[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAHp75VedQ_t0vo51QXNOKoGDdt4sp8nyreNkteHpu0=zKxO8rw@mail.gmail.com>
Date: Wed, 23 Aug 2017 00:07:57 +0300
From: Andy Shevchenko <andy.shevchenko@...il.com>
To: Logan Gunthorpe <logang@...tatee.com>
Cc: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Linux-Arch <linux-arch@...r.kernel.org>,
linux-ntb@...glegroups.com,
linux-crypto <linux-crypto@...r.kernel.org>,
Arnd Bergmann <arnd@...db.de>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Horia Geantă <horia.geanta@....com>,
Stephen Bates <sbates@...thlin.com>,
Christoph Hellwig <hch@....de>,
Alan Cox <gnomes@...rguk.ukuu.org.uk>
Subject: Re: [PATCH v7 5/7] io-64-nonatomic: add io{read|write}64[be]{_lo_hi|_hi_lo}
macros
On Tue, Aug 22, 2017 at 8:55 PM, Logan Gunthorpe <logang@...tatee.com> wrote:
> On 22/08/17 11:40 AM, Andy Shevchenko wrote:
>> Recomendation is kinda arguable. I doubt modern architectures make
>> difference between IO operations and MMIO.
>> Does, for example, PCI requires some special signal (message / wire)
>> handling when pio_*() accessors used vs. mmio_*() ones?
> Sort of. There is a difference between the two in x86.
Yes, the difference is in signaling. The question is if it's still in
use. Looks like yes...
> But really if it's
> not a valid recommendation someone should just go through and deprecate one
> of them or document better when one should be used over the other. In any
> case it's still really ugly to mix the two types in the same driver.
Mixing is ugly, totally agree.
--
With Best Regards,
Andy Shevchenko
Powered by blists - more mailing lists