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, 135 insertions, 0 deletions

diff --git a/drivers/video/cmdline.c b/drivers/video/cmdline.c
new file mode 100644
index 000000000..49ee3fb49
--- /dev/null
+++ b/drivers/video/cmdline.c
@@ -0,0 +1,135 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Based on the fbdev code in drivers/video/fbdev/core/fb_cmdline:
+ *
+ *  Copyright (C) 2014 Intel Corp
+ *  Copyright (C) 1994 Martin Schaller
+ *
+ *	2001 - Documented with DocBook
+ *	- Brad Douglas <brad@neruo.com>
+ *
+ * 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.
+ *
+ * Authors:
+ *    Daniel Vetter <daniel.vetter@ffwll.ch>
+ */
+
+#include <linux/fb.h> /* for FB_MAX */
+#include <linux/init.h>
+
+#include <video/cmdline.h>
+
+/*
+ * FB_MAX is the maximum number of framebuffer devices and also
+ * the maximum number of video= parameters. Although not directly
+ * related to each other, it makes sense to keep it that way.
+ */
+static const char *video_options[FB_MAX] __read_mostly;
+static const char *video_option __read_mostly;
+static int video_of_only __read_mostly;
+
+static const char *__video_get_option_string(const char *name)
+{
+	const char *options = NULL;
+	size_t name_len = 0;
+
+	if (name)
+		name_len = strlen(name);
+
+	if (name_len) {
+		unsigned int i;
+		const char *opt;
+
+		for (i = 0; i < ARRAY_SIZE(video_options); ++i) {
+			if (!video_options[i])
+				continue;
+			if (video_options[i][0] == '\0')
+				continue;
+			opt = video_options[i];
+			if (!strncmp(opt, name, name_len) && opt[name_len] == ':')
+				options = opt + name_len + 1;
+		}
+	}
+
+	/* No match, return global options */
+	if (!options)
+		options = video_option;
+
+	return options;
+}
+
+/**
+ * video_get_options - get kernel boot parameters
+ * @name:	name of the output as it would appear in the boot parameter
+ *		line (video=<name>:<options>)
+ *
+ * Looks up the video= options for the given name. Names are connector
+ * names with DRM, or driver names with fbdev. If no video option for
+ * the name has been specified, the function returns the global video=
+ * setting. A @name of NULL always returns the global video setting.
+ *
+ * Returns:
+ * The string of video options for the given name, or NULL if no video
+ * option has been specified.
+ */
+const char *video_get_options(const char *name)
+{
+	return __video_get_option_string(name);
+}
+EXPORT_SYMBOL(video_get_options);
+
+#if IS_ENABLED(CONFIG_FB_CORE)
+bool __video_get_options(const char *name, const char **options, bool is_of)
+{
+	bool enabled = true;
+	const char *opt = NULL;
+
+	if (video_of_only && !is_of)
+		enabled = false;
+
+	opt = __video_get_option_string(name);
+
+	if (options)
+		*options = opt;
+
+	return enabled;
+}
+EXPORT_SYMBOL(__video_get_options);
+#endif
+
+/*
+ * Process command line options for video adapters. This function is
+ * a __setup and __init function. It only stores the options. Drivers
+ * have to call video_get_options() as necessary.
+ */
+static int __init video_setup(char *options)
+{
+	if (!options || !*options)
+		goto out;
+
+	if (!strncmp(options, "ofonly", 6)) {
+		video_of_only = true;
+		goto out;
+	}
+
+	if (strchr(options, ':')) {
+		/* named */
+		size_t i;
+
+		for (i = 0; i < ARRAY_SIZE(video_options); i++) {
+			if (!video_options[i]) {
+				video_options[i] = options;
+				break;
+			}
+		}
+	} else {
+		/* global */
+		video_option = options;
+	}
+
+out:
+	return 1;
+}
+__setup("video=", video_setup);