Re: [PATCH] x86: Setup early console as early as possible

["H. Peter Anvin" <hpa@zytor.com>]

On 07/12/2010 01:57 PM, Bodo Eggert wrote:
> H. Peter Anvin <hpa@zytor.com> wrote:
>> On 07/12/2010 09:21 AM, Yinghai Lu wrote:
> 
>>> earlycon=uart8250 or console=uart8250 should be better than earlyprintk.
>>> because it is shared between different archs already.
> 
>> console= is crap, because it makes the user keep track of items they
>> should not have to deal with directly (unless they want to override the
>> defaults), i.e. the specific I/O ports used by the serial ports.
> 
> Why does console= require these parameters, if earlyprintk does not?
> 
> IMO you should fix console= to not need these parameters unless they cannot
> be guessed (in which case earlyprintk cannot guess them, either).
> 

earlycon is basically a low-level interface where you can steer
everything.  This is what you have to do on most embedded architectures,
where the firmware doesn't tell you where the console is located.

On x86, either the firmware can tell you or you can rely on 20 years of
legacy, with very few divergences.

	-hpa