Re: [PATCH tty v11 1/2] serial: 8250: Switch to nbcon console, take 2

[Jon Hunter <jonathanh@nvidia.com>]

Hi John,

On 25/08/2026 14:58, John Ogness wrote:
> Hi Jon,
> 
> On 2026-08-24, Jon Hunter <jonathanh@nvidia.com> wrote:
>> With this I see a little more output on the console, but it still
>> appears to hang later and never fully boots.
>>
>> Boot log here: https://pastebin.com/Xe1vmeUj
> 
> Thanks, this information is helpful. Below is another hack, to be used
> on the unmodified linux-next. For whatever reason, I think the queuing
> of irqwork is causing problems. Perhaps the irqwork is actually firing
> even though interrupts are supposed to be disabled here.
> 
> The following hack will disable the queuing during the cpuidle. Please
> give this a try:
> 
> ===== BEGIN CPUIDLE HACK =====
> diff --git a/drivers/cpuidle/cpuidle-tegra.c b/drivers/cpuidle/cpuidle-tegra.c
> index aca907a62bb5d..1dca9d6defbff 100644
> --- a/drivers/cpuidle/cpuidle-tegra.c
> +++ b/drivers/cpuidle/cpuidle-tegra.c
> @@ -226,6 +226,7 @@ static int tegra_cpuidle_adjust_state_index(int index, unsigned int cpu)
>   	return index;
>   }
>   
> +extern bool console_irqwork_blocked;
>   static __cpuidle int tegra_cpuidle_enter(struct cpuidle_device *dev,
>   					 struct cpuidle_driver *drv,
>   					 int index)
> @@ -238,6 +239,8 @@ static __cpuidle int tegra_cpuidle_enter(struct cpuidle_device *dev,
>   	if (dev->states_usage[index].disable)
>   		return -1;
>   
> +	console_irqwork_blocked = true;
> +
>   	if (index == TEGRA_C1) {
>   		if (do_rcu)
>   			ct_cpuidle_enter();
> @@ -256,6 +259,8 @@ static __cpuidle int tegra_cpuidle_enter(struct cpuidle_device *dev,
>   		index = ret;
>   	}
>   
> +	console_irqwork_blocked = false;
> +
>   	return index;
>   }
> ===== END CPUIDLE HACK =====
Yes that does boot. Let me know if there are any further tests I can run.

Thanks!
Jon

-- 
nvpublic