Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   <alexandru.tachici@analog.com>
To:     <linux-iio@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
        <devicetree@vger.kernel.org>
CC:     <jic23@kernel.org>, <robh+dt@kernel.org>
Subject: [PATCH v4 0/1] iio: adc: ad7124: allow more than 8 channels
Date:   Mon, 1 Mar 2021 18:15:25 +0200
Message-ID: <20210301161526.72884-1-alexandru.tachici@analog.com>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

From: Alexandru Tachici <alexandru.tachici@analog.com>

Currently AD7124-8 driver cannot use more than 8 IIO channels
because it was assigning the channel configurations bijectively
to channels specified in the device-tree. This is not possible
to do when using more than 8 channels as AD7124-8 has only 8
configuration registers.

All configurations are marked as live if they are
programmed on the device. Any change that happens from
userspace (sampling rate, filters etc.) will invalidate
them.

To allow the user to use all channels at once the driver
will keep in memory configurations for all channels but
will program only 8 of them at a time on the device.

If multiple channels have the same configuration, only
one configuration register will be used.

If there are more configurations needed than available registers
only the last 8 used configurations will be allowed to exist
on the device in a LRU fashion. (in case of raw reads).

If a read is requested on a channel whose configuration
is not programmed:
- check if there are similar configurations already programmed
	if yes: - point channel to that config
	if no:  - check if there are empty config slots
			- if yes: write config, push into queue of LRU configs
			- if no: pop one config, get it's config slot nr,
			write new config on the old slot, push new config
			in queue of LRU configs.

Alexandru Tachici (1):
  iio: adc: ad7124: allow more than 8 channels

Changelog v3 -> v4:
	- fixed multi-line comments
	- moved locking and unlocking out of the switch/case in write_raw
	- check if there were any actual changes in write_raw to the
	sample frequency/filter/scale by comparing old/new values
	before marking a config. as changed
	- maintain a bitmap in order to keep track of empty slots
	on the device

 drivers/iio/adc/ad7124.c | 470 ++++++++++++++++++++++++++-------------
 1 file changed, 313 insertions(+), 157 deletions(-)

-- 
2.20.1