Re: [PATCH v4 04/10] coresight: Appropriately disable programming clocks

[Leo Yan <leo.yan@arm.com>]

On Mon, Jul 21, 2025 at 12:40:38PM +0100, Suzuki Kuruppassery Poulose wrote:
> On 21/07/2025 11:48, Leo Yan wrote:
> > On Mon, Jul 21, 2025 at 10:15:22AM +0100, Suzuki Kuruppassery Poulose wrote:
> > 
> > [...]
> > 
> > > > diff --git a/include/linux/coresight.h b/include/linux/coresight.h
> > > > index 4ac65c68bbf44b98db22c3dad2d83a224ce5278e..dd2b4cc7a2b70cf060a3207548fe80e3824c489f 100644
> > > > --- a/include/linux/coresight.h
> > > > +++ b/include/linux/coresight.h
> > > > @@ -480,26 +480,16 @@ static inline bool is_coresight_device(void __iomem *base)
> > > >     * Returns:
> > > >     *
> > > >     * clk   - Clock is found and enabled
> > > > - * NULL  - clock is not found
> > > 
> > > This is still valid, right ?
> > 
> > No. Since this patch uses devm_clk_get_enabled() to get a clock, if the
> > pclk is not found, it returns -ENOENT (see of_parse_clkspec()).
> > 
> > Only the optional clock APIs (e.g., devm_clk_get_optional_enabled())
> > return a NULL pointer instead of -ENOENT when the clock is not found.
> 
> This will break ACPI based systems, as we may not have a "pclk" described
> for them. We should be able to tolerate "no pclk"

Get it. How about change the code like below?

  static inline struct clk *coresight_get_enable_apb_pclk(struct device *dev)
  {
      struct clk *pclk;

      pclk = devm_clk_get_enabled(dev, "apb_pclk");
      if (IS_ERR(pclk))
          pclk = devm_clk_get_enabled(dev, "apb");

      /* Tolerate no pclk for ACPI device */
      if ((pclk == ERR_PTR(-ENOENT)) && has_acpi_companion(dev))
          return NULL;

      return pclk;
  }