[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABCJKue1dDYe9aL0ExU8NMORxJN-7qhF=D+=sSKr0d2Quhb3YA@mail.gmail.com>
Date: Fri, 1 Oct 2021 13:08:47 -0700
From: Sami Tolvanen <samitolvanen@...gle.com>
To: Nick Desaulniers <ndesaulniers@...gle.com>
Cc: X86 ML <x86@...nel.org>, Kees Cook <keescook@...omium.org>,
Josh Poimboeuf <jpoimboe@...hat.com>,
Peter Zijlstra <peterz@...radead.org>,
Nathan Chancellor <nathan@...nel.org>,
Sedat Dilek <sedat.dilek@...il.com>,
linux-hardening@...r.kernel.org,
LKML <linux-kernel@...r.kernel.org>, llvm@...ts.linux.dev
Subject: Re: [PATCH v4 05/15] tracepoint: Exclude tp_stub_func from CFI checking
On Thu, Sep 30, 2021 at 11:51 AM Nick Desaulniers
<ndesaulniers@...gle.com> wrote:
>
> On Thu, Sep 30, 2021 at 11:05 AM Sami Tolvanen <samitolvanen@...gle.com> wrote:
> >
> > If allocate_probes fails, func_remove replaces the old function
> > with a pointer to tp_stub_func, which is called using a mismatching
> > function pointer that will always trip indirect call checks with
> > CONFIG_CFI_CLANG. Use DEFINE_CFI_IMMEDATE_RETURN_STUB to define
> > tp_stub_func to allow it to pass CFI checking.
> >
> > Signed-off-by: Sami Tolvanen <samitolvanen@...gle.com>
> > Reviewed-by: Nick Desaulniers <ndesaulniers@...gle.com>
> > ---
> > kernel/tracepoint.c | 5 +----
> > 1 file changed, 1 insertion(+), 4 deletions(-)
> >
> > diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c
> > index 64ea283f2f86..58acc7d86c3f 100644
> > --- a/kernel/tracepoint.c
> > +++ b/kernel/tracepoint.c
>
> looking at 4+5/15 together, I wonder if this TU should explicitly
> include linux/cfi.h?
Good point. Currently cfi.h is included in module.h, but including it
explicitly makes this less likely to break in future. I'll add an
explicit include in v5. Thanks for taking a look!
Sami
Powered by blists - more mailing lists