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;
MIME-Version: 1.0
References: <20190801231046.105022-1-nhuck@google.com> <01222982-4206-9925-0482-639a79384451@arm.com>
 <CAJkfWY6StuyMuKG7XdEJrqMsA_Xy02QZKp8r0K2jwSZwBCt+9Q@mail.gmail.com> <20190805133940.GA10425@arm.com>
In-Reply-To: <20190805133940.GA10425@arm.com>
From:   Nathan Huckleberry <nhuck@google.com>
Date:   Tue, 6 Aug 2019 14:29:16 -0700
Message-ID: <CAJkfWY5EL+MyRzSfcfJF2H8WoX73FEO0bOrwcoR4c4ekvaWvOQ@mail.gmail.com>
Subject: Re: [RFC PATCH] ARM: UNWINDER_FRAME_POINTER implementation for Clang
To:     Dave Martin <Dave.Martin@arm.com>
Cc:     Robin Murphy <robin.murphy@arm.com>,
        clang-built-linux <clang-built-linux@googlegroups.com>,
        Tri Vo <trong@google.com>, linux@armlinux.org.uk,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

I'm not sure that we should disable a broken feature instead of
attempting a fix.

CONFIG_FUNCTION_GRAPH_TRACER is dependent on CONFIG_FRAME_POINTER and
there have been reports by MediaTek that the frame pointer unwinder is
faster in some cases.

On Mon, Aug 5, 2019 at 6:39 AM Dave Martin <Dave.Martin@arm.com> wrote:
>
> On Fri, Aug 02, 2019 at 10:27:30AM -0700, Nathan Huckleberry wrote:
> > You're right. Would pushing an extra register be an adequate fix?
>
> Would forcing CONFIG_ARM_UNWIND=y for clang work as an alternative to
> this?
>
> Assuming clang supports -funwind-tables or equivalent, this may just
> work.
>
> [...]
>
> Cheers
> ---Dave