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, 289 insertions, 0 deletions

diff --git a/net/wireless/mesh.c b/net/wireless/mesh.c
new file mode 100644
index 000000000..2c6654075
--- /dev/null
+++ b/net/wireless/mesh.c
@@ -0,0 +1,289 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Portions
+ * Copyright (C) 2022-2024 Intel Corporation
+ */
+#include <linux/ieee80211.h>
+#include <linux/export.h>
+#include <net/cfg80211.h>
+#include "nl80211.h"
+#include "core.h"
+#include "rdev-ops.h"
+
+/* Default values, timeouts in ms */
+#define MESH_TTL 		31
+#define MESH_DEFAULT_ELEMENT_TTL 31
+#define MESH_MAX_RETR	 	3
+#define MESH_RET_T 		100
+#define MESH_CONF_T 		100
+#define MESH_HOLD_T 		100
+
+#define MESH_PATH_TIMEOUT	5000
+#define MESH_RANN_INTERVAL      5000
+#define MESH_PATH_TO_ROOT_TIMEOUT      6000
+#define MESH_ROOT_INTERVAL     5000
+#define MESH_ROOT_CONFIRMATION_INTERVAL 2000
+#define MESH_DEFAULT_PLINK_TIMEOUT	1800 /* timeout in seconds */
+
+/*
+ * Minimum interval between two consecutive PREQs originated by the same
+ * interface
+ */
+#define MESH_PREQ_MIN_INT	10
+#define MESH_PERR_MIN_INT	100
+#define MESH_DIAM_TRAVERSAL_TIME 50
+
+#define MESH_RSSI_THRESHOLD	0
+
+/*
+ * A path will be refreshed if it is used PATH_REFRESH_TIME milliseconds
+ * before timing out.  This way it will remain ACTIVE and no data frames
+ * will be unnecessarily held in the pending queue.
+ */
+#define MESH_PATH_REFRESH_TIME			1000
+#define MESH_MIN_DISCOVERY_TIMEOUT (2 * MESH_DIAM_TRAVERSAL_TIME)
+
+/* Default maximum number of established plinks per interface */
+#define MESH_MAX_ESTAB_PLINKS	32
+
+#define MESH_MAX_PREQ_RETRIES	4
+
+#define MESH_SYNC_NEIGHBOR_OFFSET_MAX 50
+
+#define MESH_DEFAULT_BEACON_INTERVAL	1000	/* in 1024 us units (=TUs) */
+#define MESH_DEFAULT_DTIM_PERIOD	2
+#define MESH_DEFAULT_AWAKE_WINDOW	10	/* in 1024 us units (=TUs) */
+
+const struct mesh_config default_mesh_config = {
+	.dot11MeshRetryTimeout = MESH_RET_T,
+	.dot11MeshConfirmTimeout = MESH_CONF_T,
+	.dot11MeshHoldingTimeout = MESH_HOLD_T,
+	.dot11MeshMaxRetries = MESH_MAX_RETR,
+	.dot11MeshTTL = MESH_TTL,
+	.element_ttl = MESH_DEFAULT_ELEMENT_TTL,
+	.auto_open_plinks = true,
+	.dot11MeshMaxPeerLinks = MESH_MAX_ESTAB_PLINKS,
+	.dot11MeshNbrOffsetMaxNeighbor = MESH_SYNC_NEIGHBOR_OFFSET_MAX,
+	.dot11MeshHWMPactivePathTimeout = MESH_PATH_TIMEOUT,
+	.dot11MeshHWMPpreqMinInterval = MESH_PREQ_MIN_INT,
+	.dot11MeshHWMPperrMinInterval = MESH_PERR_MIN_INT,
+	.dot11MeshHWMPnetDiameterTraversalTime = MESH_DIAM_TRAVERSAL_TIME,
+	.dot11MeshHWMPmaxPREQretries = MESH_MAX_PREQ_RETRIES,
+	.path_refresh_time = MESH_PATH_REFRESH_TIME,
+	.min_discovery_timeout = MESH_MIN_DISCOVERY_TIMEOUT,
+	.dot11MeshHWMPRannInterval = MESH_RANN_INTERVAL,
+	.dot11MeshGateAnnouncementProtocol = false,
+	.dot11MeshForwarding = true,
+	.rssi_threshold = MESH_RSSI_THRESHOLD,
+	.ht_opmode = IEEE80211_HT_OP_MODE_PROTECTION_NONHT_MIXED,
+	.dot11MeshHWMPactivePathToRootTimeout = MESH_PATH_TO_ROOT_TIMEOUT,
+	.dot11MeshHWMProotInterval = MESH_ROOT_INTERVAL,
+	.dot11MeshHWMPconfirmationInterval = MESH_ROOT_CONFIRMATION_INTERVAL,
+	.power_mode = NL80211_MESH_POWER_ACTIVE,
+	.dot11MeshAwakeWindowDuration = MESH_DEFAULT_AWAKE_WINDOW,
+	.plink_timeout = MESH_DEFAULT_PLINK_TIMEOUT,
+	.dot11MeshNolearn = false,
+};
+
+const struct mesh_setup default_mesh_setup = {
+	/* cfg80211_join_mesh() will pick a channel if needed */
+	.sync_method = IEEE80211_SYNC_METHOD_NEIGHBOR_OFFSET,
+	.path_sel_proto = IEEE80211_PATH_PROTOCOL_HWMP,
+	.path_metric = IEEE80211_PATH_METRIC_AIRTIME,
+	.auth_id = 0, /* open */
+	.ie = NULL,
+	.ie_len = 0,
+	.is_secure = false,
+	.user_mpm = false,
+	.beacon_interval = MESH_DEFAULT_BEACON_INTERVAL,
+	.dtim_period = MESH_DEFAULT_DTIM_PERIOD,
+};
+
+int __cfg80211_join_mesh(struct cfg80211_registered_device *rdev,
+			 struct net_device *dev,
+			 struct mesh_setup *setup,
+			 const struct mesh_config *conf)
+{
+	struct wireless_dev *wdev = dev->ieee80211_ptr;
+	int err;
+
+	BUILD_BUG_ON(IEEE80211_MAX_SSID_LEN != IEEE80211_MAX_MESH_ID_LEN);
+
+	lockdep_assert_wiphy(wdev->wiphy);
+
+	if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_MESH_POINT)
+		return -EOPNOTSUPP;
+
+	if (!(rdev->wiphy.flags & WIPHY_FLAG_MESH_AUTH) &&
+	      setup->is_secure)
+		return -EOPNOTSUPP;
+
+	if (wdev->u.mesh.id_len)
+		return -EALREADY;
+
+	if (!setup->mesh_id_len)
+		return -EINVAL;
+
+	if (!rdev->ops->join_mesh)
+		return -EOPNOTSUPP;
+
+	if (wdev->links[0].cac_started)
+		return -EBUSY;
+
+	if (!setup->chandef.chan) {
+		/* if no channel explicitly given, use preset channel */
+		setup->chandef = wdev->u.mesh.preset_chandef;
+	}
+
+	if (!setup->chandef.chan) {
+		/* if we don't have that either, use the first usable channel */
+		enum nl80211_band band;
+
+		for (band = 0; band < NUM_NL80211_BANDS; band++) {
+			struct ieee80211_supported_band *sband;
+			struct ieee80211_channel *chan;
+			int i;
+
+			sband = rdev->wiphy.bands[band];
+			if (!sband)
+				continue;
+
+			for (i = 0; i < sband->n_channels; i++) {
+				chan = &sband->channels[i];
+				if (chan->flags & (IEEE80211_CHAN_NO_IR |
+						   IEEE80211_CHAN_DISABLED |
+						   IEEE80211_CHAN_RADAR))
+					continue;
+				setup->chandef.chan = chan;
+				break;
+			}
+
+			if (setup->chandef.chan)
+				break;
+		}
+
+		/* no usable channel ... */
+		if (!setup->chandef.chan)
+			return -EINVAL;
+
+		setup->chandef.width = NL80211_CHAN_WIDTH_20_NOHT;
+		setup->chandef.center_freq1 = setup->chandef.chan->center_freq;
+	}
+
+	/*
+	 * check if basic rates are available otherwise use mandatory rates as
+	 * basic rates
+	 */
+	if (!setup->basic_rates) {
+		struct ieee80211_supported_band *sband =
+				rdev->wiphy.bands[setup->chandef.chan->band];
+
+		if (setup->chandef.chan->band == NL80211_BAND_2GHZ) {
+			int i;
+
+			/*
+			 * Older versions selected the mandatory rates for
+			 * 2.4 GHz as well, but were broken in that only
+			 * 1 Mbps was regarded as a mandatory rate. Keep
+			 * using just 1 Mbps as the default basic rate for
+			 * mesh to be interoperable with older versions.
+			 */
+			for (i = 0; i < sband->n_bitrates; i++) {
+				if (sband->bitrates[i].bitrate == 10) {
+					setup->basic_rates = BIT(i);
+					break;
+				}
+			}
+		} else {
+			setup->basic_rates = ieee80211_mandatory_rates(sband);
+		}
+	}
+
+	err = cfg80211_chandef_dfs_required(&rdev->wiphy,
+					    &setup->chandef,
+					    NL80211_IFTYPE_MESH_POINT);
+	if (err < 0)
+		return err;
+	if (err > 0 && !setup->userspace_handles_dfs)
+		return -EINVAL;
+
+	if (!cfg80211_reg_can_beacon(&rdev->wiphy, &setup->chandef,
+				     NL80211_IFTYPE_MESH_POINT))
+		return -EINVAL;
+
+	err = rdev_join_mesh(rdev, dev, conf, setup);
+	if (!err) {
+		memcpy(wdev->u.mesh.id, setup->mesh_id, setup->mesh_id_len);
+		wdev->u.mesh.id_len = setup->mesh_id_len;
+		wdev->u.mesh.chandef = setup->chandef;
+		wdev->u.mesh.beacon_interval = setup->beacon_interval;
+	}
+
+	return err;
+}
+
+int cfg80211_set_mesh_channel(struct cfg80211_registered_device *rdev,
+			      struct wireless_dev *wdev,
+			      struct cfg80211_chan_def *chandef)
+{
+	int err;
+
+	/*
+	 * Workaround for libertas (only!), it puts the interface
+	 * into mesh mode but doesn't implement join_mesh. Instead,
+	 * it is configured via sysfs and then joins the mesh when
+	 * you set the channel. Note that the libertas mesh isn't
+	 * compatible with 802.11 mesh.
+	 */
+	if (rdev->ops->libertas_set_mesh_channel) {
+		if (chandef->width != NL80211_CHAN_WIDTH_20_NOHT)
+			return -EINVAL;
+
+		if (!netif_running(wdev->netdev))
+			return -ENETDOWN;
+
+		err = rdev_libertas_set_mesh_channel(rdev, wdev->netdev,
+						     chandef->chan);
+		if (!err)
+			wdev->u.mesh.chandef = *chandef;
+
+		return err;
+	}
+
+	if (wdev->u.mesh.id_len)
+		return -EBUSY;
+
+	wdev->u.mesh.preset_chandef = *chandef;
+	return 0;
+}
+
+int cfg80211_leave_mesh(struct cfg80211_registered_device *rdev,
+			struct net_device *dev)
+{
+	struct wireless_dev *wdev = dev->ieee80211_ptr;
+	int err;
+
+	lockdep_assert_wiphy(wdev->wiphy);
+
+	if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_MESH_POINT)
+		return -EOPNOTSUPP;
+
+	if (!rdev->ops->leave_mesh)
+		return -EOPNOTSUPP;
+
+	if (!wdev->u.mesh.id_len)
+		return -ENOTCONN;
+
+	err = rdev_leave_mesh(rdev, dev);
+	if (!err) {
+		wdev->conn_owner_nlportid = 0;
+		wdev->u.mesh.id_len = 0;
+		wdev->u.mesh.beacon_interval = 0;
+		memset(&wdev->u.mesh.chandef, 0,
+		       sizeof(wdev->u.mesh.chandef));
+		rdev_set_qos_map(rdev, dev, NULL);
+		cfg80211_sched_dfs_chan_update(rdev);
+	}
+
+	return err;
+}