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, 153 insertions, 0 deletions

diff --git a/drivers/clocksource/timer-goldfish.c b/drivers/clocksource/timer-goldfish.c
new file mode 100644
index 000000000..3fe12c7ee
--- /dev/null
+++ b/drivers/clocksource/timer-goldfish.c
@@ -0,0 +1,153 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/interrupt.h>
+#include <linux/ioport.h>
+#include <linux/clocksource.h>
+#include <linux/clockchips.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/goldfish.h>
+#include <clocksource/timer-goldfish.h>
+
+struct goldfish_timer {
+	struct clocksource cs;
+	struct clock_event_device ced;
+	struct resource res;
+	void __iomem *base;
+};
+
+static struct goldfish_timer *ced_to_gf(struct clock_event_device *ced)
+{
+	return container_of(ced, struct goldfish_timer, ced);
+}
+
+static struct goldfish_timer *cs_to_gf(struct clocksource *cs)
+{
+	return container_of(cs, struct goldfish_timer, cs);
+}
+
+static u64 goldfish_timer_read(struct clocksource *cs)
+{
+	struct goldfish_timer *timerdrv = cs_to_gf(cs);
+	void __iomem *base = timerdrv->base;
+	u32 time_low, time_high;
+	u64 ticks;
+
+	/*
+	 * time_low: get low bits of current time and update time_high
+	 * time_high: get high bits of time at last time_low read
+	 */
+	time_low = gf_ioread32(base + TIMER_TIME_LOW);
+	time_high = gf_ioread32(base + TIMER_TIME_HIGH);
+
+	ticks = ((u64)time_high << 32) | time_low;
+
+	return ticks;
+}
+
+static int goldfish_timer_set_oneshot(struct clock_event_device *evt)
+{
+	struct goldfish_timer *timerdrv = ced_to_gf(evt);
+	void __iomem *base = timerdrv->base;
+
+	gf_iowrite32(0, base + TIMER_ALARM_HIGH);
+	gf_iowrite32(0, base + TIMER_ALARM_LOW);
+	gf_iowrite32(1, base + TIMER_IRQ_ENABLED);
+
+	return 0;
+}
+
+static int goldfish_timer_shutdown(struct clock_event_device *evt)
+{
+	struct goldfish_timer *timerdrv = ced_to_gf(evt);
+	void __iomem *base = timerdrv->base;
+
+	gf_iowrite32(0, base + TIMER_IRQ_ENABLED);
+
+	return 0;
+}
+
+static int goldfish_timer_next_event(unsigned long delta,
+				     struct clock_event_device *evt)
+{
+	struct goldfish_timer *timerdrv = ced_to_gf(evt);
+	void __iomem *base = timerdrv->base;
+	u64 now;
+
+	now = goldfish_timer_read(&timerdrv->cs);
+
+	now += delta;
+
+	gf_iowrite32(upper_32_bits(now), base + TIMER_ALARM_HIGH);
+	gf_iowrite32(lower_32_bits(now), base + TIMER_ALARM_LOW);
+
+	return 0;
+}
+
+static irqreturn_t goldfish_timer_irq(int irq, void *dev_id)
+{
+	struct goldfish_timer *timerdrv = dev_id;
+	struct clock_event_device *evt = &timerdrv->ced;
+	void __iomem *base = timerdrv->base;
+
+	gf_iowrite32(1, base + TIMER_CLEAR_INTERRUPT);
+
+	evt->event_handler(evt);
+
+	return IRQ_HANDLED;
+}
+
+int __init goldfish_timer_init(int irq, void __iomem *base)
+{
+	struct goldfish_timer *timerdrv;
+	int ret;
+
+	timerdrv = kzalloc_obj(*timerdrv);
+	if (!timerdrv)
+		return -ENOMEM;
+
+	timerdrv->base = base;
+
+	timerdrv->ced = (struct clock_event_device){
+		.name			= "goldfish_timer",
+		.features		= CLOCK_EVT_FEAT_ONESHOT,
+		.set_state_shutdown	= goldfish_timer_shutdown,
+		.set_state_oneshot      = goldfish_timer_set_oneshot,
+		.set_next_event		= goldfish_timer_next_event,
+	};
+
+	timerdrv->res = (struct resource){
+		.name  = "goldfish_timer",
+		.start = (unsigned long)base,
+		.end   = (unsigned long)base + 0xfff,
+	};
+
+	ret = request_resource(&iomem_resource, &timerdrv->res);
+	if (ret) {
+		pr_err("Cannot allocate '%s' resource\n", timerdrv->res.name);
+		return ret;
+	}
+
+	timerdrv->cs = (struct clocksource){
+		.name		= "goldfish_timer",
+		.rating		= 400,
+		.read		= goldfish_timer_read,
+		.mask		= CLOCKSOURCE_MASK(64),
+		.flags		= 0,
+		.max_idle_ns	= LONG_MAX,
+	};
+
+	clocksource_register_hz(&timerdrv->cs, NSEC_PER_SEC);
+
+	ret = request_irq(irq, goldfish_timer_irq, IRQF_TIMER,
+			  "goldfish_timer", timerdrv);
+	if (ret) {
+		pr_err("Couldn't register goldfish-timer interrupt\n");
+		return ret;
+	}
+
+	clockevents_config_and_register(&timerdrv->ced, NSEC_PER_SEC,
+					1, 0xffffffff);
+
+	return 0;
+}