Re: [PATCH 0/3] clk: let clock perform allocation in init

[Jerome Brunet]

On Tue 24 Sep 2019 at 14:39, Jerome Brunet <jbrunet@xxxxxxxxxxxx> wrote:

> This patchset is a follow up on this pinky swear [0].
> Its purpose is:
>  * Clarify the acceptable use of clk_ops init() callback
>  * Let the init() callback return an error code in case anything
>    fail.
>  * Add the terminate() counter part of of init() to release the
>    resources which may have been claimed in init()
>
> After discussing with Stephen at LPC, I decided to drop the 2 last patches
> of the RFC [1]. I can live without it for now and nobody expressed a
> critical need to get the proposed placeholder.
>
> [0]: https://lkml.kernel.org/r/CAEG3pNB-143Pr_xCTPj=tURhpiTiJqi61xfDGDVdU7zG5H-2tA@xxxxxxxxxxxxxx
> [1]: https://lkml.kernel.org/r/20190828102012.4493-1-jbrunet@xxxxxxxxxxxx
>

Hi Stephen,

Do you think we can fit this into the incoming cycle ?

Cheers
Jerome

> Jerome Brunet (3):
>   clk: actually call the clock init before any other callback of the
>     clock
>   clk: let init callback return an error code
>   clk: add terminate callback to clk_ops
>
>  drivers/clk/clk.c                     | 38 ++++++++++++++++++---------
>  drivers/clk/meson/clk-mpll.c          |  4 ++-
>  drivers/clk/meson/clk-phase.c         |  4 ++-
>  drivers/clk/meson/clk-pll.c           |  4 ++-
>  drivers/clk/meson/sclk-div.c          |  4 ++-
>  drivers/clk/microchip/clk-core.c      |  8 ++++--
>  drivers/clk/mmp/clk-frac.c            |  4 ++-
>  drivers/clk/mmp/clk-mix.c             |  4 ++-
>  drivers/clk/qcom/clk-hfpll.c          |  6 +++--
>  drivers/clk/rockchip/clk-pll.c        | 28 ++++++++++++--------
>  drivers/clk/ti/clock.h                |  2 +-
>  drivers/clk/ti/clockdomain.c          |  8 +++---
>  drivers/net/phy/mdio-mux-meson-g12a.c |  4 ++-
>  include/linux/clk-provider.h          | 13 ++++++---
>  14 files changed, 90 insertions(+), 41 deletions(-)