clk: imx6q: handle ENET PLL bypass
authorLucas Stach <l.stach@pengutronix.de>
Thu, 15 Nov 2018 14:30:28 +0000 (15:30 +0100)
committerStephen Boyd <sboyd@kernel.org>
Mon, 10 Dec 2018 19:39:16 +0000 (11:39 -0800)
The ENET PLL is different from the other i.MX6 PLLs, as it has
multiple outputs with different post-dividers, which are all
bypassed if the single bypass bit is activated. The hardware setup
looks something like this:
                                _
refclk-o---PLL---o----DIV1-----| \
       |         |             |M |----OUT1
       o-----------------------|_/
       |         |              _
       |         o----DIV2-----| \
       |         |             |M |----OUT2
       o-----------------------|_/
       |         |              _
       |         `----DIV3-----| \
       |                       |M |----OUT3
       `-----------------------|_/

The bypass bit not only bypasses the PLL, but also the attached
post-dividers. This would be reasonbly straight forward to model
with a single output, or with different bypass bits for each output,
but sadly the HW guys decided that it would be good to actuate all
3 muxes with a single bit.

So the need to have the PLL bypassed for one of the outputs always
affects 2 other (in our model) independent branches of the clock
tree.

This means the decision to bypass this PLL is a system wide design
choice and should not be changed on-the-fly, so we can treat any
bapass configuration as static. As such we can just register the
post-dividiers with a ratio that reflects the bypass status, which
allows us to bypass the PLL without breaking our abstraction model
and with it DT stability.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
drivers/clk/imx/clk-imx6q.c

index bd53c40..cd86a21 100644 (file)
@@ -225,6 +225,41 @@ static void of_assigned_ldb_sels(struct device_node *node,
        }
 }
 
+static bool pll6_bypassed(struct device_node *node)
+{
+       int index, ret, num_clocks;
+       struct of_phandle_args clkspec;
+
+       num_clocks = of_count_phandle_with_args(node, "assigned-clocks",
+                                               "#clock-cells");
+       if (num_clocks < 0)
+               return false;
+
+       for (index = 0; index < num_clocks; index++) {
+               ret = of_parse_phandle_with_args(node, "assigned-clocks",
+                                                "#clock-cells", index,
+                                                &clkspec);
+               if (ret < 0)
+                       return false;
+
+               if (clkspec.np == node &&
+                   clkspec.args[0] == IMX6QDL_PLL6_BYPASS)
+                       break;
+       }
+
+       /* PLL6 bypass is not part of the assigned clock list */
+       if (index == num_clocks)
+               return false;
+
+       ret = of_parse_phandle_with_args(node, "assigned-clock-parents",
+                                        "#clock-cells", index, &clkspec);
+
+       if (clkspec.args[0] != IMX6QDL_CLK_PLL6)
+               return true;
+
+       return false;
+}
+
 #define CCM_CCDR               0x04
 #define CCM_CCSR               0x0c
 #define CCM_CS2CDR             0x2c
@@ -503,16 +538,32 @@ static void __init imx6q_clocks_init(struct device_node *ccm_node)
        clk[IMX6QDL_CLK_USBPHY1_GATE] = imx_clk_gate("usbphy1_gate", "dummy", base + 0x10, 6);
        clk[IMX6QDL_CLK_USBPHY2_GATE] = imx_clk_gate("usbphy2_gate", "dummy", base + 0x20, 6);
 
-       clk[IMX6QDL_CLK_SATA_REF] = imx_clk_fixed_factor("sata_ref", "pll6_enet", 1, 5);
-       clk[IMX6QDL_CLK_PCIE_REF] = imx_clk_fixed_factor("pcie_ref", "pll6_enet", 1, 4);
+       /*
+        * The ENET PLL is special in that is has multiple outputs with
+        * different post-dividers that are all affected by the single bypass
+        * bit, so a single mux bit affects 3 independent branches of the clock
+        * tree. There is no good way to model this in the clock framework and
+        * dynamically changing the bypass bit, will yield unexpected results.
+        * So we treat any configuration that bypasses the ENET PLL as
+        * essentially static with the divider ratios reflecting the bypass
+        * status.
+        *
+        */
+       if (!pll6_bypassed(ccm_node)) {
+               clk[IMX6QDL_CLK_SATA_REF] = imx_clk_fixed_factor("sata_ref", "pll6_enet", 1, 5);
+               clk[IMX6QDL_CLK_PCIE_REF] = imx_clk_fixed_factor("pcie_ref", "pll6_enet", 1, 4);
+               clk[IMX6QDL_CLK_ENET_REF] = clk_register_divider_table(NULL, "enet_ref", "pll6_enet", 0,
+                                               base + 0xe0, 0, 2, 0, clk_enet_ref_table,
+                                               &imx_ccm_lock);
+       } else {
+               clk[IMX6QDL_CLK_SATA_REF] = imx_clk_fixed_factor("sata_ref", "pll6_enet", 1, 1);
+               clk[IMX6QDL_CLK_PCIE_REF] = imx_clk_fixed_factor("pcie_ref", "pll6_enet", 1, 1);
+               clk[IMX6QDL_CLK_ENET_REF] = imx_clk_fixed_factor("enet_ref", "pll6_enet", 1, 1);
+       }
 
        clk[IMX6QDL_CLK_SATA_REF_100M] = imx_clk_gate("sata_ref_100m", "sata_ref", base + 0xe0, 20);
        clk[IMX6QDL_CLK_PCIE_REF_125M] = imx_clk_gate("pcie_ref_125m", "pcie_ref", base + 0xe0, 19);
 
-       clk[IMX6QDL_CLK_ENET_REF] = clk_register_divider_table(NULL, "enet_ref", "pll6_enet", 0,
-                       base + 0xe0, 0, 2, 0, clk_enet_ref_table,
-                       &imx_ccm_lock);
-
        clk[IMX6QDL_CLK_LVDS1_SEL] = imx_clk_mux("lvds1_sel", base + 0x160, 0, 5, lvds_sels, ARRAY_SIZE(lvds_sels));
        clk[IMX6QDL_CLK_LVDS2_SEL] = imx_clk_mux("lvds2_sel", base + 0x160, 5, 5, lvds_sels, ARRAY_SIZE(lvds_sels));