Re: [PATCH] missing pci_disable_device()

[Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>]

Alan Cox wrote:

> On Iau, 2004-09-09 at 11:29, Kenji Kaneshige wrote:
>> > 	dev_warn(&pci_dev->dev, "Device was removed without properly "
>> > 				"calling pci_disable_device(), please fix.\n");
>> > 	WARN_ON(1);
>> > 
> 
> "This may need fixing" would be better than "please fix" as it may be
> a wrong warning
>

Yes.
I should have considered drivers that intentionally don't disable
devices. I'll change the message.

 
>> I changed my patch based on your feedback. But I have one
>> concern about putting "WARN_ON(1);". I'm worrying that people
>> might be surprised if stack dump is shown on their console,
>> especially if the broken driver handles many devices.
> 
> You could put
> 
> #ifdef CONFIG_DEBUG_KERNEL
> 
> #endif
> 
> around that section, then only users selecting kernel debugging would
> be bothered by it.
> 

Thank you for advice.

But I don't know if we should take this approach, because
'CONFIG_DEBUG_KERNEL' is set by default on RedHat and some
other distros.

How do you think?

Thanks,
Kenji Kaneshige