[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-Id: <74c253e6-1433-42b2-aabe-f6a85aeca099@app.fastmail.com>
Date: Thu, 27 Feb 2025 17:50:22 +0100
From: "Arnd Bergmann" <arnd@...db.de>
To: "Andy Shevchenko" <andriy.shevchenko@...ux.intel.com>,
"Arnd Bergmann" <arnd@...nel.org>
Cc: Thomas Weißschuh <thomas.weissschuh@...utronix.de>,
Linux-Arch <linux-arch@...r.kernel.org>,
"Andrew Morton" <akpm@...ux-foundation.org>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] asm-generic/io.h: rework split ioread64/iowrite64 helpers
On Thu, Feb 27, 2025, at 16:20, Andy Shevchenko wrote:
> On Thu, Feb 27, 2025 at 03:19:01PM +0100, Arnd Bergmann wrote:
>> From: Arnd Bergmann <arnd@...db.de>
>>
>> There are two incompatible sets of definitions of these eight functions:
>> On 64-bit architectures setting CONFIG_HAS_IOPORT, they turn into
>> either pair of 32-bit PIO (inl/outl) accesses or a single 64-bit MMIO
>> (readq/writeq). On other 64-bit architectures, they are always split
>> into 32-bit accesses.
>>
>> Depending on which header gets included in a driver, there are
>> additionally definitions for ioread64()/iowrite64() that are
>> expected to produce a 64-bit register MMIO access on all 64-bit
>> architectures.
>>
>> To separate the conflicting definitions, make the version in
>> include/linux/io-64-nonatomic-*.h visible on all architectures
>> but pick the one from lib/iomap.c on architectures that set
>> CONFIG_GENERIC_IOMAP in place of the default fallback.
>
> Thanks, this is good to go in my opinion,
> Acked-by: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Thanks, I've put it into my asm-generic tree now.
Arnd
Powered by blists - more mailing lists