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, 175 insertions, 0 deletions

diff --git a/drivers/video/fbdev/maxinefb.c b/drivers/video/fbdev/maxinefb.c
new file mode 100644
index 000000000..67d4bd881
--- /dev/null
+++ b/drivers/video/fbdev/maxinefb.c
@@ -0,0 +1,175 @@
+/*
+ *      linux/drivers/video/maxinefb.c
+ *
+ *	DECstation 5000/xx onboard framebuffer support ... derived from:
+ *	"HP300 Topcat framebuffer support (derived from macfb of all things)
+ *	Phil Blundell <philb@gnu.org> 1998", the original code can be
+ *      found in the file hpfb.c in the same directory.
+ *
+ *      DECstation related code Copyright (C) 1999,2000,2001 by
+ *      Michael Engel <engel@unix-ag.org> and
+ *      Karsten Merker <merker@linuxtag.org>.
+ *      This file is subject to the terms and conditions of the GNU General
+ *      Public License.  See the file COPYING in the main directory of this
+ *      archive for more details.
+ *
+ */
+
+/*
+ * Changes:
+ * 2001/01/27 removed debugging and testing code, fixed fb_ops
+ *            initialization which had caused a crash before,
+ *            general cleanup, first official release (KM)
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/mm.h>
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/fb.h>
+#include <video/maxinefb.h>
+
+/* bootinfo.h defines the machine type values, needed when checking */
+/* whether are really running on a maxine, KM                       */
+#include <asm/bootinfo.h>
+
+static struct fb_info fb_info;
+
+static const struct fb_var_screeninfo maxinefb_defined = {
+	.xres =		1024,
+	.yres =		768,
+	.xres_virtual =	1024,
+	.yres_virtual =	768,
+	.bits_per_pixel =8,
+	.activate =	FB_ACTIVATE_NOW,
+	.height =	-1,
+	.width =	-1,
+	.vmode =	FB_VMODE_NONINTERLACED,
+};
+
+static struct fb_fix_screeninfo maxinefb_fix __initdata = {
+	.id =		"Maxine",
+	.smem_len =	(1024*768),
+	.type =		FB_TYPE_PACKED_PIXELS,
+	.visual =	FB_VISUAL_PSEUDOCOLOR,
+	.line_length =	1024,
+};
+
+/* Handle the funny Inmos RamDAC/video controller ... */
+
+static void maxinefb_ims332_write_register(int regno, register unsigned int val)
+{
+	register unsigned char *regs = (char *) MAXINEFB_IMS332_ADDRESS;
+	unsigned char *wptr;
+
+	wptr = regs + 0xa0000 + (regno << 4);
+	*((volatile unsigned int *) (regs)) = (val >> 8) & 0xff00;
+	*((volatile unsigned short *) (wptr)) = val;
+}
+
+#if 0
+/* dead code: leave here for hardware interface documentation */
+static unsigned int maxinefb_ims332_read_register(int regno)
+{
+	register unsigned char *regs = (char *) MAXINEFB_IMS332_ADDRESS;
+	unsigned char *rptr;
+	register unsigned int j, k;
+
+	rptr = regs + 0x80000 + (regno << 4);
+	j = *((volatile unsigned short *) rptr);
+	k = *((volatile unsigned short *) regs);
+
+	return (j & 0xffff) | ((k & 0xff00) << 8);
+}
+#endif
+
+/* Set the palette */
+static int maxinefb_setcolreg(unsigned regno, unsigned red, unsigned green,
+			      unsigned blue, unsigned transp, struct fb_info *info)
+{
+	/* value to be written into the palette reg. */
+	unsigned long hw_colorvalue = 0;
+
+	if (regno > 255)
+		return 1;
+
+	red   >>= 8;    /* The cmap fields are 16 bits    */
+	green >>= 8;    /* wide, but the harware colormap */
+	blue  >>= 8;    /* registers are only 8 bits wide */
+
+	hw_colorvalue = (blue << 16) + (green << 8) + (red);
+
+	maxinefb_ims332_write_register(IMS332_REG_COLOR_PALETTE + regno,
+				       hw_colorvalue);
+	return 0;
+}
+
+static const struct fb_ops maxinefb_ops = {
+	.owner		= THIS_MODULE,
+	FB_DEFAULT_IOMEM_OPS,
+	.fb_setcolreg	= maxinefb_setcolreg,
+};
+
+static int __init maxinefb_init(void)
+{
+	unsigned long fboff;
+	unsigned long fb_start;
+	int i;
+
+	if (fb_get_options("maxinefb", NULL))
+		return -ENODEV;
+
+	/* Validate we're on the proper machine type */
+	if (mips_machtype != MACH_DS5000_XX) {
+		return -EINVAL;
+	}
+
+	printk(KERN_INFO "Maxinefb: Personal DECstation detected\n");
+	printk(KERN_INFO "Maxinefb: initializing onboard framebuffer\n");
+
+	/* Framebuffer display memory base address */
+	fb_start = DS5000_xx_ONBOARD_FBMEM_START;
+
+	/* Clear screen */
+	for (fboff = fb_start; fboff < fb_start + 0x1ffff; fboff++)
+		*(volatile unsigned char *)fboff = 0x0;
+
+	maxinefb_fix.smem_start = fb_start;
+
+	/* erase hardware cursor */
+	for (i = 0; i < 512; i++) {
+		maxinefb_ims332_write_register(IMS332_REG_CURSOR_RAM + i,
+					       0);
+		/*
+		   if (i&0x8 == 0)
+		   maxinefb_ims332_write_register (IMS332_REG_CURSOR_RAM + i, 0x0f);
+		   else
+		   maxinefb_ims332_write_register (IMS332_REG_CURSOR_RAM + i, 0xf0);
+		 */
+	}
+
+	fb_info.fbops = &maxinefb_ops;
+	fb_info.screen_base = (char *)maxinefb_fix.smem_start;
+	fb_info.var = maxinefb_defined;
+	fb_info.fix = maxinefb_fix;
+
+	fb_alloc_cmap(&fb_info.cmap, 256, 0);
+
+	if (register_framebuffer(&fb_info) < 0)
+		return -ENODEV;
+	return 0;
+}
+
+static void __exit maxinefb_exit(void)
+{
+	unregister_framebuffer(&fb_info);
+}
+
+module_init(maxinefb_init);
+module_exit(maxinefb_exit);
+
+MODULE_LICENSE("GPL");