Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
From:   =?utf-8?q?Alvin_=C5=A0ipraga?= <alvin@pqrs.dk>
Subject: [PATCH v6 0/3] clk: si5351: add option to adjust PLL without
 glitches
Date:   Fri, 24 Nov 2023 14:17:41 +0100
Message-Id: <20231124-alvin-clk-si5351-no-pll-reset-v6-0-69b82311cb90@bang-olufsen.dk>
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 8bit
To:     Michael Turquette <mturquette@baylibre.com>,
        Stephen Boyd <sboyd@kernel.org>,
        Rob Herring <robh+dt@kernel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
        Conor Dooley <conor+dt@kernel.org>,
        =?utf-8?q?Alvin_=C5=A0ipraga?= <alsi@bang-olufsen.dk>
Cc:     Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>,
        Rabeeh Khoury <rabeeh@solid-run.com>,
        Jacob Siverskog <jacob@teenage.engineering>,
        Sergej Sawazki <sergej@taudac.com>, linux-clk@vger.kernel.org,
        devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
        Rob Herring <robh@kernel.org>
Precedence: bulk

This series intends to address a problem I had when using the Si5351A as
a runtime adjustable audio bit clock. The basic issue is that the driver
in its current form unconditionally resets the PLL whenever adjusting
its rate. But this reset causes an unwanted ~1.4 ms LOW signal glitch in
the clock output.

As a remedy, a new property is added to control the reset behaviour of
the PLLs more precisely. In the process I also converted the bindings to
YAML.

Changes:

v5 -> v6: 
- collect Rob's Reviewed-bys
- no further changes, resend only

v4 -> v5:
- address Rob's comments:
    - min/maxItems on top-level clocks:
    - remove unnecessary else:
    - remove spurious |

v3 -> v4:

- remove spurious | per Rob's suggestion
- simplify conditional clocks/clock-names per Rob's suggestion
- remove mention of clkout[0-7] still being admissible in the commit
  body of patch 1 - while the Linux driver still tolerates this, the
  new dt-bindings do not

v2 -> v3:

- address further comments from Rob:
  - drop unnecessary refs and minItems
  - simplify if conditions for chip variants
  - ignore his comment about dropping '|', as line would be >80 columns
  - move additionalProperties: false close to type: object
  - define clocks/clock-names at top-level
- drop patch to dove-cubox dts per Krzysztof's comment - will send
  separately
- collect Sebastian's Acked-by

v1 -> v2:

- address Rob's comments on the two dt-bindings patches
- new patch to correct the clock node names in the only upstream device
  tree using si5351

---
Alvin Šipraga (3):
      dt-bindings: clock: si5351: convert to yaml
      dt-bindings: clock: si5351: add PLL reset mode property
      clk: si5351: allow PLLs to be adjusted without reset

 .../devicetree/bindings/clock/silabs,si5351.txt    | 126 ----------
 .../devicetree/bindings/clock/silabs,si5351.yaml   | 265 +++++++++++++++++++++
 drivers/clk/clk-si5351.c                           |  47 +++-
 include/linux/platform_data/si5351.h               |   2 +
 4 files changed, 311 insertions(+), 129 deletions(-)
---
base-commit: f6abdcb2444f6ebe06e19cd9eee767c7c46612ae
change-id: 20231014-alvin-clk-si5351-no-pll-reset-ecfac0a6550c