Re: [PATCH] static_call,x86: Robustify trampoline patching

[Peter Zijlstra]

On Tue, Nov 02, 2021 at 10:35:30AM -0700, Kees Cook wrote:
> On Sat, Oct 30, 2021 at 10:16:31AM +0200, Peter Zijlstra wrote:
> > foo.cfi:
> > 	endbr
> > 	xorl $0xdeadbeef, %r10d
> > 	jz foo
> > 	ud2
> > 	nop	# make it an even 16 bytes
> > foo:
> > 	# actual function text
> > 
> > 
> > Then have the address of foo, be the address of foo, like any normal
> > sane person would expect. Have direct calls to foo, go to foo, again, as
> > expected.
> > 
> > When doing an indirect call (to r11, as clang does), then, and only
> > then, do:
> > 
> > 	movl $0xdeadbeef, %r10d
> > 	subq $0x10, %r11
> > 	call *%r11
> > 
> > 	# if the r11 lives, add:
> > 	addq $0x10, %r11
> > 
> > 
> > Then only when caller and callee agree 0xdeadbeef is the password, does
> > the indirect call go through.
> > 
> > Why isn't this a suitable CFI scheme even without IBT?
> 
> The trouble is that the callee is doing the verification. There's no
> protection against calling into a callee that doesn't perform a check
> (e.g. BPF JIT, or otherwise constructed executable memory, etc). The
> caller needs to do the verification that what they're calling into is
> safe before it makes the call.

Right, Ard said the same, see new crackpot scheme here:

  https://lkml.kernel.org/r/YYE1yPClPMHvyvIt@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx