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, 191 insertions, 0 deletions

diff --git a/samples/rust/rust_configfs.rs b/samples/rust/rust_configfs.rs
new file mode 100644
index 000000000..a1bd9db60
--- /dev/null
+++ b/samples/rust/rust_configfs.rs
@@ -0,0 +1,191 @@
+// SPDX-License-Identifier: GPL-2.0
+
+//! Rust configfs sample.
+
+use kernel::alloc::flags;
+use kernel::configfs;
+use kernel::configfs::configfs_attrs;
+use kernel::new_mutex;
+use kernel::page::PAGE_SIZE;
+use kernel::prelude::*;
+use kernel::sync::Mutex;
+
+module! {
+    type: RustConfigfs,
+    name: "rust_configfs",
+    authors: ["Rust for Linux Contributors"],
+    description: "Rust configfs sample",
+    license: "GPL",
+}
+
+#[pin_data]
+struct RustConfigfs {
+    #[pin]
+    config: configfs::Subsystem<Configuration>,
+}
+
+#[pin_data]
+struct Configuration {
+    message: &'static CStr,
+    #[pin]
+    bar: Mutex<(KBox<[u8; PAGE_SIZE]>, usize)>,
+}
+
+impl Configuration {
+    fn new() -> impl PinInit<Self, Error> {
+        try_pin_init!(Self {
+            message: c"Hello World\n",
+            bar <- new_mutex!((KBox::new([0; PAGE_SIZE], flags::GFP_KERNEL)?, 0)),
+        })
+    }
+}
+
+impl kernel::InPlaceModule for RustConfigfs {
+    fn init(_module: &'static ThisModule) -> impl PinInit<Self, Error> {
+        pr_info!("Rust configfs sample (init)\n");
+
+        // Define a subsystem with the data type `Configuration`, two
+        // attributes, `message` and `bar` and child group type `Child`. `mkdir`
+        // in the directory representing this subsystem will create directories
+        // backed by the `Child` type.
+        let item_type = configfs_attrs! {
+            container: configfs::Subsystem<Configuration>,
+            data: Configuration,
+            child: Child,
+            attributes: [
+                message: 0,
+                bar: 1,
+            ],
+        };
+
+        try_pin_init!(Self {
+            config <- configfs::Subsystem::new(
+                c"rust_configfs", item_type, Configuration::new()
+            ),
+        })
+    }
+}
+
+#[vtable]
+impl configfs::GroupOperations for Configuration {
+    type Child = Child;
+
+    fn make_group(&self, name: &CStr) -> Result<impl PinInit<configfs::Group<Child>, Error>> {
+        // Define a group with data type `Child`, one attribute `baz` and child
+        // group type `GrandChild`. `mkdir` in the directory representing this
+        // group will create directories backed by the `GrandChild` type.
+        let tpe = configfs_attrs! {
+            container: configfs::Group<Child>,
+            data: Child,
+            child: GrandChild,
+            attributes: [
+                baz: 0,
+            ],
+        };
+
+        Ok(configfs::Group::new(name.try_into()?, tpe, Child::new()))
+    }
+}
+
+#[vtable]
+impl configfs::AttributeOperations<0> for Configuration {
+    type Data = Configuration;
+
+    fn show(container: &Configuration, page: &mut [u8; PAGE_SIZE]) -> Result<usize> {
+        pr_info!("Show message\n");
+        let data = container.message.to_bytes();
+        page[0..data.len()].copy_from_slice(data);
+        Ok(data.len())
+    }
+}
+
+#[vtable]
+impl configfs::AttributeOperations<1> for Configuration {
+    type Data = Configuration;
+
+    fn show(container: &Configuration, page: &mut [u8; PAGE_SIZE]) -> Result<usize> {
+        pr_info!("Show bar\n");
+        let guard = container.bar.lock();
+        let data = guard.0.as_slice();
+        let len = guard.1;
+        page[0..len].copy_from_slice(&data[0..len]);
+        Ok(len)
+    }
+
+    fn store(container: &Configuration, page: &[u8]) -> Result {
+        pr_info!("Store bar\n");
+        let mut guard = container.bar.lock();
+        guard.0[0..page.len()].copy_from_slice(page);
+        guard.1 = page.len();
+        Ok(())
+    }
+}
+
+// `pin_data` cannot handle structs without braces.
+#[pin_data]
+struct Child {}
+
+impl Child {
+    fn new() -> impl PinInit<Self, Error> {
+        try_pin_init!(Self {})
+    }
+}
+
+#[vtable]
+impl configfs::GroupOperations for Child {
+    type Child = GrandChild;
+
+    fn make_group(&self, name: &CStr) -> Result<impl PinInit<configfs::Group<GrandChild>, Error>> {
+        // Define a group with data type `GrandChild`, one attribute `gc`. As no
+        // child type is specified, it will not be possible to create subgroups
+        // in this group, and `mkdir`in the directory representing this group
+        // will return an error.
+        let tpe = configfs_attrs! {
+            container: configfs::Group<GrandChild>,
+            data: GrandChild,
+            attributes: [
+                gc: 0,
+            ],
+        };
+
+        Ok(configfs::Group::new(
+            name.try_into()?,
+            tpe,
+            GrandChild::new(),
+        ))
+    }
+}
+
+#[vtable]
+impl configfs::AttributeOperations<0> for Child {
+    type Data = Child;
+
+    fn show(_container: &Child, page: &mut [u8; PAGE_SIZE]) -> Result<usize> {
+        pr_info!("Show baz\n");
+        let data = c"Hello Baz\n".to_bytes();
+        page[0..data.len()].copy_from_slice(data);
+        Ok(data.len())
+    }
+}
+
+// `pin_data` cannot handle structs without braces.
+#[pin_data]
+struct GrandChild {}
+
+impl GrandChild {
+    fn new() -> impl PinInit<Self, Error> {
+        try_pin_init!(Self {})
+    }
+}
+
+#[vtable]
+impl configfs::AttributeOperations<0> for GrandChild {
+    type Data = GrandChild;
+
+    fn show(_container: &GrandChild, page: &mut [u8; PAGE_SIZE]) -> Result<usize> {
+        pr_info!("Show grand child\n");
+        let data = c"Hello GC\n".to_bytes();
+        page[0..data.len()].copy_from_slice(data);
+        Ok(data.len())
+    }
+}