[PATCH 1/3] arm64: Prohibit probing on arch_stack_walk()

[Masami Hiramatsu (Google)]

On Thu, 1 Dec 2022 14:47:32 +0000
Mark Rutland <mark.rutland at arm.com> wrote:

> Hi Masami,
> 
> On Thu, Dec 01, 2022 at 11:39:02PM +0900, Masami Hiramatsu (Google) wrote:
> > From: Masami Hiramatsu (Google) <mhiramat at kernel.org>
> > 
> > Make arch_stack_walk() as NOKPROBE_SYMBOL and make other inline functions
> > called from arch_stack_walk() as nokprobe_inline so that user does not
> > put probe on it, because this function can be used from return_address()
> > which is already NOKPROBE_SYMBOL.
> 
> I think it would make sense to make this noinstr rather than NOKPROBES, since
> it's also used in the bowels of ftrace, and any instrumentation is liable to
> lead to some mutual recursion.

Yeah, OK. So use noinstr instead of notrace.

> 
> > Without this, if the kernel built with CONFIG_LOCKDEP=y, just probing
> > arch_stack_walk() via <tracefs>/kprobe_events will crash the kernel on
> > arm64.
> > 
> >  # echo p arch_stack_walk >> ${TRACEFS}/kprobe_events
> >  # echo 1 > ${TRACEFS}/events/kprobes/enable
> >   kprobes: Failed to recover from reentered kprobes.
> >   kprobes: Dump kprobe:
> >   .symbol_name = arch_stack_walk, .offset = 0, .addr = arch_stack_walk+0x0/0x1c0
> >   ------------[ cut here ]------------
> >   kernel BUG at arch/arm64/kernel/probes/kprobes.c:241!
> >   kprobes: Failed to recover from reentered kprobes.
> >   kprobes: Dump kprobe:
> >   .symbol_name = arch_stack_walk, .offset = 0, .addr = arch_stack_walk+0x0/0x1c0
> >   ------------[ cut here ]------------
> >   kernel BUG at arch/arm64/kernel/probes/kprobes.c:241!
> >   PREEMPT SMP
> >   Modules linked in:
> >   CPU: 0 PID: 17 Comm: migration/0 Tainted: G                 N 6.1.0-rc5+ #6
> >   Hardware name: linux,dummy-virt (DT)
> >   Stopper: 0x0 <- 0x0
> >   pstate: 600003c5 (nZCv DAIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
> >   pc : kprobe_breakpoint_handler+0x178/0x17c
> >   lr : kprobe_breakpoint_handler+0x178/0x17c
> >   sp : ffff8000080d3090
> >   x29: ffff8000080d3090 x28: ffff0df5845798c0 x27: ffffc4f59057a774
> >   x26: ffff0df5ffbba770 x25: ffff0df58f420f18 x24: ffff49006f641000
> >   x23: ffffc4f590579768 x22: ffff0df58f420f18 x21: ffff8000080d31c0
> >   x20: ffffc4f590579768 x19: ffffc4f590579770 x18: 0000000000000006
> >   x17: 5f6b636174735f68 x16: 637261203d207264 x15: 64612e202c30203d
> >   x14: 2074657366666f2e x13: 30633178302f3078 x12: 302b6b6c61775f6b
> >   x11: 636174735f686372 x10: ffffc4f590dc5bd8 x9 : ffffc4f58eb31958
> >   x8 : 00000000ffffefff x7 : ffffc4f590dc5bd8 x6 : 80000000fffff000
> >   x5 : 000000000000bff4 x4 : 0000000000000000 x3 : 0000000000000000
> >   x2 : 0000000000000000 x1 : ffff0df5845798c0 x0 : 0000000000000064
> >   Call trace:
> >   kprobes: Failed to recover from reentered kprobes.
> >   kprobes: Dump kprobe:
> >   .symbol_name = arch_stack_walk, .offset = 0, .addr = arch_stack_walk+0x0/0x1c0
> >   ------------[ cut here ]------------
> >   kernel BUG at arch/arm64/kernel/probes/kprobes.c:241!
> > 
> > Fixes: 39ef362d2d45 ("arm64: Make return_address() use arch_stack_walk()")
> > Cc: stable at vger.kernel.org
> > Signed-off-by: Masami Hiramatsu (Google) <mhiramat at kernel.org>
> > ---
> >  arch/arm64/kernel/stacktrace.c |    7 ++++---
> >  1 file changed, 4 insertions(+), 3 deletions(-)
> > 
> > diff --git a/arch/arm64/kernel/stacktrace.c b/arch/arm64/kernel/stacktrace.c
> > index 634279b3b03d..b0e913f944b4 100644
> > --- a/arch/arm64/kernel/stacktrace.c
> > +++ b/arch/arm64/kernel/stacktrace.c
> > @@ -23,7 +23,7 @@
> >   *
> >   * The regs must be on a stack currently owned by the calling task.
> >   */
> > -static inline void unwind_init_from_regs(struct unwind_state *state,
> > +static nokprobe_inline void unwind_init_from_regs(struct unwind_state *state,
> >  					 struct pt_regs *regs)
> >  {
> >  	unwind_init_common(state, current);
> > @@ -40,7 +40,7 @@ static inline void unwind_init_from_regs(struct unwind_state *state,
> >   *
> >   * The function which invokes this must be noinline.
> >   */
> > -static __always_inline void unwind_init_from_caller(struct unwind_state *state)
> > +static nokprobe_inline void unwind_init_from_caller(struct unwind_state *state)
> 
> This code must be __always_inline to get the correct caller context, so making
> this nokprobe_inline breaks a CONFIG_KPROBES=n build of the kernel.

Oops, indeed.

> 
> This also shouldn't be necessary; nokprobe_inline *is* __always_inline when
> CONFIG_KPROBES=y.
> 
> As above, I'd prefer that we mark these helpers as __always_inline and mark the
> main entry point as noinstr.

OK.

Thank you,

> 
> Thanks,
> Mark.
> 
> >  {
> >  	unwind_init_common(state, current);
> >  
> > @@ -58,7 +58,7 @@ static __always_inline void unwind_init_from_caller(struct unwind_state *state)
> >   * duration of the unwind, or the unwind will be bogus. It is never valid to
> >   * call this for the current task.
> >   */
> > -static inline void unwind_init_from_task(struct unwind_state *state,
> > +static nokprobe_inline void unwind_init_from_task(struct unwind_state *state,
> >  					 struct task_struct *task)
> >  {
> >  	unwind_init_common(state, task);
> > @@ -218,3 +218,4 @@ noinline notrace void arch_stack_walk(stack_trace_consume_fn consume_entry,
> >  
> >  	unwind(&state, consume_entry, cookie);
> >  }
> > +NOKPROBE_SYMBOL(arch_stack_walk);
> > 


-- 
Masami Hiramatsu (Google) <mhiramat at kernel.org>