The clk framework on bootup will automatically disable all unused clocks
on bootup unless the clk_ignore_unused kernel parameter is present.
Let's add a basic debugging log statement here that shows which clocks
are disabled. There is already tracepoint present here as well, but
there's nothing like a simple, good ol' fashioned printk for simplicity.
Signed-off-by: Brian Masney <[email protected]>
---
drivers/clk/clk.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
index 57b83665e5c3..ddf5a48e72b6 100644
--- a/drivers/clk/clk.c
+++ b/drivers/clk/clk.c
@@ -1352,6 +1352,7 @@ static void __init clk_disable_unused_subtree(struct clk_core *core)
* back to .disable
*/
if (clk_core_is_enabled(core)) {
+ dev_dbg(core->dev, "Powering off unused clock %s\n", core->name);
trace_clk_disable(core);
if (core->ops->disable_unused)
core->ops->disable_unused(core->hw);
--
2.38.1
Quoting Brian Masney (2022-11-17 02:58:29)
> The clk framework on bootup will automatically disable all unused clocks
> on bootup unless the clk_ignore_unused kernel parameter is present.
> Let's add a basic debugging log statement here that shows which clocks
> are disabled. There is already tracepoint present here as well, but
> there's nothing like a simple, good ol' fashioned printk for simplicity.
>
> Signed-off-by: Brian Masney <[email protected]>
> ---
I'd like to see a documentation update instead that covers how to enable
the tracepoint on the kernel commandline and have it print to the serial
console.
On Wed, Mar 29, 2023 at 02:49:50PM -0700, Stephen Boyd wrote:
> Quoting Brian Masney (2022-11-17 02:58:29)
> > The clk framework on bootup will automatically disable all unused clocks
> > on bootup unless the clk_ignore_unused kernel parameter is present.
> > Let's add a basic debugging log statement here that shows which clocks
> > are disabled. There is already tracepoint present here as well, but
> > there's nothing like a simple, good ol' fashioned printk for simplicity.
> >
> > Signed-off-by: Brian Masney <[email protected]>
> > ---
>
> I'd like to see a documentation update instead that covers how to enable
> the tracepoint on the kernel commandline and have it print to the serial
> console.
Sure, I can do that. I see there's a section 'Disabling clock gating of
unused clocks' in Documentation/driver-api/clk.rst where I think this
would be appropriate.
Brian
On Thu, Mar 30, 2023 at 10:05:37AM -0400, Brian Masney wrote:
> On Wed, Mar 29, 2023 at 02:49:50PM -0700, Stephen Boyd wrote:
> > Quoting Brian Masney (2022-11-17 02:58:29)
> > > The clk framework on bootup will automatically disable all unused clocks
> > > on bootup unless the clk_ignore_unused kernel parameter is present.
> > > Let's add a basic debugging log statement here that shows which clocks
> > > are disabled. There is already tracepoint present here as well, but
> > > there's nothing like a simple, good ol' fashioned printk for simplicity.
> > >
> > > Signed-off-by: Brian Masney <[email protected]>
> > > ---
> >
> > I'd like to see a documentation update instead that covers how to enable
> > the tracepoint on the kernel commandline and have it print to the serial
> > console.
>
> Sure, I can do that. I see there's a section 'Disabling clock gating of
> unused clocks' in Documentation/driver-api/clk.rst where I think this
> would be appropriate.
Just to close out this thread, I submitted a patch to the docs tree.
https://lore.kernel.org/lkml/[email protected]/T/#u
Brian