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, 153 insertions, 0 deletions

diff --git a/rust/proc-macro2/extra.rs b/rust/proc-macro2/extra.rs
new file mode 100644
index 000000000..55feb5ec7
--- /dev/null
+++ b/rust/proc-macro2/extra.rs
@@ -0,0 +1,153 @@
+// SPDX-License-Identifier: Apache-2.0 OR MIT
+
+//! Items which do not have a correspondence to any API in the proc_macro crate,
+//! but are necessary to include in proc-macro2.
+
+use crate::fallback;
+use crate::imp;
+use crate::marker::{ProcMacroAutoTraits, MARKER};
+use crate::Span;
+use core::fmt::{self, Debug};
+
+/// Invalidate any `proc_macro2::Span` that exist on the current thread.
+///
+/// The implementation of `Span` uses thread-local data structures and this
+/// function clears them. Calling any method on a `Span` on the current thread
+/// created prior to the invalidation will return incorrect values or crash.
+///
+/// This function is useful for programs that process more than 2<sup>32</sup>
+/// bytes of Rust source code on the same thread. Just like rustc, proc-macro2
+/// uses 32-bit source locations, and these wrap around when the total source
+/// code processed by the same thread exceeds 2<sup>32</sup> bytes (4
+/// gigabytes). After a wraparound, `Span` methods such as `source_text()` can
+/// return wrong data.
+///
+/// # Example
+///
+/// As of late 2023, there is 200 GB of Rust code published on crates.io.
+/// Looking at just the newest version of every crate, it is 16 GB of code. So a
+/// workload that involves parsing it all would overflow a 32-bit source
+/// location unless spans are being invalidated.
+///
+/// ```
+/// use flate2::read::GzDecoder;
+/// use std::ffi::OsStr;
+/// use std::io::{BufReader, Read};
+/// use std::str::FromStr;
+/// use tar::Archive;
+///
+/// rayon::scope(|s| {
+///     for krate in every_version_of_every_crate() {
+///         s.spawn(move |_| {
+///             proc_macro2::extra::invalidate_current_thread_spans();
+///
+///             let reader = BufReader::new(krate);
+///             let tar = GzDecoder::new(reader);
+///             let mut archive = Archive::new(tar);
+///             for entry in archive.entries().unwrap() {
+///                 let mut entry = entry.unwrap();
+///                 let path = entry.path().unwrap();
+///                 if path.extension() != Some(OsStr::new("rs")) {
+///                     continue;
+///                 }
+///                 let mut content = String::new();
+///                 entry.read_to_string(&mut content).unwrap();
+///                 match proc_macro2::TokenStream::from_str(&content) {
+///                     Ok(tokens) => {/* ... */},
+///                     Err(_) => continue,
+///                 }
+///             }
+///         });
+///     }
+/// });
+/// #
+/// # fn every_version_of_every_crate() -> Vec<std::fs::File> {
+/// #     Vec::new()
+/// # }
+/// ```
+///
+/// # Panics
+///
+/// This function is not applicable to and will panic if called from a
+/// procedural macro.
+#[cfg(span_locations)]
+#[cfg_attr(docsrs, doc(cfg(feature = "span-locations")))]
+pub fn invalidate_current_thread_spans() {
+    crate::imp::invalidate_current_thread_spans();
+}
+
+/// An object that holds a [`Group`]'s `span_open()` and `span_close()` together
+/// in a more compact representation than holding those 2 spans individually.
+///
+/// [`Group`]: crate::Group
+#[derive(Copy, Clone)]
+pub struct DelimSpan {
+    inner: DelimSpanEnum,
+    _marker: ProcMacroAutoTraits,
+}
+
+#[derive(Copy, Clone)]
+enum DelimSpanEnum {
+    #[cfg(wrap_proc_macro)]
+    Compiler {
+        join: proc_macro::Span,
+        open: proc_macro::Span,
+        close: proc_macro::Span,
+    },
+    Fallback(fallback::Span),
+}
+
+impl DelimSpan {
+    pub(crate) fn new(group: &imp::Group) -> Self {
+        #[cfg(wrap_proc_macro)]
+        let inner = match group {
+            imp::Group::Compiler(group) => DelimSpanEnum::Compiler {
+                join: group.span(),
+                open: group.span_open(),
+                close: group.span_close(),
+            },
+            imp::Group::Fallback(group) => DelimSpanEnum::Fallback(group.span()),
+        };
+
+        #[cfg(not(wrap_proc_macro))]
+        let inner = DelimSpanEnum::Fallback(group.span());
+
+        DelimSpan {
+            inner,
+            _marker: MARKER,
+        }
+    }
+
+    /// Returns a span covering the entire delimited group.
+    pub fn join(&self) -> Span {
+        match &self.inner {
+            #[cfg(wrap_proc_macro)]
+            DelimSpanEnum::Compiler { join, .. } => Span::_new(imp::Span::Compiler(*join)),
+            DelimSpanEnum::Fallback(span) => Span::_new_fallback(*span),
+        }
+    }
+
+    /// Returns a span for the opening punctuation of the group only.
+    pub fn open(&self) -> Span {
+        match &self.inner {
+            #[cfg(wrap_proc_macro)]
+            DelimSpanEnum::Compiler { open, .. } => Span::_new(imp::Span::Compiler(*open)),
+            DelimSpanEnum::Fallback(span) => Span::_new_fallback(span.first_byte()),
+        }
+    }
+
+    /// Returns a span for the closing punctuation of the group only.
+    pub fn close(&self) -> Span {
+        match &self.inner {
+            #[cfg(wrap_proc_macro)]
+            DelimSpanEnum::Compiler { close, .. } => Span::_new(imp::Span::Compiler(*close)),
+            DelimSpanEnum::Fallback(span) => Span::_new_fallback(span.last_byte()),
+        }
+    }
+}
+
+impl Debug for DelimSpan {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        Debug::fmt(&self.join(), f)
+    }
+}