_trans	*tp;
+	int			error = 0;
+
+	if (xfs_bulkstat_already_done(breq->mp, breq->startino))
+		return 0;
+
+	/*
+	 * Grab an empty transaction so that we can use its recursive buffer
+	 * locking abilities to detect cycles in the inobt without deadlocking.
+	 */
+	tp = xfs_trans_alloc_empty(breq->mp);
+	error = xfs_inobt_walk(breq->mp, tp, breq->startino, breq->iwalk_flags,
+			xfs_inumbers_walk, breq->icount, &ic);
+	xfs_trans_cancel(tp);
+
+	/*
+	 * We found some inode groups, so clear the error status and return
+	 * them.  The lastino pointer will point directly at the inode that
+	 * triggered any error that occurred, so on the next call the error
+	 * will be triggered again and propagated to userspace as there will be
+	 * no formatted inode groups in the buffer.
+	 */
+	if (breq->ocount > 0)
+		error = 0;
+
+	return error;
+}
+
+/* Convert an inumbers (v5) struct to a inogrp (v1) struct. */
+void
+xfs_inumbers_to_inogrp(
+	struct xfs_inogrp		*ig1,
+	const struct xfs_inumbers	*ig)
+{
+	/* memset is needed here because of padding holes in the structure. */
+	memset(ig1, 0, sizeof(struct xfs_inogrp));
+	ig1->xi_startino = ig->xi_startino;
+	ig1->xi_alloccount = ig->xi_alloccount;
+	ig1->xi_allocmask = ig->xi_allocmask;
+}