Re: [PATCH -tip v4 10/12] x86/kprobes: Push a fake return address at kretprobe_trampoline

[Josh Poimboeuf]

On Fri, Mar 26, 2021 at 10:20:09AM -0400, Steven Rostedt wrote:
> On Fri, 26 Mar 2021 21:03:49 +0900
> Masami Hiramatsu <mhiramat@xxxxxxxxxx> wrote:
> 
> > I confirmed this is not related to this series, but occurs when I build kernels with different
> > configs without cleanup.
> > 
> > Once I build kernel with CONFIG_UNWIND_GUESS=y (for testing), and after that,
> > I build kernel again with CONFIG_UNWIND_ORC=y (but without make clean), this
> > happened. In this case, I guess ORC data might be corrupted?
> > When I cleanup and rebuild, the stacktrace seems correct.
> 
> Hmm, that should be fixed. Seems like we are missing a dependency somewhere.

Thomas reported something similar: for example arch/x86/kernel/head_64.o
doesn't get rebuilt when changing unwinders.

  https://lkml.kernel.org/r/87tuqublrb.fsf@xxxxxxxxxxxxxxxxxxxxxxx

Masahiro, any idea how we can force a full tree rebuild when changing
the unwinder?

-- 
Josh