Re: [PATCH] SH: add machine-ops and Dreamcast specific fix-up

[Adrian McMenamin <adrian@newgolddream.dyndns.info>]

On Sat, 2007-08-04 at 12:06 +0900, Paul Mundt wrote:
> On Fri, Aug 03, 2007 at 08:26:17PM +0100, Adrian McMenamin wrote:
> > diff --git a/arch/sh/boards/dreamcast/Makefile
> > b/arch/sh/boards/dreamcast/Makefile
> > index e6fcd3d..7b97546 100644
> > --- a/arch/sh/boards/dreamcast/Makefile
> > +++ b/arch/sh/boards/dreamcast/Makefile
> > @@ -2,5 +2,5 @@
> >  # Makefile for the Sega Dreamcast specific parts of the kernel
> >  #
> > 
> > -obj-y	 := setup.o irq.o rtc.o reboot.o
> > +obj-y	 := setup.o irq.o rtc.o
> > 
> 
> You've created this diff against a bogus kernel.

True. My apologies - I missed that.

> 
> > diff --git a/arch/sh/kernel/Makefile b/arch/sh/kernel/Makefile
> > index bdb30ba..7ab2359 100644
> > --- a/arch/sh/kernel/Makefile
> > +++ b/arch/sh/kernel/Makefile

....

> > 
> >  void machine_crash_shutdown(struct pt_regs *regs)
> >  {
> 
> You've missed machine_crash_shutdown().
> 

Apologies - though I am a bit puzzled as why the code compiled and
linked with the two definitions, but that's not much of an excuse.


> > +static void mach_reboot_fixups(void)
> > +{
> > +	if (mach_is_dreamcast()) {
> > +		writel(0x00007611, 0xA05F6890);
> > +	}
> > +}
> > +
> Whether it's only the dreamcast or not is irrelevant, why bother adding
> abstraction if you intend to add pointless hacks that completely
> side-steps it?
> 

I don't understand the point you are trying to make. Please explain with
more clarity. What have I completely side stepped? I have followed,
broadly, the same pattern used in i386. Just that there, afaics, they
pick up on various PCI cards as the basis on which to modify the reboot.


> > diff --git a/include/asm-sh/emergency-restart.h
> > b/include/asm-sh/emergency-restart.h
> > index 108d8c4..d6bec92 100644
> > --- a/include/asm-sh/emergency-restart.h
> > +++ b/include/asm-sh/emergency-restart.h
> > @@ -1,6 +1,6 @@
> > -#ifndef _ASM_EMERGENCY_RESTART_H
> > -#define _ASM_EMERGENCY_RESTART_H
> > +#ifndef _ASM_SH_EMERGENCY_RESTART_H
> > +#define _ASM_SH_EMERGENCY_RESTART_H
> > 
> > -#include <asm-generic/emergency-restart.h>
> > +extern void machine_emergency_restart(void);
> > 
> > -#endif /* _ASM_EMERGENCY_RESTART_H */
> > +#endif /* _ASM_SH_EMERGENCY_RESTART_H */
> > 
> Pointless. Separating out machine_emergency_restart() buys us nothing,
> leave this alone and just kill it off from the machine_ops entirely.
> You've also ignored my earlier mail where I suggested this and killing
> off some of the other ops we had no use for (as well as consolidating
> machine_crash_shutdown()). I do wish you would read these things and wait
> until there's been a resolution one way or another.

I haven't ignored it. It was just explained with your customary
clarity :)


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/