linux-kernel - Re: [PATCH v4 2/5] irqchip: sifive-plic: Don't inline plic_toggle() and plic_irq

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAAhSdy1JOFW3f_xtZSjjHPzcHWoG5C50tXzm7U=Oav9GYD=j8g@mail.gmail.com>
Date:   Sat, 19 Jan 2019 10:30:52 +0530
From:   Anup Patel <anup@...infault.org>
To:     Christoph Hellwig <hch@...radead.org>
Cc:     Palmer Dabbelt <palmer@...ive.com>,
        Albert Ou <aou@...s.berkeley.edu>,
        Daniel Lezcano <daniel.lezcano@...aro.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Jason Cooper <jason@...edaemon.net>,
        Marc Zyngier <marc.zyngier@....com>,
        Atish Patra <atish.patra@....com>,
        linux-riscv@...ts.infradead.org,
        "linux-kernel@...r.kernel.org List" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v4 2/5] irqchip: sifive-plic: Don't inline plic_toggle()
 and plic_irq_toggle()

On Tue, Jan 15, 2019 at 9:24 PM Christoph Hellwig <hch@...radead.org> wrote:
>
> On Thu, Dec 27, 2018 at 04:48:18PM +0530, Anup Patel wrote:
> > The plic_toggle() uses raw_spin_lock() and plic_irq_toggle has a
> > for loop so both these functions are not suitable for being inline
> > hence this patch removes the inline keyword.
>
> That is a weird argument for a function which has by design exactly
> two callers and is in the hot path.  The alternative to the inline
> here would be to duplicate the code.

It's strange that you see it as weird argument. Both plic_toggle()
and plic_irq_toggle() are 5+ lines functions with loops. The loop
is clear in plic_irq_toggle() whereas raw_spin_lock() in plic_toggle()
expands into inline-assembly spin-loop because raw_spin_lock()
is a macro (not function).

Further looking at disassembly of both functions, these are 55+
instructions. I think we let GCC decide whether these functions
should be inlined or not rather than us explicitly making these
functions inline.

Regards,
Anup