Re: Printing (Was: Re: Compiling problems with 2.4)

[Geert U. <ge...@li...>]

On Wed, 14 Feb 2001, Frank Rudolf Georg Petzold wrote:
> On Tue, Feb 13, 2001 at 04:24:59PM +0100, Geert Uytterhoeven wrote:
> > What happens if you replace the DPRINTK(...) by mb()?
> 
> I confirm that it somewhat works with debug enabled. Configuration:
> 604e/233, 70ns RAM, CV64/3D with 1280x1024-8, XFree 4.0.2 running.
> Somwhat means that the data is not entirely correct when arriving at the
> printer (I have a PostScript printer). Therefore I get strange results, but 
> most of it is still readable.
> 
> If I replace the printk with mb() it does not work at all.
> 
> > The fact that it works for some people could be attributed to the CPU (603e vs.
> > 604e).
> 
> It seems to be partially a tming problem, but not entirely. As I said, I once
> inserted delay loops that would only allow to write about 1 Bps to the CIA, but
> that did not help. There must be more to the dprintk call.

Does it work if you put a `mb()' in between any two accesses of CIA registers?

What is the definition of `mb()' in the APUS tree nowadays? Plain 2.4.2-pre3
has (include/asm-ppc/system.h):

| #define mb()  __asm__ __volatile__ ("sync" : : : "memory")
| #define rmb()  __asm__ __volatile__ ("sync" : : : "memory")
| #define wmb()  __asm__ __volatile__ ("eieio" : : : "memory")

Side question: does it work under Linux/m68k?

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li...

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds