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, 254 insertions, 0 deletions

diff --git a/net/mac80211/ethtool.c b/net/mac80211/ethtool.c
new file mode 100644
index 000000000..3d365626f
--- /dev/null
+++ b/net/mac80211/ethtool.c
@@ -0,0 +1,254 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * mac80211 ethtool hooks for cfg80211
+ *
+ * Copied from cfg.c - originally
+ * Copyright 2006-2010	Johannes Berg <johannes@sipsolutions.net>
+ * Copyright 2014	Intel Corporation (Author: Johannes Berg)
+ * Copyright (C) 2018, 2022-2023 Intel Corporation
+ */
+#include <linux/types.h>
+#include <net/cfg80211.h>
+#include "ieee80211_i.h"
+#include "sta_info.h"
+#include "driver-ops.h"
+
+static int ieee80211_set_ringparam(struct net_device *dev,
+				   struct ethtool_ringparam *rp,
+				   struct kernel_ethtool_ringparam *kernel_rp,
+				   struct netlink_ext_ack *extack)
+{
+	struct ieee80211_local *local = wiphy_priv(dev->ieee80211_ptr->wiphy);
+
+	if (rp->rx_mini_pending != 0 || rp->rx_jumbo_pending != 0)
+		return -EINVAL;
+
+	guard(wiphy)(local->hw.wiphy);
+
+	return drv_set_ringparam(local, rp->tx_pending, rp->rx_pending);
+}
+
+static void ieee80211_get_ringparam(struct net_device *dev,
+				    struct ethtool_ringparam *rp,
+				    struct kernel_ethtool_ringparam *kernel_rp,
+				    struct netlink_ext_ack *extack)
+{
+	struct ieee80211_local *local = wiphy_priv(dev->ieee80211_ptr->wiphy);
+
+	memset(rp, 0, sizeof(*rp));
+
+	guard(wiphy)(local->hw.wiphy);
+
+	drv_get_ringparam(local, &rp->tx_pending, &rp->tx_max_pending,
+			  &rp->rx_pending, &rp->rx_max_pending);
+}
+
+static const char ieee80211_gstrings_sta_stats[][ETH_GSTRING_LEN] = {
+	"rx_packets", "rx_bytes",
+	"rx_duplicates", "rx_fragments", "rx_dropped",
+	"tx_packets", "tx_bytes",
+	"tx_filtered", "tx_retry_failed", "tx_retries",
+	"tx_handlers_drop", "sta_state", "txrate", "rxrate",
+	"signal", "channel", "noise", "ch_time", "ch_time_busy",
+	"ch_time_ext_busy", "ch_time_rx", "ch_time_tx"
+};
+#define STA_STATS_LEN	ARRAY_SIZE(ieee80211_gstrings_sta_stats)
+
+static int ieee80211_get_sset_count(struct net_device *dev, int sset)
+{
+	struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
+	int rv = 0;
+
+	if (sset == ETH_SS_STATS)
+		rv += STA_STATS_LEN;
+
+	rv += drv_get_et_sset_count(sdata, sset);
+
+	if (rv == 0)
+		return -EOPNOTSUPP;
+	return rv;
+}
+
+static void ieee80211_get_stats(struct net_device *dev,
+				struct ethtool_stats *stats,
+				u64 *data)
+{
+	struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
+	struct ieee80211_chanctx_conf *chanctx_conf;
+	struct ieee80211_channel *channel;
+	struct sta_info *sta;
+	struct ieee80211_local *local = sdata->local;
+	struct station_info sinfo;
+	struct survey_info survey;
+	int i, q;
+#define STA_STATS_SURVEY_LEN 7
+
+	memset(data, 0, sizeof(u64) * STA_STATS_LEN);
+
+#define ADD_STA_STATS(sta)					\
+	do {							\
+		data[i++] += sinfo.rx_packets;			\
+		data[i++] += sinfo.rx_bytes;			\
+		data[i++] += (sta)->rx_stats.num_duplicates;	\
+		data[i++] += (sta)->rx_stats.fragments;		\
+		data[i++] += sinfo.rx_dropped_misc;		\
+								\
+		data[i++] += sinfo.tx_packets;			\
+		data[i++] += sinfo.tx_bytes;			\
+		data[i++] += (sta)->status_stats.filtered;	\
+		data[i++] += sinfo.tx_failed;			\
+		data[i++] += sinfo.tx_retries;			\
+	} while (0)
+
+	/* For Managed stations, find the single station based on BSSID
+	 * and use that.  For interface types, iterate through all available
+	 * stations and add stats for any station that is assigned to this
+	 * network device.
+	 */
+
+	guard(wiphy)(local->hw.wiphy);
+
+	if (sdata->vif.type == NL80211_IFTYPE_STATION) {
+		sta = sta_info_get_bss(sdata, sdata->deflink.u.mgd.bssid);
+
+		if (!(sta && !WARN_ON(sta->sdata->dev != dev)))
+			goto do_survey;
+
+		memset(&sinfo, 0, sizeof(sinfo));
+		sta_set_sinfo(sta, &sinfo, false);
+
+		i = 0;
+		ADD_STA_STATS(&sta->deflink);
+
+		data[i++] = sdata->tx_handlers_drop;
+		data[i++] = sta->sta_state;
+
+
+		if (sinfo.filled & BIT_ULL(NL80211_STA_INFO_TX_BITRATE))
+			data[i] = 100000ULL *
+				cfg80211_calculate_bitrate(&sinfo.txrate);
+		i++;
+		if (sinfo.filled & BIT_ULL(NL80211_STA_INFO_RX_BITRATE))
+			data[i] = 100000ULL *
+				cfg80211_calculate_bitrate(&sinfo.rxrate);
+		i++;
+
+		if (sinfo.filled & BIT_ULL(NL80211_STA_INFO_SIGNAL_AVG))
+			data[i] = (u8)sinfo.signal_avg;
+		i++;
+	} else {
+		list_for_each_entry(sta, &local->sta_list, list) {
+			/* Make sure this station belongs to the proper dev */
+			if (sta->sdata->dev != dev)
+				continue;
+
+			memset(&sinfo, 0, sizeof(sinfo));
+			sta_set_sinfo(sta, &sinfo, false);
+			i = 0;
+			ADD_STA_STATS(&sta->deflink);
+			data[i++] = sdata->tx_handlers_drop;
+		}
+	}
+
+do_survey:
+	i = STA_STATS_LEN - STA_STATS_SURVEY_LEN;
+	/* Get survey stats for current channel */
+	survey.filled = 0;
+
+	rcu_read_lock();
+	chanctx_conf = rcu_dereference(sdata->vif.bss_conf.chanctx_conf);
+	if (chanctx_conf)
+		channel = chanctx_conf->def.chan;
+	else if (local->open_count > 0 &&
+		 local->open_count == local->virt_monitors &&
+		 sdata->vif.type == NL80211_IFTYPE_MONITOR)
+		channel = local->monitor_chanreq.oper.chan;
+	else
+		channel = NULL;
+	rcu_read_unlock();
+
+	if (channel) {
+		q = 0;
+		do {
+			survey.filled = 0;
+			if (drv_get_survey(local, q, &survey) != 0) {
+				survey.filled = 0;
+				break;
+			}
+			q++;
+		} while (channel != survey.channel);
+	}
+
+	if (survey.filled)
+		data[i++] = survey.channel->center_freq;
+	else
+		data[i++] = 0;
+	if (survey.filled & SURVEY_INFO_NOISE_DBM)
+		data[i++] = (u8)survey.noise;
+	else
+		data[i++] = -1LL;
+	if (survey.filled & SURVEY_INFO_TIME)
+		data[i++] = survey.time;
+	else
+		data[i++] = -1LL;
+	if (survey.filled & SURVEY_INFO_TIME_BUSY)
+		data[i++] = survey.time_busy;
+	else
+		data[i++] = -1LL;
+	if (survey.filled & SURVEY_INFO_TIME_EXT_BUSY)
+		data[i++] = survey.time_ext_busy;
+	else
+		data[i++] = -1LL;
+	if (survey.filled & SURVEY_INFO_TIME_RX)
+		data[i++] = survey.time_rx;
+	else
+		data[i++] = -1LL;
+	if (survey.filled & SURVEY_INFO_TIME_TX)
+		data[i++] = survey.time_tx;
+	else
+		data[i++] = -1LL;
+
+	if (WARN_ON(i != STA_STATS_LEN))
+		return;
+
+	drv_get_et_stats(sdata, stats, &(data[STA_STATS_LEN]));
+}
+
+static void ieee80211_get_strings(struct net_device *dev, u32 sset, u8 *data)
+{
+	struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
+	int sz_sta_stats = 0;
+
+	if (sset == ETH_SS_STATS) {
+		sz_sta_stats = sizeof(ieee80211_gstrings_sta_stats);
+		memcpy(data, ieee80211_gstrings_sta_stats, sz_sta_stats);
+	}
+	drv_get_et_strings(sdata, sset, &(data[sz_sta_stats]));
+}
+
+static int ieee80211_get_regs_len(struct net_device *dev)
+{
+	return 0;
+}
+
+static void ieee80211_get_regs(struct net_device *dev,
+			       struct ethtool_regs *regs,
+			       void *data)
+{
+	struct wireless_dev *wdev = dev->ieee80211_ptr;
+
+	regs->version = wdev->wiphy->hw_version;
+	regs->len = 0;
+}
+
+const struct ethtool_ops ieee80211_ethtool_ops = {
+	.get_drvinfo = cfg80211_get_drvinfo,
+	.get_regs_len = ieee80211_get_regs_len,
+	.get_regs = ieee80211_get_regs,
+	.get_link = ethtool_op_get_link,
+	.get_ringparam = ieee80211_get_ringparam,
+	.set_ringparam = ieee80211_set_ringparam,
+	.get_strings = ieee80211_get_strings,
+	.get_ethtool_stats = ieee80211_get_stats,
+	.get_sset_count = ieee80211_get_sset_count,
+};