Re: [PATCH v3 22/29] riscv sigcontext: adding cfi state field in sigcontext

[Deepak Gupta]

On Fri, May 24, 2024 at 05:46:16PM +0800, Andy Chiu wrote:
> Hi Deepak,
>
> On Thu, Apr 4, 2024 at 7:42 AM Deepak Gupta <debug@xxxxxxxxxxxx> wrote:
> > Shadow stack needs to be saved and restored on signal delivery and signal
> > return.
> >
> > sigcontext embedded in ucontext is extendible. Adding cfi state in there
> > which can be used to save cfi state before signal delivery and restore
> > cfi state on sigreturn
> >
> > Signed-off-by: Deepak Gupta <debug@xxxxxxxxxxxx>
> > ---
> >  arch/riscv/include/uapi/asm/sigcontext.h | 5 +++++
> >  1 file changed, 5 insertions(+)
> >
> > diff --git a/arch/riscv/include/uapi/asm/sigcontext.h b/arch/riscv/include/uapi/asm/sigcontext.h
> > index cd4f175dc837..5ccdd94a0855 100644
> > --- a/arch/riscv/include/uapi/asm/sigcontext.h
> > +++ b/arch/riscv/include/uapi/asm/sigcontext.h
> > @@ -21,6 +21,10 @@ struct __sc_riscv_v_state {
> >         struct __riscv_v_ext_state v_state;
> >  } __attribute__((aligned(16)));
> >
> > +struct __sc_riscv_cfi_state {
> > +       unsigned long ss_ptr;   /* shadow stack pointer */
> > +       unsigned long rsvd;             /* keeping another word reserved in case we need it */
> > +};
> >  /*
> >   * Signal context structure
> >   *
> > @@ -29,6 +33,7 @@ struct __sc_riscv_v_state {
> >   */
> >  struct sigcontext {
> >         struct user_regs_struct sc_regs;
> > +       struct __sc_riscv_cfi_state sc_cfi_state;
>
> I am concerned about this change as this could potentially break uabi.
> Let's say there is a pre-CFI program running on this kernel. It
> receives a signal so the kernel lays out the sig-stack as presented in
> this structure. If the program accesses sc_fpregs, it would now get
> sc_cfi_state. As the offset has changed, and the pre-CFI program has
> not been re-compiled.

Yeah this is a problem if program was built with older kernel/old toolchain
(or cfi unaware toolchain). Thanks.

> >         union {
> >                 union __riscv_fp_state sc_fpregs;
> >                 struct __riscv_extra_ext_header sc_extdesc;
> > --
> > 2.43.2
>
> There may be two ways to deal with this. One is to use a different
> signal ABI for CFI-enabled programs. This may complicate the user
> space because new programs will have to determine whether it should
> use the CFI-ABI at run time. Another way is to follow what Vector does
> for signal stack. It adds a way to introduce new extensions on signal
> stack without impacting ABI.
>
> Please let me know if I misunderstand anything, thanks.

I think following how vector does would be cleaner.
Let me munch on this a little bit.

> Cheers,
> Andy