Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Content-Type: text/plain;
        charset=us-ascii
Mime-Version: 1.0 (1.0)
Subject: Re: [PATCH 8/9] perf/breakpoint: Split breakpoint "check" and "commit"
From:   Andy Lutomirski <luto@amacapital.net>
In-Reply-To: <20180516031059.GA2271@lerouge>
Date:   Tue, 15 May 2018 21:58:03 -0700
Cc:     Peter Zijlstra <peterz@infradead.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Jiri Olsa <jolsa@redhat.com>,
        Namhyung Kim <namhyung@kernel.org>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Yoshinori Sato <ysato@users.sourceforge.jp>,
        Benjamin Herrenschmidt <benh@kernel.crashing.org>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Chris Zankel <chris@zankel.net>,
        Paul Mackerras <paulus@samba.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Will Deacon <will.deacon@arm.com>,
        Michael Ellerman <mpe@ellerman.id.au>,
        Rich Felker <dalias@libc.org>, Ingo Molnar <mingo@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Andy Lutomirski <luto@kernel.org>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        Max Filippov <jcmvbkbc@gmail.com>
Content-Transfer-Encoding: quoted-printable
Message-Id: <99DA1A94-6AF6-4B9A-858A-7018D58CC38E@amacapital.net>
References: <1525634395-23380-1-git-send-email-frederic@kernel.org> <1525634395-23380-9-git-send-email-frederic@kernel.org> <20180509091703.GH12217@hirez.programming.kicks-ass.net> <20180516031059.GA2271@lerouge>
To:     Frederic Weisbecker <frederic@kernel.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On May 15, 2018, at 8:11 PM, Frederic Weisbecker <frederic@kernel.org> wro=
te:
>=20
>> On Wed, May 09, 2018 at 11:17:03AM +0200, Peter Zijlstra wrote:
>>> On Sun, May 06, 2018 at 09:19:54PM +0200, Frederic Weisbecker wrote:
>>> arch/arm/include/asm/hw_breakpoint.h     |  5 ++++-
>>> arch/arm/kernel/hw_breakpoint.c          | 22 +++-------------------
>>> arch/arm64/include/asm/hw_breakpoint.h   |  5 ++++-
>>> arch/arm64/kernel/hw_breakpoint.c        | 22 +++-------------------
>>> arch/powerpc/include/asm/hw_breakpoint.h |  5 ++++-
>>> arch/powerpc/kernel/hw_breakpoint.c      | 22 +++-------------------
>>> arch/sh/include/asm/hw_breakpoint.h |  5 ++++-
>>> arch/sh/kernel/hw_breakpoint.c           | 22 +++-------------------
>>> arch/x86/include/asm/hw_breakpoint.h     |  5 ++++-
>>> arch/x86/kernel/hw_breakpoint.c          | 23 +++--------------------
>>> arch/xtensa/include/asm/hw_breakpoint.h  |  5 ++++-
>>> arch/xtensa/kernel/hw_breakpoint.c       | 22 +++-------------------
>>=20
>> Because of those ^,
>>=20
>>> kernel/events/hw_breakpoint.c            | 11 ++++++-----
>>=20
>> would it not make sense to have a prelimenary patch doing something
>> like:
>>=20
>> __weak int hw_breakpoint_arch_check(struct perf_event *bp)
>> {
>>    return arch_validate_hwbkpt_settings(bp);
>> }
>=20
> So eventually I fear I can't do that, due to linking order.
>=20
> Say I convert x86 to implement hw_breakpoint_arch_check(), so I
> remove arch_validate_hwbkpt_settings(). On build time, the weak version
> is still compiled and can't find a declaration for arch_validate_hwbkpt_se=
ttings().
>=20
> I tried to keep the declaration while the definition has been removed but
> it seems the weak version is linked first before it gets later replaced by=

> the overriden arch version. So I get a build error.
>=20
> I could keep arch_validate_hwbkpt_settings() around on all archs and remov=
e it in
> the end with the weak version but that would defeat the purpose of removin=
g
> the mid-state in the current patch.

How about just not using weak functions?  Weak functions have annoying issue=
s like this, and they have trouble generating good code. I much prefer the p=
attern:

in arch header:
extern void arch_func(whatever);
#define arch_func arch_func

in generic header:
#ifndef arch_func
static inline void arch_func(whatever) ...
#endif=