[PATCH 0/3] clk: keystone: syscon-clk: fixes for audio refclk

Matthias Schiffer posted 3 patches 2 years, 6 months ago
drivers/clk/keystone/syscon-clk.c | 73 ++++++++++++++++++++-----------
1 file changed, 48 insertions(+), 25 deletions(-)
[PATCH 0/3] clk: keystone: syscon-clk: fixes for audio refclk
Posted by Matthias Schiffer 2 years, 6 months ago
Currently the driver uses of_clk_hw_onecell_get() for all clocks it
manages, but this is incorrect for the audio refclk, which has 0 rather
than 1 clock cell according to its binding documentation [1]; attempting
to look up the clock when referenced like this in the Device Tree leads
to errors, as uninitialized memory is passed to of_clk_hw_onecell_get()
as the index.

The actual fix is in patch 3; patches 1 and 2 are preparation and
related cleanup. I've added a Fixes: tag to all 3 patches, as they
need to be backported together.

Matthias Schiffer (3):
  clk: keystone: syscon-clk: use struct instead of array for match data
  clk: keystone: syscon-clk: specify whether a parent is required in
    match data
  clk: keystone: syscon-clk: use of_clk_hw_simple_get() for audio refclk

[1] https://www.kernel.org/doc/Documentation/devicetree/bindings/clock/ti,am62-audio-refclk.yaml


 drivers/clk/keystone/syscon-clk.c | 73 ++++++++++++++++++++-----------
 1 file changed, 48 insertions(+), 25 deletions(-)

-- 
TQ-Systems GmbH | Mühlstraße 2, Gut Delling | 82229 Seefeld, Germany
Amtsgericht München, HRB 105018
Geschäftsführer: Detlef Schneider, Rüdiger Stahl, Stefan Schneider
https://www.tq-group.com/

Re: [PATCH 0/3] clk: keystone: syscon-clk: fixes for audio refclk
Posted by Francesco Dolcini 2 years, 6 months ago
On Tue, Aug 01, 2023 at 12:36:06PM +0200, Matthias Schiffer wrote:
> The actual fix is in patch 3; patches 1 and 2 are preparation and
> related cleanup. I've added a Fixes: tag to all 3 patches, as they
> need to be backported together.

Please see this https://lore.kernel.org/all/20230728222639.110409-1-francesco@dolcini.it/
that was sent a few days ago. It fixes the same issue and it's already
reviewed.

Francesco