advice on using dev_info(), dev_err() and friends (was: [PATCH 2/9] isdn4linux: Siemens Gigaset drivers - common module)

[Tilman Schmidt <t.schmidt@phoenixsoftware.de>]

Hello Greg,

thank you for your comments. Just a few follow-up questions.

On 15.02.2006 04:27, Greg KH wrote:
> On Sat, Feb 11, 2006 at 03:52:27PM +0100, Hansjoerg Lipp wrote:
> 
>>--- linux-2.6.16-rc2/drivers/isdn/gigaset/gigaset.h	1970-01-01 01:00:00.000000000 +0100
>>+++ linux-2.6.16-rc2-gig/drivers/isdn/gigaset/gigaset.h	2006-02-11 15:20:26.000000000 +0100
[...]
>>+#undef info
>>+#define info(format, arg...) printk(KERN_INFO "%s: " format "\n", THIS_MODULE ? THIS_MODULE->name : "gigaset_hw" , ## arg)
> 
> Care to use the dev_info(), dev_err() and other dev_* friends instead of
> rolling your own?  It gives you a much easier and standardised way of
> identifying the driver and individual device that the message is
> happening for.

This turns out to be surprisingly tricky, as these macros take a device
pointer argument which mustn't be NULL either.

Could you please advise how to use these when I do not have a device
pointer available (ie. before the probe method has been called) or when
there is a risk of it being no longer valid (ie. the USB device has been
unplugged)? I can detect both cases by checking for dev==NULL, but then
what do I do if it is? Is there a dummy device structure somewhere which
I could then use instead? Somehow, replacing all our occurrences of
	info("m%sg", args);
with
	if (cs->dev)
		dev_info(cs->dev, "m%sg\n", args);
	else
		printk(KERN_INFO "gigaset: m%sg\n", args);
doesn't really appeal to me. :-)

Thanks
Tilman