Re: [RESEND][PATCH v3 14/17] static_call: Add static_cond_call()

[hpa]

On March 24, 2020 9:33:21 AM PDT, Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>On Tue, Mar 24, 2020 at 9:22 AM Andy Lutomirski <luto@xxxxxxxxxxxxxx>
>wrote:
>>
>> I havenât checked if static calls currently support return values,
>but
>> the conditional case only makes sense for functions that return void.
>>
>> Aside from that, it might be nice for passing NULL in to warn or bug
>> when the NULL pointer is stored instead of silently NOPping out the
>> call in cases where having a real implementation isnât optional.
>
>Both good points. I take back my question.
>
>And it aside from warning about passing in NULL then it doesn't work,
>I wonder if we could warn - at build time - when then using the COND
>version with a function that doesn't return void?
>
>Of course, one alternative is to just say "instead of using NOP, use
>'xorl %eax,%eax'", and then we'd have the rule that a NULL conditional
>function returns zero (or NULL).
>
>I _think_ a "xorl %eax,%eax ; retq" is just three bytes and would fit
>in the tailcall slot too.
>
>                Linus

"movl $0,%eax" is five bytes, the same length as a call. Doesn't work for a tailcall, still, although if the sequence:

    jmp tailcall
    retq

... can be generated at the tailcall site then the jmp can get patched out.

This would be equivalent to disabling tailcalls except that the stack frame is normally not unwound until between the call and the ret, so just disabling tailcalls from the compiler pov doesn't work.
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.