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: <20190617131733.2429469-1-arnd@arndb.de> <DM6PR12MB401006CA0446C71E36786E75DAE70@DM6PR12MB4010.namprd12.prod.outlook.com>
In-Reply-To: <DM6PR12MB401006CA0446C71E36786E75DAE70@DM6PR12MB4010.namprd12.prod.outlook.com>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Fri, 21 Jun 2019 11:05:11 +0200
Message-ID: <CAK8P3a3TE2hu=gjyf5gVj3HufVacFSoc4WjOVpgP3bz0LBbGKQ@mail.gmail.com>
Subject: Re: [PATCH] dmaengine: dw-edma: fix unnecessary stack usage
To:     Gustavo Pimentel <Gustavo.Pimentel@synopsys.com>
Cc:     Vinod Koul <vkoul@kernel.org>,
        Dan Williams <dan.j.williams@intel.com>,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        Russell King <rmk+kernel@armlinux.org.uk>,
        Joao Pinto <Joao.Pinto@synopsys.com>,
        "dmaengine@vger.kernel.org" <dmaengine@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, Jun 21, 2019 at 10:21 AM Gustavo Pimentel
<Gustavo.Pimentel@synopsys.com> wrote:

> On Mon, Jun 17, 2019 at 14:16:45, Arnd Bergmann <arnd@arndb.de> wrote:
>
> > Putting large constant data on the stack causes unnecessary overhead
> > and stack usage:
> >
> > drivers/dma/dw-edma/dw-edma-v0-debugfs.c:285:6: error: stack frame size of 1376 bytes in function 'dw_edma_v0_debugfs_on' [-Werror,-Wframe-larger-than=]
>
> I had that warning at the beginning of the development, that's why I
> divided the debugfs entries into several subfunctions. Perhaps my
> configuration has configured a bigger stack frame size than your
> configuration.

I suspect the problem is CONFIG_GCC_PLUGIN_STRUCTLEAK_VERBOSE,
which is a farily new configuration option that has the unintended side-effect
of sometimes preventing the compiler combining local variables from functions
inlined into the caller. Adding 'noinline_for_stack' to the sub-functions would
have also prevented this, but I think just not having the strings on the
stack is better anyway.

        Arnd