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, 307 insertions, 0 deletions

diff --git a/drivers/mtd/maps/sa1100-flash.c b/drivers/mtd/maps/sa1100-flash.c
new file mode 100644
index 000000000..4506b27a9
--- /dev/null
+++ b/drivers/mtd/maps/sa1100-flash.c
@@ -0,0 +1,307 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Flash memory access on SA11x0 based devices
+ *
+ * (C) 2000 Nicolas Pitre <nico@fluxnic.net>
+ */
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/ioport.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/slab.h>
+#include <linux/platform_device.h>
+#include <linux/err.h>
+#include <linux/io.h>
+
+#include <linux/mtd/mtd.h>
+#include <linux/mtd/map.h>
+#include <linux/mtd/partitions.h>
+#include <linux/mtd/concat.h>
+
+#include <mach/hardware.h>
+#include <linux/sizes.h>
+#include <asm/mach/flash.h>
+
+struct sa_subdev_info {
+	char name[16];
+	struct map_info map;
+	struct mtd_info *mtd;
+	struct flash_platform_data *plat;
+};
+
+struct sa_info {
+	struct mtd_info		*mtd;
+	int			num_subdev;
+	struct sa_subdev_info	subdev[];
+};
+
+static DEFINE_SPINLOCK(sa1100_vpp_lock);
+static int sa1100_vpp_refcnt;
+static void sa1100_set_vpp(struct map_info *map, int on)
+{
+	struct sa_subdev_info *subdev = container_of(map, struct sa_subdev_info, map);
+	unsigned long flags;
+
+	spin_lock_irqsave(&sa1100_vpp_lock, flags);
+	if (on) {
+		if (++sa1100_vpp_refcnt == 1)   /* first nested 'on' */
+			subdev->plat->set_vpp(1);
+	} else {
+		if (--sa1100_vpp_refcnt == 0)   /* last nested 'off' */
+			subdev->plat->set_vpp(0);
+	}
+	spin_unlock_irqrestore(&sa1100_vpp_lock, flags);
+}
+
+static void sa1100_destroy_subdev(struct sa_subdev_info *subdev)
+{
+	if (subdev->mtd)
+		map_destroy(subdev->mtd);
+	if (subdev->map.virt)
+		iounmap(subdev->map.virt);
+	release_mem_region(subdev->map.phys, subdev->map.size);
+}
+
+static int sa1100_probe_subdev(struct sa_subdev_info *subdev, struct resource *res)
+{
+	unsigned long phys;
+	unsigned int size;
+	int ret;
+
+	phys = res->start;
+	size = res->end - phys + 1;
+
+	/*
+	 * Retrieve the bankwidth from the MSC registers.
+	 * We currently only implement CS0 and CS1 here.
+	 */
+	switch (phys) {
+	default:
+		printk(KERN_WARNING "SA1100 flash: unknown base address "
+		       "0x%08lx, assuming CS0\n", phys);
+		fallthrough;
+	case SA1100_CS0_PHYS:
+		subdev->map.bankwidth = (MSC0 & MSC_RBW) ? 2 : 4;
+		break;
+
+	case SA1100_CS1_PHYS:
+		subdev->map.bankwidth = ((MSC0 >> 16) & MSC_RBW) ? 2 : 4;
+		break;
+	}
+
+	if (!request_mem_region(phys, size, subdev->name)) {
+		ret = -EBUSY;
+		goto out;
+	}
+
+	if (subdev->plat->set_vpp)
+		subdev->map.set_vpp = sa1100_set_vpp;
+
+	subdev->map.phys = phys;
+	subdev->map.size = size;
+	subdev->map.virt = ioremap(phys, size);
+	if (!subdev->map.virt) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	simple_map_init(&subdev->map);
+
+	/*
+	 * Now let's probe for the actual flash.  Do it here since
+	 * specific machine settings might have been set above.
+	 */
+	subdev->mtd = do_map_probe(subdev->plat->map_name, &subdev->map);
+	if (subdev->mtd == NULL) {
+		ret = -ENXIO;
+		goto err;
+	}
+
+	printk(KERN_INFO "SA1100 flash: CFI device at 0x%08lx, %uMiB, %d-bit\n",
+		phys, (unsigned)(subdev->mtd->size >> 20),
+		subdev->map.bankwidth * 8);
+
+	return 0;
+
+ err:
+	sa1100_destroy_subdev(subdev);
+ out:
+	return ret;
+}
+
+static void sa1100_destroy(struct sa_info *info, struct flash_platform_data *plat)
+{
+	int i;
+
+	if (info->mtd) {
+		mtd_device_unregister(info->mtd);
+		if (info->mtd != info->subdev[0].mtd)
+			mtd_concat_destroy(info->mtd);
+	}
+
+	for (i = info->num_subdev - 1; i >= 0; i--)
+		sa1100_destroy_subdev(&info->subdev[i]);
+	kfree(info);
+
+	if (plat->exit)
+		plat->exit();
+}
+
+static struct sa_info *sa1100_setup_mtd(struct platform_device *pdev,
+					struct flash_platform_data *plat)
+{
+	struct sa_info *info;
+	int nr, i, ret = 0;
+
+	/*
+	 * Count number of devices.
+	 */
+	for (nr = 0; ; nr++)
+		if (!platform_get_resource(pdev, IORESOURCE_MEM, nr))
+			break;
+
+	if (nr == 0) {
+		ret = -ENODEV;
+		goto out;
+	}
+
+	/*
+	 * Allocate the map_info structs in one go.
+	 */
+	info = kzalloc_flex(*info, subdev, nr);
+	if (!info) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	if (plat->init) {
+		ret = plat->init();
+		if (ret)
+			goto err;
+	}
+
+	/*
+	 * Claim and then map the memory regions.
+	 */
+	for (i = 0; i < nr; i++) {
+		struct sa_subdev_info *subdev = &info->subdev[i];
+		struct resource *res;
+
+		res = platform_get_resource(pdev, IORESOURCE_MEM, i);
+		if (!res)
+			break;
+
+		subdev->map.name = subdev->name;
+		sprintf(subdev->name, "%s-%d", plat->name, i);
+		subdev->plat = plat;
+
+		ret = sa1100_probe_subdev(subdev, res);
+		if (ret)
+			break;
+	}
+
+	info->num_subdev = i;
+
+	/*
+	 * ENXIO is special.  It means we didn't find a chip when we probed.
+	 */
+	if (ret != 0 && !(ret == -ENXIO && info->num_subdev > 0))
+		goto err;
+
+	/*
+	 * If we found one device, don't bother with concat support.  If
+	 * we found multiple devices, use concat if we have it available,
+	 * otherwise fail.  Either way, it'll be called "sa1100".
+	 */
+	if (info->num_subdev == 1) {
+		strcpy(info->subdev[0].name, plat->name);
+		info->mtd = info->subdev[0].mtd;
+		ret = 0;
+	} else if (info->num_subdev > 1) {
+		struct mtd_info **cdev;
+
+		cdev = kmalloc_objs(*cdev, nr);
+		if (!cdev) {
+			ret = -ENOMEM;
+			goto err;
+		}
+
+		/*
+		 * We detected multiple devices.  Concatenate them together.
+		 */
+		for (i = 0; i < info->num_subdev; i++)
+			cdev[i] = info->subdev[i].mtd;
+
+		info->mtd = mtd_concat_create(cdev, info->num_subdev,
+					      plat->name);
+		kfree(cdev);
+		if (info->mtd == NULL) {
+			ret = -ENXIO;
+			goto err;
+		}
+	}
+	info->mtd->dev.parent = &pdev->dev;
+
+	if (ret == 0)
+		return info;
+
+ err:
+	sa1100_destroy(info, plat);
+ out:
+	return ERR_PTR(ret);
+}
+
+static const char * const part_probes[] = { "cmdlinepart", "RedBoot", NULL };
+
+static int sa1100_mtd_probe(struct platform_device *pdev)
+{
+	struct flash_platform_data *plat = dev_get_platdata(&pdev->dev);
+	struct sa_info *info;
+	int err;
+
+	if (!plat)
+		return -ENODEV;
+
+	info = sa1100_setup_mtd(pdev, plat);
+	if (IS_ERR(info)) {
+		err = PTR_ERR(info);
+		goto out;
+	}
+
+	/*
+	 * Partition selection stuff.
+	 */
+	mtd_device_parse_register(info->mtd, part_probes, NULL, plat->parts,
+				  plat->nr_parts);
+
+	platform_set_drvdata(pdev, info);
+	err = 0;
+
+ out:
+	return err;
+}
+
+static void sa1100_mtd_remove(struct platform_device *pdev)
+{
+	struct sa_info *info = platform_get_drvdata(pdev);
+	struct flash_platform_data *plat = dev_get_platdata(&pdev->dev);
+
+	sa1100_destroy(info, plat);
+}
+
+static struct platform_driver sa1100_mtd_driver = {
+	.probe		= sa1100_mtd_probe,
+	.remove		= sa1100_mtd_remove,
+	.driver		= {
+		.name	= "sa1100-mtd",
+	},
+};
+
+module_platform_driver(sa1100_mtd_driver);
+
+MODULE_AUTHOR("Nicolas Pitre");
+MODULE_DESCRIPTION("SA1100 CFI map driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:sa1100-mtd");