Re: [PATCH] unify the use of FUTEX_PRIVATE

[Jens Gustedt <jens.gustedt@inria.fr>]

[-- Attachment #1: Type: text/plain, Size: 1731 bytes --]

Hello Rich,

On Sat, 24 Jun 2017 18:16:30 -0400 Rich Felker <dalias@libc.org> wrote:

> For now let's just use a macro for FUTEX_PRIVATE.

Ok, was my guess :)

> My concern was
> mainly about whether there are cases where it's assumed that the 128
> or ^128 translates to FUTEX_PRIVATE by arithmetic ops. The __wait,
> __timedwait, and __wake interfaces just treat their argument as a
> boolean now, so I think the only places there might be remaining
> assumptions of that sort is where a flag bit from the sync object is
> passed directly into a futex __syscall -- places like the condvar
> implementation. You've looked at all this code just now so you
> probably have a better idea if there are any such places;

All the usages that I found go through a Boolean operation. The only
place where -maybe- the compiler might shortcut some of it is __wake
because it is inlined. In any case, by just using the macro the
compiler see's the same 128.

> if so I
> think the assumption that the bits match up should just be replaced by
> a conditional (e.g. flags & 128 ? 0 : FUTEX_PRIVATE or such).

There is none in the arch independent code, so you could go with my
patch as it is, I think.

I only found one other place, but which I don't understand (and I am
not sure that I want to :). This is in
src/thread/x32/syscall_cp_fixup.c
where there is a special casing for SYS_futex

Thanks
Jens

-- 
:: INRIA Nancy Grand Est ::: Camus ::::::: ICube/ICPS :::
:: ::::::::::::::: office Strasbourg : +33 368854536   ::
:: :::::::::::::::::::::: gsm France : +33 651400183   ::
:: ::::::::::::::: gsm international : +49 15737185122 ::
:: http://icube-icps.unistra.fr/index.php/Jens_Gustedt ::

[-- Attachment #2: Digitale Signatur von OpenPGP --]
[-- Type: application/pgp-signature, Size: 195 bytes --]