Messages in this thread | | | From | Andy Shevchenko <> | Date | Mon, 8 Aug 2022 15:41:48 +0200 | Subject | Re: [PATCH v1 1/5] regmap: mmio: Don't unprepare attached clock |
| |
On Mon, Aug 8, 2022 at 3:19 PM Mark Brown <broonie@kernel.org> wrote: > > On Fri, Aug 05, 2022 at 11:53:17PM +0300, Andy Shevchenko wrote: > > The commit eb4a219d19fd ("regmap: Skip clk_put for attached clocks when > > freeing context") oexcluded clk_put() call on regmap freeing. But the > > same is needed for clk_unprepare() since the regmap MMIO users should > > call regmap_mmio_detach_clk() which does unprepare the clock. Update > > the code accordingly, so neither clk_put() nor clk_unprepare() would > > be called for the attached clock. > > regmap_mmio_attach_clk() prepares the clock that's passed in, we should > undo that when detaching otherwise we're leaking a prepare (as we do in > the explicit detach).
Why do we allow the user to avoid explicit detach? What is the point of having that API in the case we take care of it?
-- With Best Regards, Andy Shevchenko
| |