Feedback-ID: i56a14606:Fastmail
User-Agent: Cyrus-JMAP/3.9.0-alpha0-221-gec32977366-fm-20230306.001-gec329773
Mime-Version: 1.0
Message-Id: <2f07c421-7f9c-4e53-8bd5-46be07bb5c89@app.fastmail.com>
In-Reply-To: <9a1c49b0-2271-53c7-7f48-039f83d39e82@opensource.wdc.com>
References: <20230314121216.413434-1-schnelle@linux.ibm.com>
 <20230314121216.413434-3-schnelle@linux.ibm.com>
 <CAMuHMdVry2YViJ5oFgo9i+uStWbhy7mXKWdWvCX=qgAu1-_Y1w@mail.gmail.com>
 <c7315ca2-3ebf-7f3b-da64-9a74a995b0ae@opensource.wdc.com>
 <CAMuHMdVajEYsw8HrKw0GwV+09gbtkhjVMuKZ6RSBvq6got=jAg@mail.gmail.com>
 <9a1c49b0-2271-53c7-7f48-039f83d39e82@opensource.wdc.com>
Date:   Thu, 16 Mar 2023 16:21:58 +0100
From:   "Arnd Bergmann" <arnd@arndb.de>
To:     "Damien Le Moal" <damien.lemoal@opensource.wdc.com>,
        "Geert Uytterhoeven" <geert@linux-m68k.org>
Cc:     "Niklas Schnelle" <schnelle@linux.ibm.com>,
        "Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
        "Bjorn Helgaas" <bhelgaas@google.com>,
        =?UTF-8?Q?Uwe_Kleine-K=C3=B6nig?= <u.kleine-koenig@pengutronix.de>,
        "Mauro Carvalho Chehab" <mchehab@kernel.org>,
        "Alan Stern" <stern@rowland.harvard.edu>,
        "Rafael J . Wysocki" <rafael@kernel.org>,
        "Paul Walmsley" <paul.walmsley@sifive.com>,
        "Palmer Dabbelt" <palmer@dabbelt.com>,
        "Albert Ou" <aou@eecs.berkeley.edu>, linux-kernel@vger.kernel.org,
        Linux-Arch <linux-arch@vger.kernel.org>,
        linux-pci@vger.kernel.org, "Arnd Bergmann" <arnd@kernel.org>,
        linux-ide@vger.kernel.org
Subject: Re: [PATCH v3 02/38] ata: add HAS_IOPORT dependencies
Content-Type: text/plain
Precedence: bulk

On Thu, Mar 16, 2023, at 00:57, Damien Le Moal wrote:
> On 2023/03/15 20:36, Geert Uytterhoeven wrote:

> Ah. OK. I see now. So indeed, applying the dependency on the entire ATA_SFF
> group of drivers is very coarse.


> Can you change this to apply the dependency per driver ?

I think that will fail to build because of this function
on architectures that drop their non-functional
inb/outb helpers:

int ata_pci_bmdma_clear_simplex(struct pci_dev *pdev)
{
        unsigned long bmdma = pci_resource_start(pdev, 4);
        u8 simplex;
 
        if (bmdma == 0)
                return -ENOENT;
 
        simplex = inb(bmdma + 0x02); 
        outb(simplex & 0x60, bmdma + 0x02);
        simplex = inb(bmdma + 0x02);
        if (simplex & 0x80)
                return -EOPNOTSUPP;
        return 0;
}

This is only called from five pata drivers (ali, amd,
cmd64x, netcell, serverworks), so an easy workaround
would be to make sure those depend on HAS_IOPORT
and enclose the function definition in an #ifdef.

        Arnd