On Tue, Jan 03, 2012 at 08:54:17PM +0900, Kukjin Kim wrote:
> Russell King - ARM Linux wrote:
> >
> > On Tue, Jan 03, 2012 at 08:01:15PM +0900, Kukjin Kim wrote:
> > > If any problems, please kindly let me know.
> >
> > Right, this looks better. I'm now left with one remaining bit:
> >
> > arch/arm/mach-exynos/include/mach/system.h:#include <plat/system-reset.h>
> > arch/arm/plat-s3c24xx/cpu.c:#include <plat/system-reset.h>
> > arch/arm/mach-s3c64xx/include/mach/system.h:#include <plat/system-reset.h>
> > arch/arm/mach-s3c2410/include/mach/system.h:#include <plat/system-reset.h>
> >
>
> And,
>
> arch/arm/mach-s5p64x0/include/mach/system.h:#include <plat/system-reset.h>
> arch/arm/mach-s5pc100/include/mach/system.h:#include <plat/system-reset.h>
> arch/arm/mach-s5pv210/include/mach/system.h:#include <plat/system-reset.h>
>
> > and:
> >
> > arch/arm/plat-samsung/include/plat/system-reset.h:
> > ...
> > #include <plat/watchdog-reset.h>
> >
> > static void arch_reset(char mode, const char *cmd)
> > {
> > arch_wdt_reset();
> > }
> >
> > I assume that with all the patches I now have merged, arch_reset()
> > should never be called on any Samsung platform, and so the include of
> > plat/watchdog-reset.h and call of arch_wdt_reset() can be removed in
> > my "ARM: restart: plat-samsung: remove plat/reset.h and s5p_reset_hook"
> > patch?
> >
> Yes, if we don't need arch_reset() anymore, we can remove them in your
> patch.
>
> But I'm not sure, because happened following build error. Russell, don't we
> need arch_reset()?
>
> arch/arm/kernel/process.c: In function 'arm_machine_restart':
> arch/arm/kernel/process.c:127: error: implicit declaration of function
> 'arch_reset'
>
> > What about arch/arm/plat-s3c24xx/cpu.c's include of system-reset.h?
> > Does that need to be replaced with watchdog-reset.h?
> >
> > That then leads to "ARM: restart: remove the now empty arch_reset()"
> > removing system-reset.h and all includes of that file?
>
> Thanks.
Oh god, I see what you've done.
ARM: 7256/1: restart: S3C64XX: use new restart hook
--- a/arch/arm/mach-s3c64xx/include/mach/system.h
+++ b/arch/arm/mach-s3c64xx/include/mach/system.h
...
-#include <plat/watchdog-reset.h>
+#include <plat/system-reset.h>
...
-static void arch_reset(char mode, const char *cmd)
-{
- if (mode != 's')
- arch_wdt_reset();
-
- /* if all else fails, or mode was for soft, jump to 0 */
- soft_restart(0);
-}
So we're making everything depend on plat/system-reset.h. Why? The
end result is to remove arch_reset() entirely, and making it depend
on some common definition in some shared header file makes things
more complicated and is error prone. How do I know when everything
is fixed up as far as the shared header file goes?
It would be _far_ better if the s3c64xx changes to add the .restart
method to _all_ s3c64xx platforms also removed the _contents_ and
only the _contents_ of arch_reset(), leaving an empty function there.
--
To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
