Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20210226140305.26356-1-nsaenzjulienne@suse.de> <20210226140305.26356-10-nsaenzjulienne@suse.de>
In-Reply-To: <20210226140305.26356-10-nsaenzjulienne@suse.de>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Tue, 2 Mar 2021 10:32:06 +0100
Message-ID: <CAK8P3a1t3yCD4vXF803nL=n3Y3hD1MOPOAEZwZA+782N64PXTA@mail.gmail.com>
Subject: Re: [RFC 09/13] iommu/arm-smmu: Make use of dev_64bit_mmio_supported()
To:     Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Cc:     Linux ARM <linux-arm-kernel@lists.infradead.org>,
        DTML <devicetree@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Florian Fainelli <f.fainelli@gmail.com>,
        Rob Herring <robh+dt@kernel.org>,
        Robin Murphy <robin.murphy@arm.com>,
        Ard Biesheuvel <ardb@kernel.org>,
        Christoph Hellwig <hch@infradead.org>,
        Neil Armstrong <narmstrong@baylibre.com>,
        David Woodhouse <dwmw2@infradead.org>,
        Russell King - ARM Linux <linux@armlinux.org.uk>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Feb 26, 2021 at 3:03 PM Nicolas Saenz Julienne
<nsaenzjulienne@suse.de> wrote:

>         if (smmu->impl && unlikely(smmu->impl->write_reg))
>                 smmu->impl->write_reg(smmu, page, offset, val);
> -       else
> +       else if (dev_64bit_mmio_supported(smmu->dev))
>                 writel_relaxed(val, arm_smmu_page(smmu, page) + offset);
> +       else
> +               hi_lo_writeq_relaxed(val, arm_smmu_page(smmu, page) + offset);
>  }

This is a writel_relaxed(), not a writeq_relaxed(), so I suppose you don't
have to change it at all.

> +       else if (dev_64bit_mmio_supported(smmu->dev))
> +               return readq_relaxed(arm_smmu_page(smmu, page) + offset);
> +       else
> +               return hi_lo_readq_relaxed(arm_smmu_page(smmu, page) + offset);
> }


I see this pattern repeat across multiple drivers. I think Christoph
had originally
suggested folding the if/else logic into the writel_relaxed() that is defined in
include/linux/io-64-nonatomic-hi-lo.h, but of course that doesn't work if you
need to pass a device pointer.

It might still make sense to have another wrapper in that same file though,
something like

static inline hi_lo_writeq_relaxed_if_possible(struct device *dev, __u64 val,
                    volatile void __iomem *addr)
{
       if (dev_64bit_mmio_supported(smmu->dev)) {
              readq_relaxed(arm_smmu_page(smmu, page) + offset);
       } else {
               writel_relaxed(val >> 32, addr + 4);
               writel_relaxed(val, addr);
       }
}

         Arnd