[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABCJKuc1LNPfO=RBDpOahGBgc7TF1L4D_ujhbrNfFWJaExsjpQ@mail.gmail.com>
Date: Tue, 14 Sep 2021 13:32:34 -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>,
clang-built-linux <clang-built-linux@...glegroups.com>
Subject: Re: [PATCH v3 04/16] cfi: Add DEFINE_CFI_IMMEDIATE_RETURN_STUB
On Tue, Sep 14, 2021 at 12:36 PM Nick Desaulniers
<ndesaulniers@...gle.com> wrote:
>
> On Tue, Sep 14, 2021 at 12:10 PM Sami Tolvanen <samitolvanen@...gle.com> wrote:
> >
> > This change introduces the DEFINE_CFI_IMMEDIATE_RETURN_STUB macro,
> > which defines a stub function that immediately returns and when
> > defined in the core kernel, always passes indirect call checking
> > with CONFIG_CFI_CLANG. Note that this macro should only be used when
> > a stub cannot be called using the correct function type.
> >
> > Signed-off-by: Sami Tolvanen <samitolvanen@...gle.com>
> > ---
> > include/asm-generic/vmlinux.lds.h | 11 +++++++++++
> > include/linux/cfi.h | 14 ++++++++++++++
> > kernel/cfi.c | 24 +++++++++++++++++++++++-
> > 3 files changed, 48 insertions(+), 1 deletion(-)
> >
> > diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
> > index f2984af2b85b..5b77284f7221 100644
> > --- a/include/asm-generic/vmlinux.lds.h
> > +++ b/include/asm-generic/vmlinux.lds.h
> > @@ -407,6 +407,16 @@
> > KEEP(*(.static_call_tramp_key)) \
> > __stop_static_call_tramp_key = .;
> >
> > +#ifdef CONFIG_CFI_CLANG
> > +#define CFI_EXCLUDED_DATA \
> > + . = ALIGN(8); \
> > + __start_cfi_excluded = .; \
> > + KEEP(*(.cfi_excluded_stubs)) \
> > + __stop_cfi_excluded = .;
> > +#else
> > +#define CFI_EXCLUDED_DATA
> > +#endif
> > +
> > /*
> > * Allow architectures to handle ro_after_init data on their
> > * own by defining an empty RO_AFTER_INIT_DATA.
> > @@ -430,6 +440,7 @@
> > __start_rodata = .; \
> > *(.rodata) *(.rodata.*) \
> > SCHED_DATA \
> > + CFI_EXCLUDED_DATA \
> > RO_AFTER_INIT_DATA /* Read only after init */ \
> > . = ALIGN(8); \
> > __start___tracepoints_ptrs = .; \
> > diff --git a/include/linux/cfi.h b/include/linux/cfi.h
> > index 879744aaa6e0..9ebf67a0d421 100644
> > --- a/include/linux/cfi.h
> > +++ b/include/linux/cfi.h
> > @@ -20,6 +20,18 @@ extern void __cfi_check(uint64_t id, void *ptr, void *diag);
> > #define __CFI_ADDRESSABLE(fn, __attr) \
> > const void *__cfi_jt_ ## fn __visible __attr = (void *)&fn
> >
> > +/*
> > + * Defines a stub function that returns immediately, and when defined and
> > + * referenced in the core kernel, always passes CFI checking. This should
> > + * be used only for stubs that cannot be called using the correct function
> > + * pointer type, which should be rare.
> > + */
> > +#define DEFINE_CFI_IMMEDIATE_RETURN_STUB(fn) \
> > + void fn(void) { return; } \
> > + const void *__cfi_excl_ ## fn __visible \
> > + __attribute__((__section__(".cfi_excluded_stubs"))) \
>
> Can we use __section from include/linux/compiler_attributes.h here
> rather than open coding the function attribute?
Sure. I'll change this in v4.
Sami
Powered by blists - more mailing lists