Re: [PATCH v3 2/2] lkdtm: Add Shadow Call Stack tests

[Dan Li]

On 3/4/22 06:54, Dan Li wrote:
> On 3/3/22 11:09, Kees Cook wrote:
> > On Thu, Mar 03, 2022 at 10:42:45AM -0800, Kees Cook wrote:
> > > Though, having the IS_ENABLED in there makes me wonder if this test
> > > should instead be made _survivable_ on failure. Something like this,
> > > completely untested:
> > >
> > >
> > > And we should, actually, be able to make the "set_lr" functions be
> > > arch-specific, leaving the test itself arch-agnostic....
> >
> > Yeah, as a tested example, this works for x86_64, and based on what you
> > had, I'd expect it to work on arm64 too:
> >
> > #include <stdio.h>
> >
> > static __attribute__((noinline))
> > void set_return_addr(unsigned long *expected, unsigned long *addr)
> > {
> >      /* Use of volatile is to make sure final write isn't seen as a dead store. */
> >      unsigned long * volatile *ret_addr = (unsigned long **)__builtin_frame_address(0) + 1;
> >
> >      /* Make sure we've found the right place on the stack before writing it. */
> >      if (*ret_addr == expected)
> >          *ret_addr = addr;
> > }
> >
> > volatile int force_label;
> > int main(void)
> > {
> >      do {
> >          /* Keep labels in scope. */
> >          if (force_label)
> >              goto normal;
> >          if (force_label)
> >              goto redirected;
> >
> >          set_return_addr(&&normal, &&redirected);
> > normal:
> >          printf("I should be skipped\n");
> >          break;
>
>  From the assembly code, it seems that "&&normal" does't always equal
> to the address of label "normal" when we use clang with -O2.
>
> > redirected:
> >          printf("Redirected\n");
> >      } while (0);
>
> The address of "&&redirected" may appear in the middle of the assembly
> instructions of the printf. If we unconditionally jump to "&&normal",> it may crash directly because x0 is not set correctly.

Sorry, it should be:
The address of "&&redirected" may appear in the middle of the assembly
instructions of the printf. If we unconditionally jump to "&&redirected",
it may crash directly because x0 of printf is not set correctly.

Thanks,
Dan.
> >      return 0;
> > }
> >
> >
> > It does _not_ work under Clang, though, which I'm still looking at.
>
> AFAICT, maybe we could specify -O0 optimization to bypass this.
>
>
> BTW:
> Occasionally found, the following code works correctly, but i think
> it doesn't solve the issue :)
>
> #include <stdio.h>
>
> static __attribute__((noinline))
> void set_return_addr(unsigned long *expected, unsigned long *addr)
> {
>      /* Use of volatile is to make sure final write isn't seen as a dead store. */
>      unsigned long * volatile *ret_addr = (unsigned long **)__builtin_frame_address(0) + 1;
>
>      /* Make sure we've found the right place on the stack before writing it. */
> //    if (*ret_addr == expected)
>          *ret_addr = addr;
> }
> volatile int force_label;
> int main(void)
> {
>      do {
>          /* Keep labels in scope. */
>          if (force_label)
>              goto normal;
>          if (force_label)
>              goto redirected;
>
>          set_return_addr(&&normal, &&redirected);
> normal:
>          printf("I should be skipped\n");
>          break;
>
> redirected:
>          printf("Redirected\n");
>          printf("\n");                //add a new printf
>      } while (0);
>
>      return 0;
> }