Re: [RFC PATCH v3 2/5] x86: Support for Static Trampolines

["Madhavan T. Venkataraman" <madvenka@linux.microsoft.com>]

On 1/28/21 8:48 PM, DJ Delorie wrote:
> 
> Sure, you still need to use the __CET_ENDBR macro in your trampoline.
> And it might be tricky to compute the number of iterations per page (I
> understand about the packing ;)
> 
> I'm just saying you don't need both CET and non-CET trampolines.
> 
> A CET-enabled trampoline should work in a CET-disabled system.
> 

You are right. I only needed two versions because of the nops for aligment.

However, I think I can address your concern. Now that we are able to
define ENDBR_PRESENT, we can define the trampoline size in the header like
this

#ifdef ENDBR_PRESENT
#define	UNIX64_TRAMP_SIZE	40
#else
#define UNIX64_TRAMP_SIZE	32
#endif

Then, in the trampoline code table:

C(trampoline_code_table):
        .rept   UNIX64_TRAMP_MAP_SIZE / UNIX64_TRAMP_SIZE
        _CET_ENDBR
        subq    $16, %rsp               /* Make space on the stack */
        movq    %r10, (%rsp)            /* Save %r10 on stack */
        movq    4077(%rip), %r10        /* Copy data into %r10 */
        movq    %r10, 8(%rsp)           /* Save data on stack */
        movq    4073(%rip), %r10        /* Copy code into %r10 */
        jmp     *%r10                   /* Jump to code */
	nop
	nop
#ifdef ENDBR_PRESENT
        nop
        nop
        nop
        nop
#endif
        .endr
ENDF(C(trampoline_code_table))


This is cool. We only need one table.

Thanks.

Madhavan