Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Sun, 4 Sep 2022 17:29:00 +0100
From:   Jonathan Cameron <jic23@kernel.org>
To:     Vincent Whitchurch <vincent.whitchurch@axis.com>
Cc:     kernel <kernel@axis.com>, "lars@metafoo.de" <lars@metafoo.de>,
        Axel Jonsson <Axel.Jonsson@axis.com>,
        "linux-iio@vger.kernel.org" <linux-iio@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 2/2] iio: adc: mcp320x: add triggered buffer support
Message-ID: <20220904172900.099ed1d0@jic23-huawei>
In-Reply-To: <Yw8SJil96IT6fP0n@axis.com>
References: <20220824104002.2749075-1-vincent.whitchurch@axis.com>
        <20220824104002.2749075-3-vincent.whitchurch@axis.com>
        <20220828182430.533bf8e0@jic23-huawei>
        <Yw8SJil96IT6fP0n@axis.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Wed, 31 Aug 2022 09:47:50 +0200
Vincent Whitchurch <vincent.whitchurch@axis.com> wrote:

> On Sun, Aug 28, 2022 at 07:24:30PM +0200, Jonathan Cameron wrote:
> > On Wed, 24 Aug 2022 12:40:02 +0200
> > Vincent Whitchurch <vincent.whitchurch@axis.com> wrote:  
> > > Add support for triggered buffers.  Just read the channels in a loop to
> > > keep things simple.  
> > 
> > Just curious, but what other options are there?  A quick datasheet scroll
> > through didn't seem to suggest you can overlap setup of next read with
> > reading out the previous (common on SPI ADCs).  
> 
> You're right that the hardware doesn't support any special method to
> read out multiple channels; I can mention that in the commit message.
> But I think you could construct a spi_message which a bunch of
> spi_transfers which toggle the CS appropriately between them to read out
> multiple channels in one go?  (Note that the variants have different
> data formats, and many of the variants only have one channel.)

That might be a little more efficient on a suitably advanced
SPI controller, or where the overhead of jumping between drivers
is large, but in many cases it won't save that much.

Would need an experiment on a platform someone cares about to decide
if it is worthwhile.

Jonathan