Registering RTC access methods in platform_init()

Registering RTC access methods in platform_init()

[David Müller (ELSOFT AG)]

Hello

I have a question regarding the registering of the RTC access methods in
platform_init() in ppc4xx_setup.c.
The problem is quite similar to the one i stated in
one of my previous e-mails "Additional IRQ handlers in ppc4xx_init_IRQ()".

Currently this is done in this way:

#ifdef CONFIG_PPC_RTC
/* new timer stuff */
	ppc_md.time_init = todc_time_init;
	ppc_md.set_rtc_time = todc_set_rtc_time;
	ppc_md.get_rtc_time = todc_get_rtc_time;

#if !defined(CONFIG_PIP405) && !defined(CONFIG_MIP405)
	ppc_md.nvram_read_val = todc_direct_read_val;
	ppc_md.nvram_write_val = todc_direct_write_val;
#else
	ppc_md.nvram_read_val = todc_mc146818_read_val;
	ppc_md.nvram_write_val = todc_mc146818_write_val;
#endif
#endif


All other ports using a MC146818 RTC have their own xxx_setup.c file
(Sandpoint, Adirondack, ...).

I suppose the preferred way to solve this is to replace this whole init
code by a call to a board specific init routine?




Dave


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Registering RTC access methods in platform_init()

[Armin Kuster]

"David Müller (ELSOFT AG)" wrote:
>
> Hello
>
>
> I suppose the preferred way to solve this is to replace this whole init
> code by a call to a board specific init routine?
>
> Dave
>

Yes.. its seems to be the way they are moving, I ended up doing it for
the walnut a few months back.

before you know it there will be no ppc4xx_setup.c

-- armin

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/