ARM: 9485/1: mm: acquire mmap write lock around show_pte() for user faultsgrafted

When CONFIG_DEBUG_USER=y, and cmdline "user_debug=31" is set,
a user fault may trigger show_pte() without any lock.
If another thread in the same process concurrently calls munmap(),
the page table pages may be freed while show_pte() is still traversing
them, causing a use-after-free in show_pte().

If CONFIG_ARM_LPAE=y, this may cause a kernel panic if the pages table
of PMD are freed when show_pte() is running.

Acquire mmap_write_lock() around show_pte() for user faults to fix the
contention.

For user faults, additionally restrict that show_pte() is called only
when the addr is a user-space address (addr < TASK_SIZE). This is because
the lock of tsk->mm only protects the virtual memory of user address space,
furthermore, dumping the page tables of a kernel-space address for user
faults is unnecessary and may have security implications.

Keep everything unchanged for kernel faults, because the kernel is
already in the "oops" state, acquiring a lock may risk a deadlock.

Co-developed-by: Qi Xi <xiqi2@huawei.com>

Fixes: 6d021b724481 ("ARM: dump pgd, pmd and pte states on unhandled data abort faults")
Link: https://lore.kernel.org/20260716014022.2823-1-xieyuanbin1@huawei.com
Acked-by: Lorenzo Stoakes (ARM) <ljs@kernel.org>
Reviewed-by: Linus Walleij <linusw@kernel.org>
Signed-off-by: Qi Xi <xiqi2@huawei.com>
Signed-off-by: Xie Yuanbin <xieyuanbin1@huawei.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>

1 files changed, 232 insertions, 0 deletions

diff --git a/tools/testing/fault-injection/failcmd.sh b/tools/testing/fault-injection/failcmd.sh
new file mode 100755
index 000000000..c4f243275
--- /dev/null
+++ b/tools/testing/fault-injection/failcmd.sh
@@ -0,0 +1,232 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+#
+# NAME
+#	failcmd.sh - run a command with injecting slab/page allocation failures
+#
+# SYNOPSIS
+#	failcmd.sh --help
+#	failcmd.sh [<options>] command [arguments]
+#
+# DESCRIPTION
+#	Run command with injecting slab/page allocation failures by fault
+#	injection.
+#
+#	NOTE: you need to run this script as root.
+#
+
+usage()
+{
+	cat >&2 <<EOF
+Usage: $0 [options] command [arguments]
+
+OPTIONS
+	-p percent
+	--probability=percent
+		likelihood of failure injection, in percent.
+		Default value is 1
+
+	-t value
+	--times=value
+		specifies how many times failures may happen at most.
+		Default value is 1
+
+	--oom-kill-allocating-task=value
+		set /proc/sys/vm/oom_kill_allocating_task to specified value
+		before running the command.
+		Default value is 1
+
+	-h, --help
+		Display a usage message and exit
+
+	--interval=value, --space=value, --verbose=value, --task-filter=value,
+	--stacktrace-depth=value, --require-start=value, --require-end=value,
+	--reject-start=value, --reject-end=value, --ignore-gfp-wait=value
+		See Documentation/fault-injection/fault-injection.rst for more
+		information
+
+	failslab options:
+	--cache-filter=value
+
+	fail_page_alloc options:
+	--ignore-gfp-highmem=value, --min-order=value
+
+ENVIRONMENT
+	FAILCMD_TYPE
+		The following values for FAILCMD_TYPE are recognized:
+
+		failslab
+			inject slab allocation failures
+		fail_page_alloc
+			inject page allocation failures
+
+		If FAILCMD_TYPE is not defined, then failslab is used.
+EOF
+}
+
+exit_if_not_hex() {
+    local value="$1"
+    if ! [[ $value =~ ^0x[0-9a-fA-F]+$ ]]; then
+        echo "Error: The provided value '$value' is not a valid hexadecimal number." >&2
+        exit 1
+    fi
+}
+
+if [ $UID != 0 ]; then
+	echo must be run as root >&2
+	exit 1
+fi
+
+DEBUGFS=`mount -t debugfs | head -1 | awk '{ print $3}'`
+
+if [ ! -d "$DEBUGFS" ]; then
+	echo debugfs is not mounted >&2
+	exit 1
+fi
+
+FAILCMD_TYPE=${FAILCMD_TYPE:-failslab}
+FAULTATTR=$DEBUGFS/$FAILCMD_TYPE
+
+if [ ! -d $FAULTATTR ]; then
+	echo $FAILCMD_TYPE is not available >&2
+	exit 1
+fi
+
+LONGOPTS=probability:,interval:,times:,space:,verbose:,task-filter:
+LONGOPTS=$LONGOPTS,stacktrace-depth:,require-start:,require-end:
+LONGOPTS=$LONGOPTS,reject-start:,reject-end:,oom-kill-allocating-task:,help
+
+if [ $FAILCMD_TYPE = failslab ]; then
+	LONGOPTS=$LONGOPTS,ignore-gfp-wait:,cache-filter:
+elif [ $FAILCMD_TYPE = fail_page_alloc ]; then
+	LONGOPTS=$LONGOPTS,ignore-gfp-wait:,ignore-gfp-highmem:,min-order:
+fi
+
+TEMP=`getopt -o p:i:t:s:v:h --long $LONGOPTS -n 'failcmd.sh' -- "$@"`
+
+if [ $? != 0 ]; then
+	usage
+	exit 1
+fi
+
+eval set -- "$TEMP"
+
+fault_attr_default()
+{
+	echo N > $FAULTATTR/task-filter
+	echo 0 > $FAULTATTR/probability
+	echo 1 > $FAULTATTR/times
+}
+
+fault_attr_default
+
+oom_kill_allocating_task_saved=`cat /proc/sys/vm/oom_kill_allocating_task`
+
+restore_values()
+{
+	fault_attr_default
+	echo $oom_kill_allocating_task_saved \
+		> /proc/sys/vm/oom_kill_allocating_task
+}
+
+#
+# Default options
+#
+declare -i oom_kill_allocating_task=1
+declare task_filter=Y
+declare -i probability=1
+declare -i times=1
+
+while true; do
+	case "$1" in
+	-p|--probability)
+		probability=$2
+		shift 2
+		;;
+	-i|--interval)
+		echo $2 > $FAULTATTR/interval
+		shift 2
+		;;
+	-t|--times)
+		times=$2
+		shift 2
+		;;
+	-s|--space)
+		echo $2 > $FAULTATTR/space
+		shift 2
+		;;
+	-v|--verbose)
+		echo $2 > $FAULTATTR/verbose
+		shift 2
+		;;
+	--task-filter)
+		task_filter=$2
+		shift 2
+		;;
+	--stacktrace-depth)
+		echo $2 > $FAULTATTR/stacktrace-depth
+		shift 2
+		;;
+	--require-start)
+		exit_if_not_hex "$2"
+		echo $2 > $FAULTATTR/require-start
+		shift 2
+		;;
+	--require-end)
+		exit_if_not_hex "$2"
+		echo $2 > $FAULTATTR/require-end
+		shift 2
+		;;
+	--reject-start)
+		exit_if_not_hex "$2"
+		echo $2 > $FAULTATTR/reject-start
+		shift 2
+		;;
+	--reject-end)
+		exit_if_not_hex "$2"
+		echo $2 > $FAULTATTR/reject-end
+		shift 2
+		;;
+	--oom-kill-allocating-task)
+		oom_kill_allocating_task=$2
+		shift 2
+		;;
+	--ignore-gfp-wait)
+		echo $2 > $FAULTATTR/ignore-gfp-wait
+		shift 2
+		;;
+	--cache-filter)
+		echo $2 > $FAULTATTR/cache_filter
+		shift 2
+		;;
+	--ignore-gfp-highmem)
+		echo $2 > $FAULTATTR/ignore-gfp-highmem
+		shift 2
+		;;
+	--min-order)
+		echo $2 > $FAULTATTR/min-order
+		shift 2
+		;;
+	-h|--help)
+		usage
+		exit 0
+		shift
+		;;
+	--)
+		shift
+		break
+		;;
+	esac
+done
+
+[ -z "$1" ] && exit 0
+
+echo $oom_kill_allocating_task > /proc/sys/vm/oom_kill_allocating_task
+echo $task_filter > $FAULTATTR/task-filter
+echo $probability > $FAULTATTR/probability
+echo $times > $FAULTATTR/times
+
+trap "restore_values" SIGINT SIGTERM EXIT
+
+cmd="echo 1 > /proc/self/make-it-fail && exec $@"
+bash -c "$cmd"