[U-Boot] [PATCH 06/10] sunxi: Add a33 dram init code

[Ian Campbell]

On Thu, 2015-04-16 at 09:27 +0200, Hans de Goede wrote:
> Hi,
> 
> On 15-04-15 21:56, Ian Campbell wrote:
> > On Tue, 2015-04-14 at 18:06 +0200, Hans de Goede wrote:
> >> From: Vishnu Patekar <vishnupatekar0510 at gmail.com>
> >>
> >> Based on Allwinner dram init code from the a33 bsp:
> >> https://github.com/allwinner-zh/bootloader/blob/master/basic_loader/bsp/bsp_for_a33/init_dram/mctl_hal.c
> >>
> >> Initial u-boot port by Vishnu Patekar, major cleanup / rewrite by
> >> Hans de Goede.
> >>
> >> Signed-off-by: Vishnu Patekar <vishnupatekar0510 at gmail.com>
> >> Signed-off-by: Hans de Goede <hdegoede at redhat.com>
> >
> >> +	/* Set dram timing */
> >> +	reg_val = (twtp << 24) | (tfaw << 16) | (trasmax << 8) | (tras << 0);
> >> +	writel(reg_val, &mctl_ctl->dramtmg0);
> >> +	reg_val = (txp << 16) | (trtp << 8) | (trc << 0);
> >> +	writel(reg_val, &mctl_ctl->dramtmg1);
> >> +	reg_val = (tcwl << 24) | (tcl << 16) | (trd2wr << 8) | (twr2rd << 0);
> >> +	writel(reg_val, &mctl_ctl->dramtmg2);
> >> +	reg_val = (tmrw << 16) | (tmrd << 12) | (tmod << 0);
> >> +	writel(reg_val, &mctl_ctl->dramtmg3);
> >> +	reg_val = (trcd << 24) | (tccd << 16) | (trrd << 8) | (trp << 0);
> >> +	writel(reg_val, &mctl_ctl->dramtmg4);
> >> +	reg_val = (tcksrx << 24) | (tcksre << 16) | (tckesr << 8) | (tcke << 0);
> >> +	writel(reg_val, &mctl_ctl->dramtmg5);
> >
> > There's a lot of magic numbers here (and in the following code),
> > although in this particular context (with the named var) unless they are
> > the same elsewhere I'm not sure #defines would improve things much, but
> > I think some of the other stuff likely would.
> >
> > Assuming you have any idea what the bits are, I suppose that per usual
> > we don't really know because -ENODOC?
> 
> Right the problem here is -ENODOC, the magic values come from the allwinnner
> code and in the cases where we do not have named variables (as we do in
> the above blurb) we've no clue what we're doing really, so I think adding
> defines there will only obfuscate things.
> 
> Can you give a short description of the cases where you believe that
> adding defines would be a good idea ?

Anywhere where we know the meanings ;-) If that's nowhere then the magic
numbers are fine (which is what my final paragraph was trying to say,
but not very clearly).

Ian.