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, 340 insertions, 0 deletions

diff --git a/fs/afs/vl_alias.c b/fs/afs/vl_alias.c
new file mode 100644
index 000000000..fc9676abd
--- /dev/null
+++ b/fs/afs/vl_alias.c
@@ -0,0 +1,340 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* AFS cell alias detection
+ *
+ * Copyright (C) 2020 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ */
+
+#include <linux/slab.h>
+#include <linux/sched.h>
+#include <linux/namei.h>
+#include <keys/rxrpc-type.h>
+#include "internal.h"
+
+/*
+ * Sample a volume.
+ */
+static struct afs_volume *afs_sample_volume(struct afs_cell *cell, struct key *key,
+					    const char *name, unsigned int namelen)
+{
+	struct afs_volume *volume;
+	struct afs_fs_context fc = {
+		.type		= 0, /* Explicitly leave it to the VLDB */
+		.volnamesz	= namelen,
+		.volname	= name,
+		.net		= cell->net,
+		.cell		= cell,
+		.key		= key, /* This might need to be something */
+	};
+
+	volume = afs_create_volume(&fc);
+	_leave(" = %p", volume);
+	return volume;
+}
+
+/*
+ * Compare the address lists of a pair of fileservers.
+ */
+static int afs_compare_fs_alists(const struct afs_server *server_a,
+				 const struct afs_server *server_b)
+{
+	const struct afs_addr_list *la, *lb;
+	int a = 0, b = 0, addr_matches = 0;
+
+	la = rcu_dereference(server_a->endpoint_state)->addresses;
+	lb = rcu_dereference(server_b->endpoint_state)->addresses;
+
+	while (a < la->nr_addrs && b < lb->nr_addrs) {
+		unsigned long pa = (unsigned long)la->addrs[a].peer;
+		unsigned long pb = (unsigned long)lb->addrs[b].peer;
+		long diff = pa - pb;
+
+		if (diff < 0) {
+			a++;
+		} else if (diff > 0) {
+			b++;
+		} else {
+			addr_matches++;
+			a++;
+			b++;
+		}
+	}
+
+	return addr_matches;
+}
+
+/*
+ * Compare the fileserver lists of two volumes.  The server lists are sorted in
+ * order of ascending UUID.
+ */
+static int afs_compare_volume_slists(const struct afs_volume *vol_a,
+				     const struct afs_volume *vol_b)
+{
+	const struct afs_server_list *la, *lb;
+	int i, a = 0, b = 0, uuid_matches = 0, addr_matches = 0;
+
+	la = rcu_dereference(vol_a->servers);
+	lb = rcu_dereference(vol_b->servers);
+
+	for (i = 0; i < AFS_MAXTYPES; i++)
+		if (vol_a->vids[i] != vol_b->vids[i])
+			return 0;
+
+	while (a < la->nr_servers && b < lb->nr_servers) {
+		const struct afs_server *server_a = la->servers[a].server;
+		const struct afs_server *server_b = lb->servers[b].server;
+		int diff = memcmp(&server_a->uuid, &server_b->uuid, sizeof(uuid_t));
+
+		if (diff < 0) {
+			a++;
+		} else if (diff > 0) {
+			b++;
+		} else {
+			uuid_matches++;
+			addr_matches += afs_compare_fs_alists(server_a, server_b);
+			a++;
+			b++;
+		}
+	}
+
+	_leave(" = %d [um %d]", addr_matches, uuid_matches);
+	return addr_matches;
+}
+
+/*
+ * Compare root.cell volumes.
+ */
+static int afs_compare_cell_roots(struct afs_cell *cell)
+{
+	struct afs_cell *p;
+
+	_enter("");
+
+	rcu_read_lock();
+
+	hlist_for_each_entry_rcu(p, &cell->net->proc_cells, proc_link) {
+		if (p == cell || p->alias_of)
+			continue;
+		if (!p->root_volume)
+			continue; /* Ignore cells that don't have a root.cell volume. */
+
+		if (afs_compare_volume_slists(cell->root_volume, p->root_volume) != 0)
+			goto is_alias;
+	}
+
+	rcu_read_unlock();
+	_leave(" = 0");
+	return 0;
+
+is_alias:
+	rcu_read_unlock();
+	cell->alias_of = afs_use_cell(p, afs_cell_trace_use_alias);
+	return 1;
+}
+
+/*
+ * Query the new cell for a volume from a cell we're already using.
+ */
+static int afs_query_for_alias_one(struct afs_cell *cell, struct key *key,
+				   struct afs_cell *p)
+{
+	struct afs_volume *volume, *pvol = NULL;
+	int ret;
+
+	/* Arbitrarily pick a volume from the list. */
+	read_seqlock_excl(&p->volume_lock);
+	if (!RB_EMPTY_ROOT(&p->volumes))
+		pvol = afs_get_volume(rb_entry(p->volumes.rb_node,
+					       struct afs_volume, cell_node),
+				      afs_volume_trace_get_query_alias);
+	read_sequnlock_excl(&p->volume_lock);
+	if (!pvol)
+		return 0;
+
+	_enter("%s:%s", cell->name, pvol->name);
+
+	/* And see if it's in the new cell. */
+	volume = afs_sample_volume(cell, key, pvol->name, pvol->name_len);
+	if (IS_ERR(volume)) {
+		afs_put_volume(pvol, afs_volume_trace_put_query_alias);
+		if (PTR_ERR(volume) != -ENOMEDIUM)
+			return PTR_ERR(volume);
+		/* That volume is not in the new cell, so not an alias */
+		return 0;
+	}
+
+	/* The new cell has a like-named volume also - compare volume ID,
+	 * server and address lists.
+	 */
+	ret = 0;
+	if (pvol->vid == volume->vid) {
+		rcu_read_lock();
+		if (afs_compare_volume_slists(volume, pvol))
+			ret = 1;
+		rcu_read_unlock();
+	}
+
+	afs_put_volume(volume, afs_volume_trace_put_query_alias);
+	afs_put_volume(pvol, afs_volume_trace_put_query_alias);
+	return ret;
+}
+
+/*
+ * Query the new cell for volumes we know exist in cells we're already using.
+ */
+static int afs_query_for_alias(struct afs_cell *cell, struct key *key)
+{
+	struct afs_cell *p;
+
+	_enter("%s", cell->name);
+
+	if (mutex_lock_interruptible(&cell->net->proc_cells_lock) < 0)
+		return -ERESTARTSYS;
+
+	hlist_for_each_entry(p, &cell->net->proc_cells, proc_link) {
+		if (p == cell || p->alias_of)
+			continue;
+		if (RB_EMPTY_ROOT(&p->volumes))
+			continue;
+		if (p->root_volume)
+			continue; /* Ignore cells that have a root.cell volume. */
+		afs_use_cell(p, afs_cell_trace_use_check_alias);
+		mutex_unlock(&cell->net->proc_cells_lock);
+
+		if (afs_query_for_alias_one(cell, key, p) != 0)
+			goto is_alias;
+
+		if (mutex_lock_interruptible(&cell->net->proc_cells_lock) < 0) {
+			afs_unuse_cell(p, afs_cell_trace_unuse_check_alias);
+			return -ERESTARTSYS;
+		}
+
+		afs_unuse_cell(p, afs_cell_trace_unuse_check_alias);
+	}
+
+	mutex_unlock(&cell->net->proc_cells_lock);
+	_leave(" = 0");
+	return 0;
+
+is_alias:
+	cell->alias_of = p; /* Transfer our ref */
+	return 1;
+}
+
+/*
+ * Look up a VLDB record for a volume.
+ */
+static char *afs_vl_get_cell_name(struct afs_cell *cell, struct key *key)
+{
+	struct afs_vl_cursor vc;
+	char *cell_name = ERR_PTR(-EDESTADDRREQ);
+	bool skipped = false, not_skipped = false;
+	int ret;
+
+	if (!afs_begin_vlserver_operation(&vc, cell, key))
+		return ERR_PTR(-ERESTARTSYS);
+
+	while (afs_select_vlserver(&vc)) {
+		if (!test_bit(AFS_VLSERVER_FL_IS_YFS, &vc.server->flags)) {
+			vc.call_error = -EOPNOTSUPP;
+			skipped = true;
+			continue;
+		}
+		not_skipped = true;
+		cell_name = afs_yfsvl_get_cell_name(&vc);
+	}
+
+	ret = afs_end_vlserver_operation(&vc);
+	if (skipped && !not_skipped)
+		ret = -EOPNOTSUPP;
+	return ret < 0 ? ERR_PTR(ret) : cell_name;
+}
+
+static int yfs_check_canonical_cell_name(struct afs_cell *cell, struct key *key)
+{
+	struct afs_cell *master;
+	size_t name_len;
+	char *cell_name;
+
+	cell_name = afs_vl_get_cell_name(cell, key);
+	if (IS_ERR(cell_name))
+		return PTR_ERR(cell_name);
+
+	if (strcmp(cell_name, cell->name) == 0) {
+		kfree(cell_name);
+		return 0;
+	}
+
+	name_len = strlen(cell_name);
+	if (!name_len || name_len > AFS_MAXCELLNAME)
+		master = ERR_PTR(-EOPNOTSUPP);
+	else
+		master = afs_lookup_cell(cell->net, cell_name, name_len, NULL,
+					 AFS_LOOKUP_CELL_ALIAS_CHECK,
+					 afs_cell_trace_use_lookup_canonical);
+	kfree(cell_name);
+	if (IS_ERR(master))
+		return PTR_ERR(master);
+
+	cell->alias_of = master; /* Transfer our ref */
+	return 1;
+}
+
+static int afs_do_cell_detect_alias(struct afs_cell *cell, struct key *key)
+{
+	struct afs_volume *root_volume;
+	int ret;
+
+	_enter("%s", cell->name);
+
+	ret = yfs_check_canonical_cell_name(cell, key);
+	if (ret != -EOPNOTSUPP)
+		return ret;
+
+	/* Try and get the root.cell volume for comparison with other cells */
+	root_volume = afs_sample_volume(cell, key, "root.cell", 9);
+	if (!IS_ERR(root_volume)) {
+		cell->root_volume = root_volume;
+		return afs_compare_cell_roots(cell);
+	}
+
+	if (PTR_ERR(root_volume) != -ENOMEDIUM)
+		return PTR_ERR(root_volume);
+
+	/* Okay, this cell doesn't have an root.cell volume.  We need to
+	 * locate some other random volume and use that to check.
+	 */
+	return afs_query_for_alias(cell, key);
+}
+
+/*
+ * Check to see if a new cell is an alias of a cell we already have.  At this
+ * point we have the cell's volume server list.
+ *
+ * Returns 0 if we didn't detect an alias, 1 if we found an alias and an error
+ * if we had problems gathering the data required.  In the case the we did
+ * detect an alias, cell->alias_of is set to point to the assumed master.
+ */
+int afs_cell_detect_alias(struct afs_cell *cell, struct key *key)
+{
+	struct afs_net *net = cell->net;
+	int ret;
+
+	if (mutex_lock_interruptible(&net->cells_alias_lock) < 0)
+		return -ERESTARTSYS;
+
+	if (test_bit(AFS_CELL_FL_CHECK_ALIAS, &cell->flags)) {
+		ret = afs_do_cell_detect_alias(cell, key);
+		if (ret >= 0)
+			clear_bit_unlock(AFS_CELL_FL_CHECK_ALIAS, &cell->flags);
+	} else {
+		ret = cell->alias_of ? 1 : 0;
+	}
+
+	mutex_unlock(&net->cells_alias_lock);
+
+	if (ret == 1)
+		pr_notice("kAFS: Cell %s is an alias of %s\n",
+			  cell->name, cell->alias_of->name);
+	return ret;
+}