Re: [PATCH net-next 1/4] indirect call wrappers: helpers to speed-up indirect calls of builtin

[Edward Cree <ecree@solarflare.com>]

On 03/12/18 11:40, Paolo Abeni wrote:
> This header define a bunch of helpers that allow avoiding the
> retpoline overhead when calling builtin functions via function pointers.
> It boils down to explicitly comparing the function pointers to
> known builtin functions and eventually invoke directly the latter.
>
> The macros defined here implement the boilerplate for the above schema
> and will be used by the next patches.
>
> rfc -> v1:
>  - use branch prediction hint, as suggested by Eric
>
> Suggested-by: Eric Dumazet <Eric Dumazet edumazet@google.com>
> Signed-off-by: Paolo Abeni <pabeni@redhat.com>
> ---
I'm not sure I see the reason why this is done with numbers and
 'name ## NR', adding extra distance between the callsite and the
 list of callees.  In particular it means that each callable needs
 to specify its index.
Wouldn't it be simpler just to have
    #define INDIRECT_CALL_1(f, f1, ...) \
        (likely(f == f1) ? f1(__VA_ARGS__) : f(__VA_ARGS__))
    #define INDIRECT_CALL_2(f, f2, f1, ...) \
        (likely(f == f2) ? f2(__VA_ARGS__) : INDIRECT_CALL_1(f, f1, __VA_ARGS__))
etc.?  Removing the need for INDIRECT_CALLABLE_DECLARE_* entirely.

At least the commit message should explain the rationale for not
 doing things this way.

-Ed

PS: this has reminded me of my desire to try runtime creation of
 these kinds of branch tables with self-modifying code; is there
 any documentation on how to go about writing to kernel .text at
 runtime?  Last time I had a try at it I got very confused.