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, 197 insertions, 0 deletions

diff --git a/arch/mips/vdso/Makefile b/arch/mips/vdso/Makefile
new file mode 100644
index 000000000..00d3ba2c4
--- /dev/null
+++ b/arch/mips/vdso/Makefile
@@ -0,0 +1,197 @@
+# SPDX-License-Identifier: GPL-2.0
+# Objects to go into the VDSO.
+
+# Include the generic Makefile to check the built vdso.
+include $(srctree)/lib/vdso/Makefile.include
+
+obj-vdso-y := elf.o sigreturn.o
+
+# Common compiler flags between ABIs.
+ccflags-vdso := \
+	$(filter -I%,$(KBUILD_CFLAGS)) \
+	$(filter -E%,$(KBUILD_CFLAGS)) \
+	$(filter -mmicromips,$(KBUILD_CFLAGS)) \
+	$(filter -march=%,$(KBUILD_CFLAGS)) \
+	$(filter -m%-float,$(KBUILD_CFLAGS)) \
+	$(filter -mno-loongson-%,$(KBUILD_CFLAGS)) \
+	$(CLANG_FLAGS) \
+	-D__VDSO__
+
+ifndef CONFIG_64BIT
+ccflags-vdso += -DBUILD_VDSO32
+endif
+
+#
+# The -fno-jump-tables flag only prevents the compiler from generating
+# jump tables but does not prevent the compiler from emitting absolute
+# offsets.
+cflags-vdso := $(ccflags-vdso) \
+	$(filter -W%,$(filter-out -Wa$(comma)%,$(KBUILD_CFLAGS))) \
+	$(filter -std=%,$(KBUILD_CFLAGS)) \
+	-O3 -g -fPIC -fno-strict-aliasing -fno-common -fno-builtin -G 0 \
+	-mrelax-pic-calls $(call cc-option, -mexplicit-relocs) \
+	-fno-stack-protector -fno-jump-tables -DDISABLE_BRANCH_PROFILING \
+	$(call cc-option, -fno-asynchronous-unwind-tables)
+aflags-vdso := $(ccflags-vdso) \
+	-D__ASSEMBLY__ -Wa,-gdwarf-2
+
+ifneq ($(c-gettimeofday-y),)
+obj-vdso-y += vgettimeofday.o
+CFLAGS_vgettimeofday.o = -include $(c-gettimeofday-y)
+
+# config-n32-o32-env.c prepares the environment to build a 32bit vDSO
+# library on a 64bit kernel.
+# Note: Needs to be included before than the generic library.
+CFLAGS_vgettimeofday-o32.o = -include $(src)/config-n32-o32-env.c -include $(c-gettimeofday-y)
+CFLAGS_vgettimeofday-n32.o = -include $(src)/config-n32-o32-env.c -include $(c-gettimeofday-y)
+endif
+
+CFLAGS_REMOVE_vgettimeofday.o = $(CC_FLAGS_FTRACE)
+
+# VDSO linker flags.
+ldflags-y := -Bsymbolic --no-undefined -soname=linux-vdso.so.1 \
+	$(filter -E%,$(KBUILD_CFLAGS)) -shared \
+	-G 0 --eh-frame-hdr --hash-style=sysv --build-id=sha1 -T
+
+CFLAGS_REMOVE_vdso.o = $(CC_FLAGS_FTRACE)
+
+# Check that we don't have PIC 'jalr t9' calls left
+quiet_cmd_vdso_mips_check = VDSOCHK $@
+      cmd_vdso_mips_check = if $(OBJDUMP) --disassemble $@ | grep -E -h "jalr.*t9" > /dev/null; \
+		       then (echo >&2 "$@: PIC 'jalr t9' calls are not supported"; \
+			     rm -f $@; /bin/false); fi
+
+#
+# Shared build commands.
+#
+
+quiet_cmd_vdsold_and_vdso_check = LD      $@
+      cmd_vdsold_and_vdso_check = $(cmd_ld); $(cmd_vdso_check); $(cmd_vdso_mips_check)
+
+quiet_cmd_vdsoas_o_S = AS      $@
+      cmd_vdsoas_o_S = $(CC) $(a_flags) -c -o $@ $<
+
+# Strip rule for the raw .so files
+$(obj)/%.so.raw: OBJCOPYFLAGS := -S
+$(obj)/%.so.raw: $(obj)/%.so.dbg.raw FORCE
+	$(call if_changed,objcopy)
+
+hostprogs := genvdso
+
+quiet_cmd_genvdso = GENVDSO $@
+define cmd_genvdso
+	$(foreach file,$(filter %.raw,$^),cp $(file) $(file:%.raw=%) &&) \
+	$(obj)/genvdso $(<:%.raw=%) $(<:%.dbg.raw=%) $@ $(VDSO_NAME)
+endef
+
+#
+# Build native VDSO.
+#
+
+native-abi := $(filter -mabi=%,$(KBUILD_CFLAGS))
+
+targets += $(obj-vdso-y)
+targets += vdso.lds
+targets += vdso.so.dbg.raw vdso.so.raw
+targets += vdso.so.dbg vdso.so
+targets += vdso-image.c
+
+obj-vdso := $(obj-vdso-y:%.o=$(obj)/%.o)
+
+$(obj-vdso): KBUILD_CFLAGS := $(cflags-vdso) $(native-abi)
+$(obj-vdso): KBUILD_AFLAGS := $(aflags-vdso) $(native-abi)
+
+$(obj)/vdso.lds: KBUILD_CPPFLAGS := $(ccflags-vdso) $(native-abi)
+
+$(obj)/vdso.so.dbg.raw: $(obj)/vdso.lds $(obj-vdso) FORCE
+	$(call if_changed,vdsold_and_vdso_check)
+
+$(obj)/vdso-image.c: $(obj)/vdso.so.dbg.raw $(obj)/vdso.so.raw \
+                     $(obj)/genvdso FORCE
+	$(call if_changed,genvdso)
+
+obj-y += vdso-image.o
+
+#
+# Build O32 VDSO.
+#
+
+# Define these outside the ifdef to ensure they are picked up by clean.
+targets += $(obj-vdso-y:%.o=%-o32.o)
+targets += vdso-o32.lds
+targets += vdso-o32.so.dbg.raw vdso-o32.so.raw
+targets += vdso-o32.so.dbg vdso-o32.so
+targets += vdso-o32-image.c
+
+ifdef CONFIG_MIPS32_O32
+
+obj-vdso-o32 := $(obj-vdso-y:%.o=$(obj)/%-o32.o)
+
+$(obj-vdso-o32): KBUILD_CFLAGS := $(cflags-vdso) -mabi=32
+$(obj-vdso-o32): KBUILD_AFLAGS := $(aflags-vdso) -mabi=32
+
+$(obj)/%-o32.o: $(src)/%.S FORCE
+	$(call if_changed_dep,vdsoas_o_S)
+
+$(obj)/%-o32.o: $(src)/%.c FORCE
+	$(call cmd,force_checksrc)
+	$(call if_changed_rule,cc_o_c)
+
+$(obj)/vdso-o32.lds: KBUILD_CPPFLAGS := $(ccflags-vdso) -mabi=32
+$(obj)/vdso-o32.lds: $(src)/vdso.lds.S FORCE
+	$(call if_changed_dep,cpp_lds_S)
+
+$(obj)/vdso-o32.so.dbg.raw: $(obj)/vdso-o32.lds $(obj-vdso-o32) FORCE
+	$(call if_changed,vdsold_and_vdso_check)
+
+$(obj)/vdso-o32-image.c: VDSO_NAME := o32
+$(obj)/vdso-o32-image.c: $(obj)/vdso-o32.so.dbg.raw $(obj)/vdso-o32.so.raw \
+                         $(obj)/genvdso FORCE
+	$(call if_changed,genvdso)
+
+obj-y += vdso-o32-image.o
+
+endif
+
+#
+# Build N32 VDSO.
+#
+
+targets += $(obj-vdso-y:%.o=%-n32.o)
+targets += vdso-n32.lds
+targets += vdso-n32.so.dbg.raw vdso-n32.so.raw
+targets += vdso-n32.so.dbg vdso-n32.so
+targets += vdso-n32-image.c
+
+ifdef CONFIG_MIPS32_N32
+
+obj-vdso-n32 := $(obj-vdso-y:%.o=$(obj)/%-n32.o)
+
+$(obj-vdso-n32): KBUILD_CFLAGS := $(cflags-vdso) -mabi=n32
+$(obj-vdso-n32): KBUILD_AFLAGS := $(aflags-vdso) -mabi=n32
+
+$(obj)/%-n32.o: $(src)/%.S FORCE
+	$(call if_changed_dep,vdsoas_o_S)
+
+$(obj)/%-n32.o: $(src)/%.c FORCE
+	$(call cmd,force_checksrc)
+	$(call if_changed_rule,cc_o_c)
+
+$(obj)/vdso-n32.lds: KBUILD_CPPFLAGS := $(ccflags-vdso) -mabi=n32
+$(obj)/vdso-n32.lds: $(src)/vdso.lds.S FORCE
+	$(call if_changed_dep,cpp_lds_S)
+
+$(obj)/vdso-n32.so.dbg.raw: $(obj)/vdso-n32.lds $(obj-vdso-n32) FORCE
+	$(call if_changed,vdsold_and_vdso_check)
+
+$(obj)/vdso-n32-image.c: VDSO_NAME := n32
+$(obj)/vdso-n32-image.c: $(obj)/vdso-n32.so.dbg.raw $(obj)/vdso-n32.so.raw \
+                         $(obj)/genvdso FORCE
+	$(call if_changed,genvdso)
+
+obj-y += vdso-n32-image.o
+
+endif
+
+# FIXME: Need install rule for debug.
+# Needs to deal with dependency for generation of dbg by cmd_genvdso...