Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 5 May 2020 11:38:40 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc:     x86@kernel.org, linux-kernel@vger.kernel.org, rostedt@goodmis.org,
        mhiramat@kernel.org, bristot@redhat.com, jbaron@akamai.com,
        torvalds@linux-foundation.org, tglx@linutronix.de,
        mingo@kernel.org, namit@vmware.com, hpa@zytor.com, luto@kernel.org,
        ard.biesheuvel@linaro.org, jpoimboe@redhat.com,
        pbonzini@redhat.com, mathieu.desnoyers@efficios.com
Subject: Re: [PATCH v4 14/18] static_call: Add static_cond_call()
Message-ID: <20200505093840.GF5298@hirez.programming.kicks-ass.net>
References: <20200501202849.647891881@infradead.org>
 <20200501202944.593400184@infradead.org>
 <1238787e-d97d-f09b-d76d-2df2dc273f4b@rasmusvillemoes.dk>
 <20200503125813.GL3762@hirez.programming.kicks-ass.net>
 <a53369f3-665a-af0e-efad-09ae456af847@rasmusvillemoes.dk>
 <20200504201445.GQ3762@hirez.programming.kicks-ass.net>
 <6cb31c72-ce0e-7923-daca-ce5ef414940b@rasmusvillemoes.dk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <6cb31c72-ce0e-7923-daca-ce5ef414940b@rasmusvillemoes.dk>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, May 05, 2020 at 09:50:26AM +0200, Rasmus Villemoes wrote:
> On 04/05/2020 22.14, Peter Zijlstra wrote:

> Anyway, it's hard to judge what version of the !HAVE_STATIC_CALL
> implementation is best when there's no !HAVE_STATIC_CALL use cases to
> look at. I just want to ensure that whatever limitations or gotchas
> (e.g., "arguments are evaluated regardless of NULLness of func", or
> alternatively "arguments must not have side effects") it ends up with
> get documented.

I can certainly try and write a better comment for it.

> > #define __static_cond_call(name) \
> > ({ \
> > 	void *func = READ_ONCE(name.func); \
> > 	if (!func) \
> > 		func = &__static_call_nop; \
> > 	(typeof(__SCT__##name)*)func; \
> > })
> 
> I think you can just make it
> 
> #define __static_cond_call(name) \
> ( \
> (typeof(__SCT__##name)*) ((void *)name.func ? : (void *)__static_call_nop) \
> )

I _think_ the compiler sees the two as exactly the same, but then, I've
not seen the inside of a modern compiler.

> but that simplification is not enough to make gcc change its mind about
> how to compile it :( But I'm guessing that various sanitizers or static
> checkers might complain about the UB.

Yeah, we'll see.