Hi Mike,

On Wed, Jul 06, 2016 at 07:33:08PM -0700, Michael Turquette wrote:

Hi Maxime,

Quoting Maxime Ripard (2016-06-29 12:05:34)

quoted

+static void __init sun8i_h3_ccu_setup(struct device_node *node)
+{
+       void __iomem *reg;
+       u32 val;
+
+       reg = of_io_request_and_map(node, 0, of_node_full_name(node));
+       if (IS_ERR(reg)) {
+               pr_err("%s: Could not map the clock registers\n",
+                      of_node_full_name(node));
+               return;
+       }
+
+       /* Force the PLL-Audio-1x divider to 4 */
+       val = readl(reg + SUN8I_H3_PLL_AUDIO_REG);
+       val &= ~GENMASK(4, 0);
+       writel(val | 3, reg + SUN8I_H3_PLL_AUDIO_REG);
+
+       sunxi_ccu_probe(node, reg, &sun8i_h3_ccu_desc);
+}
+CLK_OF_DECLARE(sun8i_h3_ccu, "allwinner,sun8i-h3-ccu",
+              sun8i_h3_ccu_setup);

There are several examples of drivers that split the clocks between
"early" CLK_OF_DECLARE clocks and "late" module clocks. If you really
need early clocks (which is less likely on a 64-bit platform with
architected timers), it would be nice to pair that with a proper
platform_driver (using builtin_platform_driver most likely).

I think we discussed that already, but yeah, we do have timers that
are not the architected ones (and this is a ARMv7 platform). I have
the feeling that splitting the two doesn't really bring any benefit,
but complexify a lot the driver.

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20160708/23fdd4dc/attachment.sig>