On Thu, Dec 10, 2009 at 10:43 PM, Shilimkar, Santosh
<santosh.shilimkar@xxxxxx> wrote:
>> >
>> > +void __init omap4_check_revision(void)
>> > +{
>> > + u32 idcode;
>> > + u16 hawkeye;
>> > + u8 rev;
>> > + char *rev_name = "ES1.0";
>> should'nt you decide this inside the rev check path?
>
> This is initialized value by default. Since the chip is ES1.0 additionaly I am not doing it.
> e.g. for ES2.0, it would be done inside. > > +
ok, no strong opinions about this, but I think it might be nice if
ES2.0 addition will modify less code.
>> > + /*
>> > + * The IC rev detection is done with hawkeye and rev.
>> > + * Note that rev does not map directly to defined processor
>> > + * revision numbers as ES1.0 uses value 0.
>> > + */
>> > + idcode = read_tap_reg(OMAP_TAP_IDCODE);
>> > + hawkeye = (idcode >> 12) & 0xffff;
>> > + rev = (idcode >> 28) & 0xff;
>> > +
>> > + if ((hawkeye == 0xb852) && (rev == 0x0)) {
>> > + omap_revision = OMAP4430_REV_ES1_0;
>> > + pr_info("OMAP%04x %s\n", omap_rev() >> 16, rev_name);
>> wondering about the >>16 part.
> Why ?
omap_rev() >>16 means that you are using higher word of this and I
think we have CLASS() macro for that? if that is not enough, should'nt
we add adequate macro to do that?
[...]
Regards,
Nishanth Menon
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
