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, 80 insertions, 0 deletions

diff --git a/scripts/dtc/update-dtc-source.sh b/scripts/dtc/update-dtc-source.sh
new file mode 100755
index 000000000..946275415
--- /dev/null
+++ b/scripts/dtc/update-dtc-source.sh
@@ -0,0 +1,80 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+# Simple script to update the version of DTC carried by the Linux kernel
+#
+# This script assumes that the dtc and the linux git trees are in the
+# same directory. After building dtc in the dtc directory, it copies the
+# source files and generated source file(s) into the scripts/dtc directory
+# in the kernel and creates a git commit updating them to the new
+# version.
+#
+# Usage: from the top level Linux source tree, run:
+# $ ./scripts/dtc/update-dtc-source.sh
+#
+# The script will change into the dtc tree, build and test dtc, copy the
+# relevant files into the kernel tree and create a git commit. The commit
+# message will need to be modified to reflect the version of DTC being
+# imported
+#
+# TODO:
+# This script is pretty basic, but it is seldom used so a few manual tasks
+# aren't a big deal. If anyone is interested in making it more robust, the
+# the following would be nice:
+# * Actually fail to complete if any testcase fails.
+#   - The dtc "make check" target needs to return a failure
+# * Extract the version number from the dtc repo for the commit message
+# * Build dtc in the kernel tree
+# * run 'make check" on dtc built from the kernel tree
+
+set -ev
+
+DTC_UPSTREAM_PATH=`pwd`/../dtc
+DTC_LINUX_PATH=`pwd`/scripts/dtc
+
+DTC_SOURCE="checks.c data.c dtc.c dtc.h flattree.c fstree.c livetree.c srcpos.c \
+		srcpos.h treesource.c util.c util.h version_gen.h \
+		dtc-lexer.l dtc-parser.y"
+LIBFDT_SOURCE="fdt.c fdt.h fdt_addresses.c fdt_empty_tree.c \
+		fdt_overlay.c fdt_ro.c fdt_rw.c fdt_strerror.c fdt_sw.c \
+		fdt_wip.c libfdt.h libfdt_env.h libfdt_internal.h"
+FDTOVERLAY_SOURCE=fdtoverlay.c
+
+get_last_dtc_version() {
+	git log --oneline scripts/dtc/ | grep 'upstream' | head -1 | sed -e 's/^.* \(.*\)/\1/'
+}
+
+last_dtc_ver=$(get_last_dtc_version)
+
+# Build DTC
+cd $DTC_UPSTREAM_PATH
+make clean
+make check
+dtc_version=$(git describe HEAD)
+dtc_log=$(git log --oneline ${last_dtc_ver}..)
+
+
+# Copy the files into the Linux tree
+cd $DTC_LINUX_PATH
+for f in $DTC_SOURCE $FDTOVERLAY_SOURCE; do
+	cp ${DTC_UPSTREAM_PATH}/${f} ${f}
+	git add ${f}
+done
+for f in $LIBFDT_SOURCE; do
+       cp ${DTC_UPSTREAM_PATH}/libfdt/${f} libfdt/${f}
+       git add libfdt/${f}
+done
+
+sed -i -- 's/#include <libfdt_env.h>/#include "libfdt_env.h"/g' ./libfdt/libfdt.h
+sed -i -- 's/#include <fdt.h>/#include "fdt.h"/g' ./libfdt/libfdt.h
+git add ./libfdt/libfdt.h
+
+commit_msg=$(cat << EOF
+scripts/dtc: Update to upstream version ${dtc_version}
+
+This adds the following commits from upstream:
+
+${dtc_log}
+EOF
+)
+
+git commit -e -v -s -m "${commit_msg}"