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
In-Reply-To: <20180823152143.GA3049@infradead.org>
References: <20180822185034.30167-1-jcmvbkbc@gmail.com> <20180823152143.GA3049@infradead.org>
From:   Max Filippov <jcmvbkbc@gmail.com>
Date:   Thu, 23 Aug 2018 11:31:02 -0700
Message-ID: <CAMo8BfL7WpM-o=Hp-sNRiG4ANFA9cSoTrKTRZYjeecNa+b3dFQ@mail.gmail.com>
Subject: Re: [PULL 00/14] Xtensa updates for 4.19
To:     Christoph Hellwig <hch@infradead.org>
Cc:     Linus Torvalds <torvalds@linux-foundation.org>,
        LKML <linux-kernel@vger.kernel.org>,
        linux-xtensa@linux-xtensa.org, Chris Zankel <chris@zankel.net>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Aug 23, 2018 at 8:21 AM, Christoph Hellwig <hch@infradead.org> wrote:
>>       xtensa: platform-specific handling of coherent memory
>
> How is this supposed to be used?  For the nommu version there only
> are __weak stubs, but no actual implementation.

The idea is that nommu xtensa platforms that have two views of the
physical memory may map one of them as cached and the other as
uncached. They will provide implementations for these four functions.

A platform with single view of the physical memory but with CPU that
has region translation still may create two views of the memory.

Both situations are platform-specific as we don't have fixed address
space layout for nommu systems.

I had an example here, but the addresses in this example are incorrect
as I misread the spec, so I didn't add it to the PR:
https://github.com/jcmvbkbc/linux-xtensa/commit/21ec04090f59647e69fb775ae9d447fcc778701e

-- 
Thanks.
-- Max