> Then, I noticed that invoke_syscall generates instructions to add random offset
> in sp when RANDOMIZE_KSTACK_OFFSET=y, which is true in the above case.

I'm also seeing this behavior when compiling with
RANDOMIZE_KSTACK_OFFSET=y. I wonder if a special hint type
could/should be added to allow for skipping the reliability check for
stack frames with this randomized offset? Forgive me if this is a
naive suggestion.

Thanks,
Dylan

Sorry for the delay. Someone else also pointed out the same thing. I just have not
had the chance to look at this. I plan to soon.

Thanks.

Madhavan

On 4/3/23 17:26, Dylan Hatch wrote:
>> Then, I noticed that invoke_syscall generates instructions to add random offset
>> in sp when RANDOMIZE_KSTACK_OFFSET=y, which is true in the above case.
> 
> I'm also seeing this behavior when compiling with
> RANDOMIZE_KSTACK_OFFSET=y. I wonder if a special hint type
> could/should be added to allow for skipping the reliability check for
> stack frames with this randomized offset? Forgive me if this is a
> naive suggestion.
> 
> Thanks,
> Dylan