hostap hangs when reseting..

Denis Vlasenko vda
Wed May 14 03:40:35 PDT 2003


On 14 May 2003 09:47, Denis Vlasenko wrote:
>         printk(KERN_WARNING "%s: %s: resetting card\n", dev_info, dev->name);
>         local->hw_resetting = 1;
>         hfa384x_disable_interrupts(dev);
>         if (local->func->cor_sreset)
>
> add a printk hunt?
>
> +printk(KERN_DEBUG "cor_sreset()...\n");
>                 local->func->cor_sreset(local);
> +printk(KERN_DEBUG "hw_shutdown()...\n");
>         prism2_hw_shutdown(dev, 1);
> +printk(KERN_DEBUG "hw_config()...\n");
>         prism2_hw_config(dev, 0);
>         local->hw_resetting = 0;
> +printk(KERN_DEBUG "reset done\n");
> }

I reviewed my logs. Seems like we indeed hang exactly after
printk(KERN_WARNING "%s: %s: resetting card...) thing.

Jouni, in which context do we call prism2_hw_reset()?
I am thinking of adding verbose logging here:

        local->hw_resetting = 1;
        hfa384x_disable_interrupts(dev);
        if (local->func->cor_sreset)
                local->func->cor_sreset(local);
        prism2_hw_shutdown(dev, 1);
        prism2_hw_config(dev, 0);
        local->hw_resetting = 0;

but since we deal with hard lockup, I'd rather like to make sure
printks will hit the disk - even if that means 5 sec delays
between cor_sreset / hw_shutdown / hw_config. Should I yield CPU
to another task? If yes, how?
--
vda




More information about the Hostap mailing list