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: <2e80d7bc-32a0-cc40-00a9-8a383a1966c2@huawei.com>
 <c1489f55-369d-2cff-ff36-b10fb5d3ee79@kernel.org> <8207cd51-5b94-2f15-de9f-d85c9c385bca@huawei.com>
 <6115fa56-a471-1e9f-edbb-e643fa4e7e11@kernel.org> <7c955142-1fcb-d99e-69e4-1e0d3d9eb8c3@huawei.com>
 <CAK8P3a0f9hnKGd6GJ8qFZSu+J-n4fY23TCGxQkmgJaxbpre50Q@mail.gmail.com>
 <90af535f-00ef-c1e3-ec20-aae2bd2a0d88@kernel.org> <CAK8P3a2Grd0JsBNsB19oAxrAFtOdpvjrpGcfeArKe7zD_jrUZw@mail.gmail.com>
 <ae0a1bf1-948f-7df0-9efb-cd1e94e27d2d@huawei.com>
In-Reply-To: <ae0a1bf1-948f-7df0-9efb-cd1e94e27d2d@huawei.com>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Fri, 6 Mar 2020 16:16:52 +0100
Message-ID: <CAK8P3a2wdCrBP=a8ZypWoC=HyCU3oYYNeCddWM7oT+xM9gTPhw@mail.gmail.com>
Subject: Re: About commit "io: change inX() to have their own IO barrier overrides"
To:     John Garry <john.garry@huawei.com>
Cc:     Sinan Kaya <okaya@kernel.org>, "xuwei (O)" <xuwei5@hisilicon.com>,
        Bjorn Helgaas <bhelgaas@google.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Jiaxun Yang <jiaxun.yang@flygoat.com>,
        linux-arch <linux-arch@vger.kernel.org>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        Will Deacon <will@kernel.org>,
        Catalin Marinas <catalin.marinas@arm.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, Mar 6, 2020 at 11:40 AM John Garry <john.garry@huawei.com> wrote:
> On 06/03/2020 07:54, Arnd Bergmann wrote:
> > On Fri, Mar 6, 2020 at 4:44 AM Sinan Kaya <okaya@kernel.org> wrote:
> -- a/lib/logic_pio.c
> +++ b/lib/logic_pio.c
> @@ -229,13 +229,21 @@ unsigned long
> logic_pio_trans_cpuaddr(resource_size_t addr)
>   }
>
>   #if defined(CONFIG_INDIRECT_PIO) && defined(PCI_IOBASE)
> +
> +#define logic_in_to_cpu_b(x) (x)
> +#define logic_in_to_cpu_w(x) __le16_to_cpu(x)
> +#define logic_in_to_cpu_l(x) __le32_to_cpu(x)
> +
>   #define BUILD_LOGIC_IO(bw, type)                                      \
>   type logic_in##bw(unsigned long addr)                                 \
>   {                                                                     \
>          type ret = (type)~0;                                           \
>                                                                         \
>          if (addr < MMIO_UPPER_LIMIT) {                                 \
> -               ret = read##bw(PCI_IOBASE + addr);                     \
> +               void __iomem *_addr = PCI_IOBASE + addr;               \
> +               __io_pbr();                                            \
> +               ret = logic_in_to_cpu_##bw(__raw_read##bw(_addr));     \
> +               __io_par(ret);                                         \
>          } else if (addr >= MMIO_UPPER_LIMIT && addr < IO_SPACE_LIMIT) {\
>                  struct logic_pio_hwaddr *entry = find_io_rang
>
> We could prob combine the le_to_cpu and __raw_read into a single macro.

What is the purpose of splitting out the byteswap rather than leaving the
open-coded rather than __le16_to_cpu(__raw_readw(PCI_IOBASE + addr))?

If this is needed to work across architectures, how about adding
an intermediate __raw_inw() etc in asm-generic/io.h like

#ifndef __raw_inw
#define __raw_inw(a) __raw_readw(PCI_IOBASE + addr));
#endif

#include <linux/logic_pio.h>

#ifndef inw
static inline u16 inw(unsigned long addr)
{
        u16 val;

        __io_pbr();
        val = __le16_to_cpu(__raw_inw(addr));
        __io_par(val);
        return val;
}
#endif

       Arnd