On 01/02/2016 12:25, Matthias Brugger wrote:

On 27/01/16 12:29, John Crispin wrote:

[...]

quoted

+static int pwrap_mt2701_init_reg_clock(struct pmic_wrapper *wrp)
+{
+    switch (wrp->slave->type) {
+    case PMIC_MT6397:
+        pwrap_writel(wrp, 0xc, PWRAP_RDDMY);
+        pwrap_writel(wrp, 0x4, PWRAP_CSHEXT_WRITE);
+        pwrap_writel(wrp, 0x0, PWRAP_CSHEXT_READ);
+        pwrap_writel(wrp, 0x2, PWRAP_CSLEXT_START);
+        pwrap_writel(wrp, 0x2, PWRAP_CSLEXT_END);
+        break;
+
+    case PMIC_MT6323:
+        pwrap_writel(wrp, 0x8, PWRAP_RDDMY);
+        pwrap_dew_write(wrp, PWRAP_DEW_RDDMY_NO, 0x8);
+        pwrap_writel(wrp, 0x5, PWRAP_CSHEXT_WRITE);
+        pwrap_writel(wrp, 0x0, PWRAP_CSHEXT_READ);
+        pwrap_writel(wrp, 0x2, PWRAP_CSLEXT_START);
+        pwrap_writel(wrp, 0x2, PWRAP_CSLEXT_END);
+        break;
+    }
+
+    return 0;
+}
+

I forgot this:
So mt2701 can be used with both PMICs? Just wondering, as from what I
know, SoC and PMIC a tightly coupled.

yes, it can be run with old and new style pmics. this is why the MAN_CMD
SPI_W bit has changed. the old BIT(13) is used to tell the wrapper what
slave type it has.

there is a way to runtime set the pmic but there is still a detail i
need to look at so it is not part of this series

In general this series looks quite good to me.

i'll spin V5 during the day. busy adding dual gmac support to the
ethernet driver just now.

	John