[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Linux/m68k config again (was Re: maintainer communication)



On Mon, 23 Dec 2013, Thorsten Glaser wrote:

> Finn Thain dixit:
> 
> > On Mon, 23 Dec 2013, Thorsten Glaser wrote:
> > 
> > > Finn Thain dixit:
> > > 
> > > > Why is CONFIG_SERIAL_PMACZILOG to be disabled?
> > > 
> > > See the discussion in the thread before this message.
> > 
> > I've seen no discussion of this on debian-68k or linux-m68k.
> 
> Oh right, that was a Cc-list conversation including Geert. Sorry.
> 

Well, the question remains, why?

BTW, the Mac LC III has no ADB driver. The same goes for certain other 
machines, i.e. the other models with Egret ADB chip as well as the early 
PowerBooks. On these machines the only available console is 
CONFIG_SERIAL_PMACZILOG.

> > > > And why was CONFIG_EARLY_PRINTK disabled?
> > >
> > > It was never enabled.
> >
> > Yes it was. See "Modernizing a Macintosh LC III": 
> > http://lists.debian.org/debian-68k/2013/11/msg00070.html
> 
> tg@ara5:~ $ fgrep EARLY /boot/config-3.10-2-m68k
> CONFIG_EARLY_PRINTK=y
> 
> Interesting. Hm. Must have been the Debian people, then, or a(n 
> unrelated) change to something CONFIG_EARLY_PRINTK depends on which in 
> turn deactivated the latter. I wasn't it.

It appears that I'm partly to blame for this. Sorry.

The default was changed to "n" in mainline 3.11, because the default "y" 
triggered a bug. The bug was subsequently fixed.

I was surprised to see a Debian kernel with CONFIG_EARLY_PRINTK=n (even 
after the change of default upstream) because I had thought that Debian 
explicitly enabled CONFIG_EARLY_PRINTK (perhaps I'm recalling x86? Or 
perhaps my memory is totally unreliable...)

Anyway, the upstream default is arguably now wrong, but because 
CONFIG_EARLY_PRINTK relates to the arch/m68k/kernel/head.S CONSOLE macro 
mess, that is a discussion for the upstream mailing list.

Finn


Reply to: