Re: [PATCH RESEND v5] arm64: dts: qcom: Add coresight nodes for x1e80100

[Johan Hovold <johan@kernel.org>]

On Tue, Feb 11, 2025 at 09:37:43AM +0800, Jie Gan wrote:
> On 2/10/2025 9:47 PM, Johan Hovold wrote:

> > This patch is now in 6.14-rc1 and results in the logs being spammed with
> > 200+ messages about cyclic dependencies. I'd expect something like this
> > to at least be mentioned in the commit message:
> > 
> > [    0.250732] /soc@0/stm@10002000: Fixed dependency cycle(s) with /soc@0/funnel@10041000
> > [    0.250854] /soc@0/tpda@10004000: Fixed dependency cycle(s) with /soc@0/funnel@10041000

> > ...

> > [    0.301582] /soc@0/funnel@10d13000: Fixed dependency cycle(s) with /soc@0/funnel@10d04000
> > [    0.301639] /soc@0/funnel@10d13000: Fixed dependency cycle(s) with /soc@0/tpda@10d12000
> > 
> > [    0.317624] Callback from call_rcu_tasks() invoked.
> > 
> > Has anyone looked into what is causing this and if anything can be done
> > about it?

> These lines have been printed by following API, 
> fw_devlink_create_devlink, it's pr_info:
> 
> ...
> 	device_links_write_lock();
> 	if (__fw_devlink_relax_cycles(link->consumer, sup_handle)) {
> 		__fwnode_link_cycle(link);
> 		pr_debug("----- cycle: end -----\n");
> 		pr_info("%pfwf: Fixed dependency cycle(s) with %pfwf\n",
> 			link->consumer, sup_handle);
> 	}
> ...

Right, and we had a few of these already, but not several hundreds. I
meant to ask if anyone has looked into whether this indicates a problem
with the patch or binding (or devlink code) or if these notifications
are simply expected.

> I am not sure we can change it to pr_debug to decrease annoying messages.

That may be an option, but I assume the current printk level is used so
that people get a chance to spot potential issues. Otherwise, we should
definitely silence it.

Johan