RE: [PATCH 8/8] clk: ppc-corenet: Add support for the platform PLL
From: Yuantian Tang <hidden>
Date: 2015-01-21 09:37:55
Hello Emil,
From the RM and your code, I didn't see any difference between the two type=
of PLL. Could you provide some use cases or feature that prove this is necessary? If there did have some features that current function didn't contain, can w= e expend the current one to include it? BTW: if you did need this, please update the binding as well, if any: Documentation/devicetree/bindings/clock/qoriq-clock.txt Thanks, Yuantian
-----Original Message----- From: Emil Medve [mailto:Emilian.Medve@Freescale.com] Sent: Wednesday, January 21, 2015 5:03 PM To: Tang Yuantian-B29983; linuxppc-dev@lists.ozlabs.org; Wood Scott-B0742=
1;
mturquette@linaro.org; haokexin@gmail.com Subject: Re: [PATCH 8/8] clk: ppc-corenet: Add support for the platform P=
LL
=20 Hello Yuan-Tian, =20 =20 On 01/21/2015 02:35 AM, Tang Yuantian-B29983 wrote:quoted
Hi Emil, I don't think it is the best to add a function that is very similar to =
existing one.
quoted
If you think the function name is not appropriate, rename it.=20 It's not a naming matter. As I said, core_pll_init() assumptions and deci=
sions
based on the number of clocks in the DT. My hunch is some of these assump=
tions
are not necessary and/or should be explicit based on the node/device comp=
atible.
Having a standalone platform PLL initialization function wasn't a unilate=
ral lack of
foresight =20 =20 Cheers, =20 =20quoted
Thanks, Yuantianquoted
-----Original Message----- From: Emil Medve [mailto:Emilian.Medve@Freescale.com] Sent: Wednesday, January 21, 2015 4:20 PM To: Tang Yuantian-B29983; linuxppc-dev@lists.ozlabs.org; Wood Scott-B07421; mturquette@linaro.org; haokexin@gmail.com Subject: Re: [PATCH 8/8] clk: ppc-corenet: Add support for the platform PLL Hello Yuan-Tian, On 01/20/2015 11:42 PM, Tang Yuantian-B29983 wrote:quoted
Which platform are you trying to use this on?CoreNet chassis v1 and v2 SoC(s)quoted
Can this be initialized by core pll function core_pll_init()? I just saw most of this function is silimar to the core_pll_init().Yes, the flow is similar, but core_pll_init() makes assumptions that it shouldn't or are not relevant to the platform PLL Cheers,quoted
Thanks, Yuantianquoted
-----Original Message----- From: Emil Medve [mailto:Emilian.Medve@Freescale.com] Sent: Tuesday, January 20, 2015 6:10 PM To: linuxppc-dev@lists.ozlabs.org; Wood Scott-B07421; mturquette@linaro.org; haokexin@gmail.com; Tang Yuantian-B29983 Cc: Medve Emilian-EMMEDVE1 Subject: [PATCH 8/8] clk: ppc-corenet: Add support for the platform PLL Change-Id: Iac11ed95f274485a86d2c11f32a3dc502bcd020f Signed-off-by: Emil Medve <redacted> --- drivers/clk/clk-ppc-corenet.c | 85 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 85 insertions(+)diff --git a/drivers/clk/clk-ppc-corenet.cb/drivers/clk/clk-ppc-corenet.c index 91816b1..ff425e1 100644--- a/drivers/clk/clk-ppc-corenet.c +++ b/drivers/clk/clk-ppc-corenet.c@@ -7,6 +7,9 @@ * * clock driver for Freescale PowerPC corenet SoCs. */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + #include <linux/clk-provider.h> #include <linux/io.h> #include <linux/kernel.h>@@ -261,9 +264,91 @@ static void __init sysclk_init(structdevice_node*node)quoted
quoted
of_clk_add_provider(np, of_clk_src_simple_get, clk); } +static void __init pltfrm_pll_init(struct device_node *np) { + void __iomem *base; + uint32_t mult; + const char *parent_name, *clk_name; + int i, _errno; + struct clk_onecell_data *cod; + + base =3D of_iomap(np, 0); + if (!base) { + pr_err("%s(): %s: of_iomap() failed\n", __func__, np->name); + return; + } + + /* Get the multiple of PLL */ + mult =3D ioread32be(base); + + iounmap(base); + + /* Check if this PLL is disabled */ + if (mult & PLL_KILL) { + pr_debug("%s(): %s: Disabled\n", __func__, np->name); + return; + } + mult =3D (mult & GENMASK(6, 1)) >> 1; + + parent_name =3D of_clk_get_parent_name(np, 0); + if (!parent_name) { + pr_err("%s(): %s: of_clk_get_parent_name() failed\n", + __func__, np->name); + return; + } + + i =3D of_property_count_strings(np, "clock-output-names"); + if (i < 0) { + pr_err("%s(): %s: of_property_count_strings(clock-output-names) =3D %d\n", + __func__, np->name, i); + return; + } + + cod =3D kmalloc(sizeof(*cod) + i * sizeof(struct clk *), GFP_KERNE=
L);
quoted
quoted
quoted
quoted
+ if (!cod) + return; + cod->clks =3D (struct clk **)(cod + 1); + cod->clk_num =3D i; + + for (i =3D 0; i < cod->clk_num; i++) { + _errno =3D of_property_read_string_index(np,"clock-output-names",quoted
quoted
quoted
quoted
+ i, &clk_name); + if (_errno < 0) { + pr_err("%s(): %s: of_property_read_string_index(clock-output-names) =3D %d\n", + __func__, np->name, _errno); + goto return_clk_unregister; + } + + cod->clks[i] =3D clk_register_fixed_factor(NULL, clk_name, + parent_name, 0, mult, 1 + i); + if (IS_ERR(cod->clks[i])) { + pr_err("%s(): %s: clk_register_fixed_factor(%s) =3D %ld\n", + __func__, np->name, + clk_name, PTR_ERR(cod->clks[i])); + goto return_clk_unregister; + } + } + + _errno =3D of_clk_add_provider(np, of_clk_src_onecell_get, cod); + if (_errno < 0) { + pr_err("%s(): %s: of_clk_add_provider() =3D %d\n", + __func__, np->name, _errno); + goto return_clk_unregister; + } + + return; + +return_clk_unregister: + while (--i >=3D 0) + clk_unregister(cod->clks[i]); + kfree(cod); +} + CLK_OF_DECLARE(qoriq_sysclk_1, "fsl,qoriq-sysclk-1.0", sysclk_init); CLK_OF_DECLARE(qoriq_sysclk_2, "fsl,qoriq-sysclk-2.0", sysclk_init); CLK_OF_DECLARE(qoriq_core_pll_1, "fsl,qoriq-core-pll-1.0", core_pll_init); CLK_OF_DECLARE(qoriq_core_pll_2, "fsl,qoriq-core-pll-2.0", core_pll_init); CLK_OF_DECLARE(qoriq_core_mux_1, "fsl,qoriq-core-mux-1.0", core_mux_init); CLK_OF_DECLARE(qoriq_core_mux_2, "fsl,qoriq-core-mux-2.0", core_mux_init); +CLK_OF_DECLARE(qoriq_pltfrm_pll_1, "fsl,qoriq-platform-pll-1.0", +pltfrm_pll_init); CLK_OF_DECLARE(qoriq_pltfrm_pll_2, +"fsl,qoriq-platform-pll-2.0", pltfrm_pll_init); -- 2.2.2