* [PATCH v1 0/2] Add support for display port clocks and clock ops
@ 2018-10-09 13:57 Taniya Das
2018-10-09 13:57 ` [PATCH v1 1/2] clk: qcom: rcg2: Add support for display port " Taniya Das
2018-10-09 13:57 ` [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks Taniya Das
0 siblings, 2 replies; 16+ messages in thread
From: Taniya Das @ 2018-10-09 13:57 UTC (permalink / raw)
To: Stephen Boyd, Michael Turquette
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu, Taniya Das
New display port clock ops supported for display port clocks.
Also add support for the display port related branches and RCGs.
Taniya Das (2):
clk: qcom: rcg2: Add support for display port clock ops
clk: qcom : dispcc: Add support for display port clocks
drivers/clk/qcom/clk-rcg.h | 1 +
drivers/clk/qcom/clk-rcg2.c | 86 +++++++++
drivers/clk/qcom/dispcc-sdm845.c | 232 +++++++++++++++++++++++++
include/dt-bindings/clock/qcom,dispcc-sdm845.h | 11 ++
4 files changed, 330 insertions(+)
--
Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member
of the Code Aurora Forum, hosted by the Linux Foundation.
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v1 1/2] clk: qcom: rcg2: Add support for display port clock ops
2018-10-09 13:57 [PATCH v1 0/2] Add support for display port clocks and clock ops Taniya Das
@ 2018-10-09 13:57 ` Taniya Das
2018-10-09 20:46 ` Stephen Boyd
2018-10-09 13:57 ` [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks Taniya Das
1 sibling, 1 reply; 16+ messages in thread
From: Taniya Das @ 2018-10-09 13:57 UTC (permalink / raw)
To: Stephen Boyd, Michael Turquette
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu, Taniya Das
New display port clock ops supported for display port clocks.
Signed-off-by: Taniya Das <tdas@codeaurora.org>
---
drivers/clk/qcom/clk-rcg.h | 1 +
drivers/clk/qcom/clk-rcg2.c | 86 +++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 87 insertions(+)
diff --git a/drivers/clk/qcom/clk-rcg.h b/drivers/clk/qcom/clk-rcg.h
index e5eca8a..e0db0f7 100644
--- a/drivers/clk/qcom/clk-rcg.h
+++ b/drivers/clk/qcom/clk-rcg.h
@@ -162,6 +162,7 @@ struct clk_rcg2 {
extern const struct clk_ops clk_pixel_ops;
extern const struct clk_ops clk_gfx3d_ops;
extern const struct clk_ops clk_rcg2_shared_ops;
+extern const struct clk_ops clk_dp_ops;
struct clk_rcg_dfs_data {
struct clk_rcg2 *rcg;
diff --git a/drivers/clk/qcom/clk-rcg2.c b/drivers/clk/qcom/clk-rcg2.c
index 6e3bd19..ca6142f 100644
--- a/drivers/clk/qcom/clk-rcg2.c
+++ b/drivers/clk/qcom/clk-rcg2.c
@@ -10,6 +10,7 @@
#include <linux/export.h>
#include <linux/clk-provider.h>
#include <linux/delay.h>
+#include <linux/rational.h>
#include <linux/regmap.h>
#include <linux/math64.h>
#include <linux/slab.h>
@@ -1124,3 +1125,88 @@ int qcom_cc_register_rcg_dfs(struct regmap *regmap,
return 0;
}
EXPORT_SYMBOL_GPL(qcom_cc_register_rcg_dfs);
+
+static int clk_rcg2_dp_set_rate(struct clk_hw *hw, unsigned long rate,
+ unsigned long parent_rate)
+{
+ struct clk_rcg2 *rcg = to_clk_rcg2(hw);
+ struct freq_tbl f = { 0 };
+ unsigned long src_rate;
+ unsigned long num, den;
+ u32 mask = BIT(rcg->hid_width) - 1;
+ u32 hid_div, cfg;
+ int i, num_parents = clk_hw_get_num_parents(hw);
+
+ src_rate = clk_hw_get_rate(clk_hw_get_parent(hw));
+ if (src_rate <= 0) {
+ pr_err("Invalid RCG parent rate\n");
+ return -EINVAL;
+ }
+
+ rational_best_approximation(src_rate, rate,
+ (unsigned long)(1 << 16) - 1,
+ (unsigned long)(1 << 16) - 1, &den, &num);
+
+ if (!num || !den) {
+ pr_err("Invalid MN values derived for requested rate %lu\n",
+ rate);
+ return -EINVAL;
+ }
+
+ regmap_read(rcg->clkr.regmap, rcg->cmd_rcgr + CFG_REG, &cfg);
+ hid_div = cfg;
+ cfg &= CFG_SRC_SEL_MASK;
+ cfg >>= CFG_SRC_SEL_SHIFT;
+
+ for (i = 0; i < num_parents; i++)
+ if (cfg == rcg->parent_map[i].cfg) {
+ f.src = rcg->parent_map[i].src;
+ break;
+ }
+
+ f.pre_div = hid_div;
+ f.pre_div >>= CFG_SRC_DIV_SHIFT;
+ f.pre_div &= mask;
+
+ if (num == den) {
+ f.m = 0;
+ f.n = 0;
+ } else {
+ f.m = num;
+ f.n = den;
+ }
+
+ return clk_rcg2_configure(rcg, &f);
+}
+
+static int clk_rcg2_dp_set_rate_and_parent(struct clk_hw *hw,
+ unsigned long rate, unsigned long parent_rate, u8 index)
+{
+ return clk_rcg2_dp_set_rate(hw, rate, parent_rate);
+}
+
+static int clk_rcg2_dp_determine_rate(struct clk_hw *hw,
+ struct clk_rate_request *req)
+{
+ if (!hw)
+ return -EINVAL;
+
+ if (!clk_hw_get_parent(hw)) {
+ pr_err("Missing the parent for the DP RCG\n");
+ return -EINVAL;
+ }
+
+ req->best_parent_rate = clk_hw_get_rate(clk_hw_get_parent(hw));
+ return 0;
+}
+
+const struct clk_ops clk_dp_ops = {
+ .is_enabled = clk_rcg2_is_enabled,
+ .get_parent = clk_rcg2_get_parent,
+ .set_parent = clk_rcg2_set_parent,
+ .recalc_rate = clk_rcg2_recalc_rate,
+ .set_rate = clk_rcg2_dp_set_rate,
+ .set_rate_and_parent = clk_rcg2_dp_set_rate_and_parent,
+ .determine_rate = clk_rcg2_dp_determine_rate,
+};
+EXPORT_SYMBOL_GPL(clk_dp_ops);
--
Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member
of the Code Aurora Forum, hosted by the Linux Foundation.
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-09 13:57 [PATCH v1 0/2] Add support for display port clocks and clock ops Taniya Das
2018-10-09 13:57 ` [PATCH v1 1/2] clk: qcom: rcg2: Add support for display port " Taniya Das
@ 2018-10-09 13:57 ` Taniya Das
2018-10-09 20:34 ` Stephen Boyd
1 sibling, 1 reply; 16+ messages in thread
From: Taniya Das @ 2018-10-09 13:57 UTC (permalink / raw)
To: Stephen Boyd, Michael Turquette
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu, Taniya Das
SDM845 dispcc supports RCG and CBCRs for display port, so add support for
the same.
Signed-off-by: Taniya Das <tdas@codeaurora.org>
---
drivers/clk/qcom/dispcc-sdm845.c | 232 +++++++++++++++++++++++++
include/dt-bindings/clock/qcom,dispcc-sdm845.h | 11 ++
2 files changed, 243 insertions(+)
diff --git a/drivers/clk/qcom/dispcc-sdm845.c b/drivers/clk/qcom/dispcc-sdm845.c
index 0cc4909..6d3136a 100644
--- a/drivers/clk/qcom/dispcc-sdm845.c
+++ b/drivers/clk/qcom/dispcc-sdm845.c
@@ -29,6 +29,8 @@ enum {
P_DSI1_PHY_PLL_OUT_DSICLK,
P_GPLL0_OUT_MAIN,
P_GPLL0_OUT_MAIN_DIV,
+ P_DP_PHY_PLL_LINK_CLK,
+ P_DP_PHY_PLL_VCO_DIV_CLK,
};
static const struct parent_map disp_cc_parent_map_0[] = {
@@ -45,6 +47,20 @@ enum {
"core_bi_pll_test_se",
};
+static const struct parent_map disp_cc_parent_map_1[] = {
+ { P_BI_TCXO, 0 },
+ { P_DP_PHY_PLL_LINK_CLK, 1 },
+ { P_DP_PHY_PLL_VCO_DIV_CLK, 2 },
+ { P_CORE_BI_PLL_TEST_SE, 7 },
+};
+
+static const char * const disp_cc_parent_names_1[] = {
+ "bi_tcxo",
+ "dp_link_clk_divsel_ten",
+ "dp_vco_divided_clk_src_mux",
+ "core_bi_pll_test_se",
+};
+
static const struct parent_map disp_cc_parent_map_2[] = {
{ P_BI_TCXO, 0 },
{ P_CORE_BI_PLL_TEST_SE, 7 },
@@ -128,6 +144,100 @@ enum {
},
};
+static const struct freq_tbl ftbl_disp_cc_mdss_dp_aux_clk_src[] = {
+ F(19200000, P_BI_TCXO, 1, 0, 0),
+ { }
+};
+
+static struct clk_rcg2 disp_cc_mdss_dp_aux_clk_src = {
+ .cmd_rcgr = 0x219c,
+ .mnd_width = 0,
+ .hid_width = 5,
+ .parent_map = disp_cc_parent_map_2,
+ .freq_tbl = ftbl_disp_cc_mdss_dp_aux_clk_src,
+ .clkr.hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_aux_clk_src",
+ .parent_names = disp_cc_parent_names_2,
+ .num_parents = 2,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_rcg2_ops,
+ },
+};
+
+static const struct freq_tbl ftbl_disp_cc_mdss_dp_crypto_clk_src[] = {
+ F(108000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
+ F(180000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
+ F(360000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
+ F(540000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
+ { }
+};
+
+static struct clk_rcg2 disp_cc_mdss_dp_crypto_clk_src = {
+ .cmd_rcgr = 0x2154,
+ .mnd_width = 0,
+ .hid_width = 5,
+ .parent_map = disp_cc_parent_map_1,
+ .freq_tbl = ftbl_disp_cc_mdss_dp_crypto_clk_src,
+ .clkr.hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_crypto_clk_src",
+ .parent_names = disp_cc_parent_names_1,
+ .num_parents = 4,
+ .flags = CLK_GET_RATE_NOCACHE,
+ .ops = &clk_rcg2_ops,
+ },
+};
+
+static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
+ F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
+ F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
+ F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
+ F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
+ { }
+};
+
+static struct clk_rcg2 disp_cc_mdss_dp_link_clk_src = {
+ .cmd_rcgr = 0x2138,
+ .mnd_width = 0,
+ .hid_width = 5,
+ .parent_map = disp_cc_parent_map_1,
+ .freq_tbl = ftbl_disp_cc_mdss_dp_link_clk_src,
+ .clkr.hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_link_clk_src",
+ .parent_names = disp_cc_parent_names_1,
+ .num_parents = 4,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_rcg2_ops,
+ },
+};
+
+static struct clk_rcg2 disp_cc_mdss_dp_pixel1_clk_src = {
+ .cmd_rcgr = 0x2184,
+ .mnd_width = 16,
+ .hid_width = 5,
+ .parent_map = disp_cc_parent_map_1,
+ .clkr.hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_pixel1_clk_src",
+ .parent_names = disp_cc_parent_names_1,
+ .num_parents = 4,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_dp_ops,
+ },
+};
+
+static struct clk_rcg2 disp_cc_mdss_dp_pixel_clk_src = {
+ .cmd_rcgr = 0x216c,
+ .mnd_width = 16,
+ .hid_width = 5,
+ .parent_map = disp_cc_parent_map_1,
+ .clkr.hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_pixel_clk_src",
+ .parent_names = disp_cc_parent_names_1,
+ .num_parents = 4,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_dp_ops,
+ },
+};
+
static const struct freq_tbl ftbl_disp_cc_mdss_esc0_clk_src[] = {
F(19200000, P_BI_TCXO, 1, 0, 0),
{ }
@@ -391,6 +501,115 @@ enum {
},
};
+static struct clk_branch disp_cc_mdss_dp_aux_clk = {
+ .halt_reg = 0x2054,
+ .halt_check = BRANCH_HALT,
+ .clkr = {
+ .enable_reg = 0x2054,
+ .enable_mask = BIT(0),
+ .hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_aux_clk",
+ .parent_names = (const char *[]){
+ "disp_cc_mdss_dp_aux_clk_src",
+ },
+ .num_parents = 1,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_branch2_ops,
+ },
+ },
+};
+
+static struct clk_branch disp_cc_mdss_dp_crypto_clk = {
+ .halt_reg = 0x2048,
+ .halt_check = BRANCH_HALT,
+ .clkr = {
+ .enable_reg = 0x2048,
+ .enable_mask = BIT(0),
+ .hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_crypto_clk",
+ .parent_names = (const char *[]){
+ "disp_cc_mdss_dp_crypto_clk_src",
+ },
+ .num_parents = 1,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_branch2_ops,
+ },
+ },
+};
+
+static struct clk_branch disp_cc_mdss_dp_link_clk = {
+ .halt_reg = 0x2040,
+ .halt_check = BRANCH_HALT,
+ .clkr = {
+ .enable_reg = 0x2040,
+ .enable_mask = BIT(0),
+ .hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_link_clk",
+ .parent_names = (const char *[]){
+ "disp_cc_mdss_dp_link_clk_src",
+ },
+ .num_parents = 1,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_branch2_ops,
+ },
+ },
+};
+
+/* reset state of disp_cc_mdss_dp_link_div_clk_src divider is 0x3 (div 4) */
+static struct clk_branch disp_cc_mdss_dp_link_intf_clk = {
+ .halt_reg = 0x2044,
+ .halt_check = BRANCH_HALT,
+ .clkr = {
+ .enable_reg = 0x2044,
+ .enable_mask = BIT(0),
+ .hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_link_intf_clk",
+ .parent_names = (const char *[]){
+ "disp_cc_mdss_dp_link_clk_src",
+ },
+ .num_parents = 1,
+ .flags = CLK_GET_RATE_NOCACHE,
+ .ops = &clk_branch2_ops,
+ },
+ },
+};
+
+static struct clk_branch disp_cc_mdss_dp_pixel1_clk = {
+ .halt_reg = 0x2050,
+ .halt_check = BRANCH_HALT,
+ .clkr = {
+ .enable_reg = 0x2050,
+ .enable_mask = BIT(0),
+ .hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_pixel1_clk",
+ .parent_names = (const char *[]){
+ "disp_cc_mdss_dp_pixel1_clk_src",
+ },
+ .num_parents = 1,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_branch2_ops,
+ },
+ },
+};
+
+static struct clk_branch disp_cc_mdss_dp_pixel_clk = {
+ .halt_reg = 0x204c,
+ .halt_check = BRANCH_HALT,
+ .clkr = {
+ .enable_reg = 0x204c,
+ .enable_mask = BIT(0),
+ .hw.init = &(struct clk_init_data){
+ .name = "disp_cc_mdss_dp_pixel_clk",
+ .parent_names = (const char *[]){
+ "disp_cc_mdss_dp_pixel_clk_src",
+ },
+ .num_parents = 1,
+ .flags = CLK_SET_RATE_PARENT,
+ .ops = &clk_branch2_ops,
+ },
+ },
+};
+
static struct clk_branch disp_cc_mdss_esc0_clk = {
.halt_reg = 0x2038,
.halt_check = BRANCH_HALT,
@@ -589,6 +808,19 @@ enum {
[DISP_CC_MDSS_BYTE1_INTF_CLK] = &disp_cc_mdss_byte1_intf_clk.clkr,
[DISP_CC_MDSS_BYTE1_DIV_CLK_SRC] =
&disp_cc_mdss_byte1_div_clk_src.clkr,
+ [DISP_CC_MDSS_DP_AUX_CLK] = &disp_cc_mdss_dp_aux_clk.clkr,
+ [DISP_CC_MDSS_DP_AUX_CLK_SRC] = &disp_cc_mdss_dp_aux_clk_src.clkr,
+ [DISP_CC_MDSS_DP_CRYPTO_CLK] = &disp_cc_mdss_dp_crypto_clk.clkr,
+ [DISP_CC_MDSS_DP_CRYPTO_CLK_SRC] =
+ &disp_cc_mdss_dp_crypto_clk_src.clkr,
+ [DISP_CC_MDSS_DP_LINK_CLK] = &disp_cc_mdss_dp_link_clk.clkr,
+ [DISP_CC_MDSS_DP_LINK_CLK_SRC] = &disp_cc_mdss_dp_link_clk_src.clkr,
+ [DISP_CC_MDSS_DP_LINK_INTF_CLK] = &disp_cc_mdss_dp_link_intf_clk.clkr,
+ [DISP_CC_MDSS_DP_PIXEL1_CLK] = &disp_cc_mdss_dp_pixel1_clk.clkr,
+ [DISP_CC_MDSS_DP_PIXEL1_CLK_SRC] =
+ &disp_cc_mdss_dp_pixel1_clk_src.clkr,
+ [DISP_CC_MDSS_DP_PIXEL_CLK] = &disp_cc_mdss_dp_pixel_clk.clkr,
+ [DISP_CC_MDSS_DP_PIXEL_CLK_SRC] = &disp_cc_mdss_dp_pixel_clk_src.clkr,
[DISP_CC_MDSS_ESC0_CLK] = &disp_cc_mdss_esc0_clk.clkr,
[DISP_CC_MDSS_ESC0_CLK_SRC] = &disp_cc_mdss_esc0_clk_src.clkr,
[DISP_CC_MDSS_ESC1_CLK] = &disp_cc_mdss_esc1_clk.clkr,
diff --git a/include/dt-bindings/clock/qcom,dispcc-sdm845.h b/include/dt-bindings/clock/qcom,dispcc-sdm845.h
index 11eed4b..526d98a 100644
--- a/include/dt-bindings/clock/qcom,dispcc-sdm845.h
+++ b/include/dt-bindings/clock/qcom,dispcc-sdm845.h
@@ -35,6 +35,17 @@
#define DISP_CC_PLL0 25
#define DISP_CC_MDSS_BYTE0_DIV_CLK_SRC 26
#define DISP_CC_MDSS_BYTE1_DIV_CLK_SRC 27
+#define DISP_CC_MDSS_DP_AUX_CLK 28
+#define DISP_CC_MDSS_DP_AUX_CLK_SRC 29
+#define DISP_CC_MDSS_DP_CRYPTO_CLK 30
+#define DISP_CC_MDSS_DP_CRYPTO_CLK_SRC 31
+#define DISP_CC_MDSS_DP_LINK_CLK 32
+#define DISP_CC_MDSS_DP_LINK_CLK_SRC 33
+#define DISP_CC_MDSS_DP_LINK_INTF_CLK 34
+#define DISP_CC_MDSS_DP_PIXEL1_CLK 35
+#define DISP_CC_MDSS_DP_PIXEL1_CLK_SRC 36
+#define DISP_CC_MDSS_DP_PIXEL_CLK 37
+#define DISP_CC_MDSS_DP_PIXEL_CLK_SRC 38
/* DISP_CC Reset */
#define DISP_CC_MDSS_RSCC_BCR 0
--
Qualcomm INDIA, on behalf of Qualcomm Innovation Center, Inc.is a member
of the Code Aurora Forum, hosted by the Linux Foundation.
^ permalink raw reply related [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-09 13:57 ` [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks Taniya Das
@ 2018-10-09 20:34 ` Stephen Boyd
2018-10-19 10:34 ` Taniya Das
0 siblings, 1 reply; 16+ messages in thread
From: Stephen Boyd @ 2018-10-09 20:34 UTC (permalink / raw)
To: Michael Turquette, Taniya Das
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu, Taniya Das
Quoting Taniya Das (2018-10-09 06:57:47)
> diff --git a/drivers/clk/qcom/dispcc-sdm845.c b/drivers/clk/qcom/dispcc-sdm845.c
> index 0cc4909..6d3136a 100644
> --- a/drivers/clk/qcom/dispcc-sdm845.c
> +++ b/drivers/clk/qcom/dispcc-sdm845.c
> @@ -128,6 +144,100 @@ enum {
> },
> };
>
> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_aux_clk_src[] = {
> + F(19200000, P_BI_TCXO, 1, 0, 0),
> + { }
> +};
> +
> +static struct clk_rcg2 disp_cc_mdss_dp_aux_clk_src = {
> + .cmd_rcgr = 0x219c,
> + .mnd_width = 0,
> + .hid_width = 5,
> + .parent_map = disp_cc_parent_map_2,
> + .freq_tbl = ftbl_disp_cc_mdss_dp_aux_clk_src,
> + .clkr.hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_aux_clk_src",
> + .parent_names = disp_cc_parent_names_2,
> + .num_parents = 2,
> + .flags = CLK_SET_RATE_PARENT,
> + .ops = &clk_rcg2_ops,
> + },
> +};
> +
> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_crypto_clk_src[] = {
> + F(108000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
> + F(180000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
> + F(360000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
> + F(540000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
> + { }
> +};
> +
> +static struct clk_rcg2 disp_cc_mdss_dp_crypto_clk_src = {
> + .cmd_rcgr = 0x2154,
> + .mnd_width = 0,
> + .hid_width = 5,
> + .parent_map = disp_cc_parent_map_1,
> + .freq_tbl = ftbl_disp_cc_mdss_dp_crypto_clk_src,
> + .clkr.hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_crypto_clk_src",
> + .parent_names = disp_cc_parent_names_1,
> + .num_parents = 4,
> + .flags = CLK_GET_RATE_NOCACHE,
Why?
> + .ops = &clk_rcg2_ops,
> + },
> +};
> +
> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
> + F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> + F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> + F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> + F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
Are these in kHz? They really look like it and that's bad. Why do we
need them at all? Just to make sure the display driver picks these exact
frequencies? It seems like we could just pass whatever number comes in
up to the parent and see what it can do.
> + { }
> +};
> +
> +static struct clk_rcg2 disp_cc_mdss_dp_link_clk_src = {
> + .cmd_rcgr = 0x2138,
> + .mnd_width = 0,
> + .hid_width = 5,
> + .parent_map = disp_cc_parent_map_1,
> + .freq_tbl = ftbl_disp_cc_mdss_dp_link_clk_src,
> + .clkr.hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_link_clk_src",
> + .parent_names = disp_cc_parent_names_1,
> + .num_parents = 4,
> + .flags = CLK_SET_RATE_PARENT,
> + .ops = &clk_rcg2_ops,
> + },
> +};
> +
> +static struct clk_rcg2 disp_cc_mdss_dp_pixel1_clk_src = {
> + .cmd_rcgr = 0x2184,
> + .mnd_width = 16,
> + .hid_width = 5,
> + .parent_map = disp_cc_parent_map_1,
> + .clkr.hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_pixel1_clk_src",
> + .parent_names = disp_cc_parent_names_1,
> + .num_parents = 4,
> + .flags = CLK_SET_RATE_PARENT,
> + .ops = &clk_dp_ops,
> + },
> +};
> +
> +static struct clk_rcg2 disp_cc_mdss_dp_pixel_clk_src = {
> + .cmd_rcgr = 0x216c,
> + .mnd_width = 16,
> + .hid_width = 5,
> + .parent_map = disp_cc_parent_map_1,
> + .clkr.hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_pixel_clk_src",
> + .parent_names = disp_cc_parent_names_1,
> + .num_parents = 4,
> + .flags = CLK_SET_RATE_PARENT,
> + .ops = &clk_dp_ops,
> + },
> +};
> +
> static const struct freq_tbl ftbl_disp_cc_mdss_esc0_clk_src[] = {
> F(19200000, P_BI_TCXO, 1, 0, 0),
> { }
> @@ -391,6 +501,115 @@ enum {
> },
> };
>
> +static struct clk_branch disp_cc_mdss_dp_aux_clk = {
> + .halt_reg = 0x2054,
> + .halt_check = BRANCH_HALT,
> + .clkr = {
> + .enable_reg = 0x2054,
> + .enable_mask = BIT(0),
> + .hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_aux_clk",
> + .parent_names = (const char *[]){
> + "disp_cc_mdss_dp_aux_clk_src",
> + },
> + .num_parents = 1,
> + .flags = CLK_SET_RATE_PARENT,
> + .ops = &clk_branch2_ops,
> + },
> + },
> +};
> +
> +static struct clk_branch disp_cc_mdss_dp_crypto_clk = {
> + .halt_reg = 0x2048,
> + .halt_check = BRANCH_HALT,
> + .clkr = {
> + .enable_reg = 0x2048,
> + .enable_mask = BIT(0),
> + .hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_crypto_clk",
> + .parent_names = (const char *[]){
> + "disp_cc_mdss_dp_crypto_clk_src",
> + },
> + .num_parents = 1,
> + .flags = CLK_SET_RATE_PARENT,
> + .ops = &clk_branch2_ops,
> + },
> + },
> +};
> +
> +static struct clk_branch disp_cc_mdss_dp_link_clk = {
> + .halt_reg = 0x2040,
> + .halt_check = BRANCH_HALT,
> + .clkr = {
> + .enable_reg = 0x2040,
> + .enable_mask = BIT(0),
> + .hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_link_clk",
> + .parent_names = (const char *[]){
> + "disp_cc_mdss_dp_link_clk_src",
> + },
> + .num_parents = 1,
> + .flags = CLK_SET_RATE_PARENT,
> + .ops = &clk_branch2_ops,
> + },
> + },
> +};
> +
> +/* reset state of disp_cc_mdss_dp_link_div_clk_src divider is 0x3 (div 4) */
Not sure what this comment is for. But it's interesting nonetheless.
> +static struct clk_branch disp_cc_mdss_dp_link_intf_clk = {
> + .halt_reg = 0x2044,
> + .halt_check = BRANCH_HALT,
> + .clkr = {
> + .enable_reg = 0x2044,
> + .enable_mask = BIT(0),
> + .hw.init = &(struct clk_init_data){
> + .name = "disp_cc_mdss_dp_link_intf_clk",
> + .parent_names = (const char *[]){
> + "disp_cc_mdss_dp_link_clk_src",
> + },
> + .num_parents = 1,
> + .flags = CLK_GET_RATE_NOCACHE,
Why?
> + .ops = &clk_branch2_ops,
> + },
> + },
> +};
> +
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 1/2] clk: qcom: rcg2: Add support for display port clock ops
2018-10-09 13:57 ` [PATCH v1 1/2] clk: qcom: rcg2: Add support for display port " Taniya Das
@ 2018-10-09 20:46 ` Stephen Boyd
2018-10-19 10:31 ` Taniya Das
0 siblings, 1 reply; 16+ messages in thread
From: Stephen Boyd @ 2018-10-09 20:46 UTC (permalink / raw)
To: Michael Turquette, Taniya Das
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu, Taniya Das
Quoting Taniya Das (2018-10-09 06:57:46)
> diff --git a/drivers/clk/qcom/clk-rcg2.c b/drivers/clk/qcom/clk-rcg2.c
> index 6e3bd19..ca6142f 100644
> --- a/drivers/clk/qcom/clk-rcg2.c
> +++ b/drivers/clk/qcom/clk-rcg2.c
> @@ -10,6 +10,7 @@
> #include <linux/export.h>
> #include <linux/clk-provider.h>
> #include <linux/delay.h>
> +#include <linux/rational.h>
Can you also select RATIONAL in the Kconfig language? Yes the clk
subsystem is already selecting it, but it's nice to be explicit in the
subdrivers.
> #include <linux/regmap.h>
> #include <linux/math64.h>
> #include <linux/slab.h>
> @@ -1124,3 +1125,88 @@ int qcom_cc_register_rcg_dfs(struct regmap *regmap,
> return 0;
> }
> EXPORT_SYMBOL_GPL(qcom_cc_register_rcg_dfs);
> +
> +static int clk_rcg2_dp_set_rate(struct clk_hw *hw, unsigned long rate,
> + unsigned long parent_rate)
> +{
> + struct clk_rcg2 *rcg = to_clk_rcg2(hw);
> + struct freq_tbl f = { 0 };
> + unsigned long src_rate;
> + unsigned long num, den;
> + u32 mask = BIT(rcg->hid_width) - 1;
> + u32 hid_div, cfg;
> + int i, num_parents = clk_hw_get_num_parents(hw);
> +
> + src_rate = clk_hw_get_rate(clk_hw_get_parent(hw));
What is this doing? We shouldn't need to check the parent clk for any
rate here when we're setting the rate.
> + if (src_rate <= 0) {
> + pr_err("Invalid RCG parent rate\n");
> + return -EINVAL;
> + }
> +
> + rational_best_approximation(src_rate, rate,
> + (unsigned long)(1 << 16) - 1,
Use GENMASK?
> + (unsigned long)(1 << 16) - 1, &den, &num);
Same?
> +
> + if (!num || !den) {
> + pr_err("Invalid MN values derived for requested rate %lu\n",
> + rate);
> + return -EINVAL;
> + }
> +
> + regmap_read(rcg->clkr.regmap, rcg->cmd_rcgr + CFG_REG, &cfg);
> + hid_div = cfg;
> + cfg &= CFG_SRC_SEL_MASK;
> + cfg >>= CFG_SRC_SEL_SHIFT;
> +
> + for (i = 0; i < num_parents; i++)
> + if (cfg == rcg->parent_map[i].cfg) {
> + f.src = rcg->parent_map[i].src;
> + break;
> + }
> +
> + f.pre_div = hid_div;
> + f.pre_div >>= CFG_SRC_DIV_SHIFT;
> + f.pre_div &= mask;
> +
> + if (num == den) {
> + f.m = 0;
> + f.n = 0;
> + } else {
> + f.m = num;
> + f.n = den;
> + }
> +
> + return clk_rcg2_configure(rcg, &f);
> +}
> +
> +static int clk_rcg2_dp_set_rate_and_parent(struct clk_hw *hw,
> + unsigned long rate, unsigned long parent_rate, u8 index)
> +{
> + return clk_rcg2_dp_set_rate(hw, rate, parent_rate);
> +}
> +
> +static int clk_rcg2_dp_determine_rate(struct clk_hw *hw,
> + struct clk_rate_request *req)
> +{
> + if (!hw)
> + return -EINVAL;
> +
> + if (!clk_hw_get_parent(hw)) {
> + pr_err("Missing the parent for the DP RCG\n");
> + return -EINVAL;
> + }
Let me Harry Potter this stuff. Expelliarmus!
> +
> + req->best_parent_rate = clk_hw_get_rate(clk_hw_get_parent(hw));
Presumably we should ask the parent clk for the rate that is requested
here by calling determine rate up and see if the parent can do it. Sure,
this clk does nothing, so we don't really need any sort of op here then
and we can just flag the clk as CLK_SET_RATE_PARENT and let the core do
the rest.
> + return 0;
> +}
> +
> +const struct clk_ops clk_dp_ops = {
> + .is_enabled = clk_rcg2_is_enabled,
> + .get_parent = clk_rcg2_get_parent,
> + .set_parent = clk_rcg2_set_parent,
> + .recalc_rate = clk_rcg2_recalc_rate,
> + .set_rate = clk_rcg2_dp_set_rate,
> + .set_rate_and_parent = clk_rcg2_dp_set_rate_and_parent,
> + .determine_rate = clk_rcg2_dp_determine_rate,
> +};
> +EXPORT_SYMBOL_GPL(clk_dp_ops);
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 1/2] clk: qcom: rcg2: Add support for display port clock ops
2018-10-09 20:46 ` Stephen Boyd
@ 2018-10-19 10:31 ` Taniya Das
0 siblings, 0 replies; 16+ messages in thread
From: Taniya Das @ 2018-10-19 10:31 UTC (permalink / raw)
To: Stephen Boyd, Michael Turquette
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu
Hello Stephen,
On 10/10/2018 2:16 AM, Stephen Boyd wrote:
> Quoting Taniya Das (2018-10-09 06:57:46)
>> diff --git a/drivers/clk/qcom/clk-rcg2.c b/drivers/clk/qcom/clk-rcg2.c
>> index 6e3bd19..ca6142f 100644
>> --- a/drivers/clk/qcom/clk-rcg2.c
>> +++ b/drivers/clk/qcom/clk-rcg2.c
>> @@ -10,6 +10,7 @@
>> #include <linux/export.h>
>> #include <linux/clk-provider.h>
>> #include <linux/delay.h>
>> +#include <linux/rational.h>
>
> Can you also select RATIONAL in the Kconfig language? Yes the clk
> subsystem is already selecting it, but it's nice to be explicit in the
> subdrivers.
>
Sure, would take care of adding it in KCONFIG.
>> #include <linux/regmap.h>
>> #include <linux/math64.h>
>> #include <linux/slab.h>
>> @@ -1124,3 +1125,88 @@ int qcom_cc_register_rcg_dfs(struct regmap *regmap,
>> return 0;
>> }
>> EXPORT_SYMBOL_GPL(qcom_cc_register_rcg_dfs);
>> +
>> +static int clk_rcg2_dp_set_rate(struct clk_hw *hw, unsigned long rate,
>> + unsigned long parent_rate)
>> +{
>> + struct clk_rcg2 *rcg = to_clk_rcg2(hw);
>> + struct freq_tbl f = { 0 };
>> + unsigned long src_rate;
>> + unsigned long num, den;
>> + u32 mask = BIT(rcg->hid_width) - 1;
>> + u32 hid_div, cfg;
>> + int i, num_parents = clk_hw_get_num_parents(hw);
>> +
>> + src_rate = clk_hw_get_rate(clk_hw_get_parent(hw));
>
> What is this doing? We shouldn't need to check the parent clk for any
> rate here when we're setting the rate.
>
Hmmm, let me get back on this.
>> + if (src_rate <= 0) {
>> + pr_err("Invalid RCG parent rate\n");
>> + return -EINVAL;
>> + }
>> +
>> + rational_best_approximation(src_rate, rate,
>> + (unsigned long)(1 << 16) - 1,
>
> Use GENMASK?
>
>> + (unsigned long)(1 << 16) - 1, &den, &num);
>
> Same? >
Sure, would use GENMASK.
>> +
>> + if (!num || !den) {
>> + pr_err("Invalid MN values derived for requested rate %lu\n",
>> + rate);
>> + return -EINVAL;
>> + }
>> +
>> + regmap_read(rcg->clkr.regmap, rcg->cmd_rcgr + CFG_REG, &cfg);
>> + hid_div = cfg;
>> + cfg &= CFG_SRC_SEL_MASK;
>> + cfg >>= CFG_SRC_SEL_SHIFT;
>> +
>> + for (i = 0; i < num_parents; i++)
>> + if (cfg == rcg->parent_map[i].cfg) {
>> + f.src = rcg->parent_map[i].src;
>> + break;
>> + }
>> +
>> + f.pre_div = hid_div;
>> + f.pre_div >>= CFG_SRC_DIV_SHIFT;
>> + f.pre_div &= mask;
>> +
>> + if (num == den) {
>> + f.m = 0;
>> + f.n = 0;
>> + } else {
>> + f.m = num;
>> + f.n = den;
>> + }
>> +
>> + return clk_rcg2_configure(rcg, &f);
>> +}
>> +
>> +static int clk_rcg2_dp_set_rate_and_parent(struct clk_hw *hw,
>> + unsigned long rate, unsigned long parent_rate, u8 index)
>> +{
>> + return clk_rcg2_dp_set_rate(hw, rate, parent_rate);
>> +}
>> +
>> +static int clk_rcg2_dp_determine_rate(struct clk_hw *hw,
>> + struct clk_rate_request *req)
>> +{
>> + if (!hw)
>> + return -EINVAL;
>> +
>> + if (!clk_hw_get_parent(hw)) {
>> + pr_err("Missing the parent for the DP RCG\n");
>> + return -EINVAL;
>> + }
>
> Let me Harry Potter this stuff. Expelliarmus!
>
>> +
>> + req->best_parent_rate = clk_hw_get_rate(clk_hw_get_parent(hw));
>
> Presumably we should ask the parent clk for the rate that is requested
> here by calling determine rate up and see if the parent can do it. Sure,
> this clk does nothing, so we don't really need any sort of op here then
> and we can just flag the clk as CLK_SET_RATE_PARENT and let the core do
> the rest.
>
>> + return 0;
>> +}
>> +
Would remove this.
>> +const struct clk_ops clk_dp_ops = {
>> + .is_enabled = clk_rcg2_is_enabled,
>> + .get_parent = clk_rcg2_get_parent,
>> + .set_parent = clk_rcg2_set_parent,
>> + .recalc_rate = clk_rcg2_recalc_rate,
>> + .set_rate = clk_rcg2_dp_set_rate,
>> + .set_rate_and_parent = clk_rcg2_dp_set_rate_and_parent,
>> + .determine_rate = clk_rcg2_dp_determine_rate,
>> +};
>> +EXPORT_SYMBOL_GPL(clk_dp_ops);
--
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation.
--
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-09 20:34 ` Stephen Boyd
@ 2018-10-19 10:34 ` Taniya Das
2018-10-28 10:34 ` Taniya Das
0 siblings, 1 reply; 16+ messages in thread
From: Taniya Das @ 2018-10-19 10:34 UTC (permalink / raw)
To: Stephen Boyd, Michael Turquette
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu
Hello Stephen,
On 10/10/2018 2:04 AM, Stephen Boyd wrote:
> Quoting Taniya Das (2018-10-09 06:57:47)
>> diff --git a/drivers/clk/qcom/dispcc-sdm845.c b/drivers/clk/qcom/dispcc-sdm845.c
>> index 0cc4909..6d3136a 100644
>> --- a/drivers/clk/qcom/dispcc-sdm845.c
>> +++ b/drivers/clk/qcom/dispcc-sdm845.c
>> @@ -128,6 +144,100 @@ enum {
>> },
>> };
>>
>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_aux_clk_src[] = {
>> + F(19200000, P_BI_TCXO, 1, 0, 0),
>> + { }
>> +};
>> +
>> +static struct clk_rcg2 disp_cc_mdss_dp_aux_clk_src = {
>> + .cmd_rcgr = 0x219c,
>> + .mnd_width = 0,
>> + .hid_width = 5,
>> + .parent_map = disp_cc_parent_map_2,
>> + .freq_tbl = ftbl_disp_cc_mdss_dp_aux_clk_src,
>> + .clkr.hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_aux_clk_src",
>> + .parent_names = disp_cc_parent_names_2,
>> + .num_parents = 2,
>> + .flags = CLK_SET_RATE_PARENT,
>> + .ops = &clk_rcg2_ops,
>> + },
>> +};
>> +
>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_crypto_clk_src[] = {
>> + F(108000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>> + F(180000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>> + F(360000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>> + { }
>> +};
>> +
>> +static struct clk_rcg2 disp_cc_mdss_dp_crypto_clk_src = {
>> + .cmd_rcgr = 0x2154,
>> + .mnd_width = 0,
>> + .hid_width = 5,
>> + .parent_map = disp_cc_parent_map_1,
>> + .freq_tbl = ftbl_disp_cc_mdss_dp_crypto_clk_src,
>> + .clkr.hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_crypto_clk_src",
>> + .parent_names = disp_cc_parent_names_1,
>> + .num_parents = 4,
>> + .flags = CLK_GET_RATE_NOCACHE,
>
> Why?
>
>> + .ops = &clk_rcg2_ops,
>> + },
>> +};
>> +
>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
>> + F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>> + F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>> + F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>
> Are these in kHz? They really look like it and that's bad. Why do we
> need them at all? Just to make sure the display driver picks these exact
> frequencies? It seems like we could just pass whatever number comes in
> up to the parent and see what it can do.
>
Let me check back the reason we had to make this change.
>> + { }
>> +};
>> +
>> +static struct clk_rcg2 disp_cc_mdss_dp_link_clk_src = {
>> + .cmd_rcgr = 0x2138,
>> + .mnd_width = 0,
>> + .hid_width = 5,
>> + .parent_map = disp_cc_parent_map_1,
>> + .freq_tbl = ftbl_disp_cc_mdss_dp_link_clk_src,
>> + .clkr.hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_link_clk_src",
>> + .parent_names = disp_cc_parent_names_1,
>> + .num_parents = 4,
>> + .flags = CLK_SET_RATE_PARENT,
>> + .ops = &clk_rcg2_ops,
>> + },
>> +};
>> +
>> +static struct clk_rcg2 disp_cc_mdss_dp_pixel1_clk_src = {
>> + .cmd_rcgr = 0x2184,
>> + .mnd_width = 16,
>> + .hid_width = 5,
>> + .parent_map = disp_cc_parent_map_1,
>> + .clkr.hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_pixel1_clk_src",
>> + .parent_names = disp_cc_parent_names_1,
>> + .num_parents = 4,
>> + .flags = CLK_SET_RATE_PARENT,
>> + .ops = &clk_dp_ops,
>> + },
>> +};
>> +
>> +static struct clk_rcg2 disp_cc_mdss_dp_pixel_clk_src = {
>> + .cmd_rcgr = 0x216c,
>> + .mnd_width = 16,
>> + .hid_width = 5,
>> + .parent_map = disp_cc_parent_map_1,
>> + .clkr.hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_pixel_clk_src",
>> + .parent_names = disp_cc_parent_names_1,
>> + .num_parents = 4,
>> + .flags = CLK_SET_RATE_PARENT,
>> + .ops = &clk_dp_ops,
>> + },
>> +};
>> +
>> static const struct freq_tbl ftbl_disp_cc_mdss_esc0_clk_src[] = {
>> F(19200000, P_BI_TCXO, 1, 0, 0),
>> { }
>> @@ -391,6 +501,115 @@ enum {
>> },
>> };
>>
>> +static struct clk_branch disp_cc_mdss_dp_aux_clk = {
>> + .halt_reg = 0x2054,
>> + .halt_check = BRANCH_HALT,
>> + .clkr = {
>> + .enable_reg = 0x2054,
>> + .enable_mask = BIT(0),
>> + .hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_aux_clk",
>> + .parent_names = (const char *[]){
>> + "disp_cc_mdss_dp_aux_clk_src",
>> + },
>> + .num_parents = 1,
>> + .flags = CLK_SET_RATE_PARENT,
>> + .ops = &clk_branch2_ops,
>> + },
>> + },
>> +};
>> +
>> +static struct clk_branch disp_cc_mdss_dp_crypto_clk = {
>> + .halt_reg = 0x2048,
>> + .halt_check = BRANCH_HALT,
>> + .clkr = {
>> + .enable_reg = 0x2048,
>> + .enable_mask = BIT(0),
>> + .hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_crypto_clk",
>> + .parent_names = (const char *[]){
>> + "disp_cc_mdss_dp_crypto_clk_src",
>> + },
>> + .num_parents = 1,
>> + .flags = CLK_SET_RATE_PARENT,
>> + .ops = &clk_branch2_ops,
>> + },
>> + },
>> +};
>> +
>> +static struct clk_branch disp_cc_mdss_dp_link_clk = {
>> + .halt_reg = 0x2040,
>> + .halt_check = BRANCH_HALT,
>> + .clkr = {
>> + .enable_reg = 0x2040,
>> + .enable_mask = BIT(0),
>> + .hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_link_clk",
>> + .parent_names = (const char *[]){
>> + "disp_cc_mdss_dp_link_clk_src",
>> + },
>> + .num_parents = 1,
>> + .flags = CLK_SET_RATE_PARENT,
>> + .ops = &clk_branch2_ops,
>> + },
>> + },
>> +};
>> +
>> +/* reset state of disp_cc_mdss_dp_link_div_clk_src divider is 0x3 (div 4) */
>
> Not sure what this comment is for. But it's interesting nonetheless.
>
>> +static struct clk_branch disp_cc_mdss_dp_link_intf_clk = {
>> + .halt_reg = 0x2044,
>> + .halt_check = BRANCH_HALT,
>> + .clkr = {
>> + .enable_reg = 0x2044,
>> + .enable_mask = BIT(0),
>> + .hw.init = &(struct clk_init_data){
>> + .name = "disp_cc_mdss_dp_link_intf_clk",
>> + .parent_names = (const char *[]){
>> + "disp_cc_mdss_dp_link_clk_src",
>> + },
>> + .num_parents = 1,
>> + .flags = CLK_GET_RATE_NOCACHE,
>
> Why?
>
It was a requirement, but let me get back on this too.
>> + .ops = &clk_branch2_ops,
>> + },
>> + },
>> +};
>> +
--
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation.
--
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-19 10:34 ` Taniya Das
@ 2018-10-28 10:34 ` Taniya Das
2018-10-29 18:43 ` Stephen Boyd
0 siblings, 1 reply; 16+ messages in thread
From: Taniya Das @ 2018-10-28 10:34 UTC (permalink / raw)
To: Stephen Boyd, Michael Turquette
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu,
linux-arm-msm-owner
Hello Stephen,
On 2018-10-19 16:04, Taniya Das wrote:
> Hello Stephen,
>
> On 10/10/2018 2:04 AM, Stephen Boyd wrote:
>> Quoting Taniya Das (2018-10-09 06:57:47)
>>> diff --git a/drivers/clk/qcom/dispcc-sdm845.c
>>> b/drivers/clk/qcom/dispcc-sdm845.c
>>> index 0cc4909..6d3136a 100644
>>> --- a/drivers/clk/qcom/dispcc-sdm845.c
>>> +++ b/drivers/clk/qcom/dispcc-sdm845.c
>>> @@ -128,6 +144,100 @@ enum {
>>> },
>>> };
>>>
>>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_aux_clk_src[] = {
>>> + F(19200000, P_BI_TCXO, 1, 0, 0),
>>> + { }
>>> +};
>>> +
>>> +static struct clk_rcg2 disp_cc_mdss_dp_aux_clk_src = {
>>> + .cmd_rcgr = 0x219c,
>>> + .mnd_width = 0,
>>> + .hid_width = 5,
>>> + .parent_map = disp_cc_parent_map_2,
>>> + .freq_tbl = ftbl_disp_cc_mdss_dp_aux_clk_src,
>>> + .clkr.hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_aux_clk_src",
>>> + .parent_names = disp_cc_parent_names_2,
>>> + .num_parents = 2,
>>> + .flags = CLK_SET_RATE_PARENT,
>>> + .ops = &clk_rcg2_ops,
>>> + },
>>> +};
>>> +
>>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_crypto_clk_src[] =
>>> {
>>> + F(108000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>>> + F(180000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>>> + F(360000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>>> + { }
>>> +};
>>> +
>>> +static struct clk_rcg2 disp_cc_mdss_dp_crypto_clk_src = {
>>> + .cmd_rcgr = 0x2154,
>>> + .mnd_width = 0,
>>> + .hid_width = 5,
>>> + .parent_map = disp_cc_parent_map_1,
>>> + .freq_tbl = ftbl_disp_cc_mdss_dp_crypto_clk_src,
>>> + .clkr.hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_crypto_clk_src",
>>> + .parent_names = disp_cc_parent_names_1,
>>> + .num_parents = 4,
>>> + .flags = CLK_GET_RATE_NOCACHE,
>>
>> Why?
>>
>>> + .ops = &clk_rcg2_ops,
>>> + },
>>> +};
>>> +
>>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
>>> + F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>> + F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>> + F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>
>> Are these in kHz? They really look like it and that's bad. Why do we
>> need them at all? Just to make sure the display driver picks these
>> exact
>> frequencies? It seems like we could just pass whatever number comes in
>> up to the parent and see what it can do.
>>
>
> Let me check back the reason we had to make this change.
We will need this flag since we reset/power-down the PLL every time we
disconnect/connect the DP cable or during suspend/resume. Only with this
flag, the calls to the PLL driver are properly called.
>
>>> + { }
>>> +};
>>> +
>>> +static struct clk_rcg2 disp_cc_mdss_dp_link_clk_src = {
>>> + .cmd_rcgr = 0x2138,
>>> + .mnd_width = 0,
>>> + .hid_width = 5,
>>> + .parent_map = disp_cc_parent_map_1,
>>> + .freq_tbl = ftbl_disp_cc_mdss_dp_link_clk_src,
>>> + .clkr.hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_link_clk_src",
>>> + .parent_names = disp_cc_parent_names_1,
>>> + .num_parents = 4,
>>> + .flags = CLK_SET_RATE_PARENT,
>>> + .ops = &clk_rcg2_ops,
>>> + },
>>> +};
>>> +
>>> +static struct clk_rcg2 disp_cc_mdss_dp_pixel1_clk_src = {
>>> + .cmd_rcgr = 0x2184,
>>> + .mnd_width = 16,
>>> + .hid_width = 5,
>>> + .parent_map = disp_cc_parent_map_1,
>>> + .clkr.hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_pixel1_clk_src",
>>> + .parent_names = disp_cc_parent_names_1,
>>> + .num_parents = 4,
>>> + .flags = CLK_SET_RATE_PARENT,
>>> + .ops = &clk_dp_ops,
>>> + },
>>> +};
>>> +
>>> +static struct clk_rcg2 disp_cc_mdss_dp_pixel_clk_src = {
>>> + .cmd_rcgr = 0x216c,
>>> + .mnd_width = 16,
>>> + .hid_width = 5,
>>> + .parent_map = disp_cc_parent_map_1,
>>> + .clkr.hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_pixel_clk_src",
>>> + .parent_names = disp_cc_parent_names_1,
>>> + .num_parents = 4,
>>> + .flags = CLK_SET_RATE_PARENT,
>>> + .ops = &clk_dp_ops,
>>> + },
>>> +};
>>> +
>>> static const struct freq_tbl ftbl_disp_cc_mdss_esc0_clk_src[] = {
>>> F(19200000, P_BI_TCXO, 1, 0, 0),
>>> { }
>>> @@ -391,6 +501,115 @@ enum {
>>> },
>>> };
>>>
>>> +static struct clk_branch disp_cc_mdss_dp_aux_clk = {
>>> + .halt_reg = 0x2054,
>>> + .halt_check = BRANCH_HALT,
>>> + .clkr = {
>>> + .enable_reg = 0x2054,
>>> + .enable_mask = BIT(0),
>>> + .hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_aux_clk",
>>> + .parent_names = (const char *[]){
>>> + "disp_cc_mdss_dp_aux_clk_src",
>>> + },
>>> + .num_parents = 1,
>>> + .flags = CLK_SET_RATE_PARENT,
>>> + .ops = &clk_branch2_ops,
>>> + },
>>> + },
>>> +};
>>> +
>>> +static struct clk_branch disp_cc_mdss_dp_crypto_clk = {
>>> + .halt_reg = 0x2048,
>>> + .halt_check = BRANCH_HALT,
>>> + .clkr = {
>>> + .enable_reg = 0x2048,
>>> + .enable_mask = BIT(0),
>>> + .hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_crypto_clk",
>>> + .parent_names = (const char *[]){
>>> + "disp_cc_mdss_dp_crypto_clk_src",
>>> + },
>>> + .num_parents = 1,
>>> + .flags = CLK_SET_RATE_PARENT,
>>> + .ops = &clk_branch2_ops,
>>> + },
>>> + },
>>> +};
>>> +
>>> +static struct clk_branch disp_cc_mdss_dp_link_clk = {
>>> + .halt_reg = 0x2040,
>>> + .halt_check = BRANCH_HALT,
>>> + .clkr = {
>>> + .enable_reg = 0x2040,
>>> + .enable_mask = BIT(0),
>>> + .hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_link_clk",
>>> + .parent_names = (const char *[]){
>>> + "disp_cc_mdss_dp_link_clk_src",
>>> + },
>>> + .num_parents = 1,
>>> + .flags = CLK_SET_RATE_PARENT,
>>> + .ops = &clk_branch2_ops,
>>> + },
>>> + },
>>> +};
>>> +
>>> +/* reset state of disp_cc_mdss_dp_link_div_clk_src divider is 0x3
>>> (div 4) */
>>
>> Not sure what this comment is for. But it's interesting nonetheless.
>>
>>> +static struct clk_branch disp_cc_mdss_dp_link_intf_clk = {
>>> + .halt_reg = 0x2044,
>>> + .halt_check = BRANCH_HALT,
>>> + .clkr = {
>>> + .enable_reg = 0x2044,
>>> + .enable_mask = BIT(0),
>>> + .hw.init = &(struct clk_init_data){
>>> + .name = "disp_cc_mdss_dp_link_intf_clk",
>>> + .parent_names = (const char *[]){
>>> + "disp_cc_mdss_dp_link_clk_src",
>>> + },
>>> + .num_parents = 1,
>>> + .flags = CLK_GET_RATE_NOCACHE,
>>
>> Why?
>>
>
> It was a requirement, but let me get back on this too.
>
I had a discussion with the Display Port teams and below is the requirement,
This flag is required since we reset/power-down the PLL every time they
disconnect/connect the DP cable or during suspend/resume.
Only with this flag, the calls to the PLL driver properly.
>>> + .ops = &clk_branch2_ops,
>>> + },
>>> + },
>>> +};
>>> +
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-28 10:34 ` Taniya Das
@ 2018-10-29 18:43 ` Stephen Boyd
2018-10-30 6:01 ` Taniya Das
2019-02-02 0:05 ` chandanu
0 siblings, 2 replies; 16+ messages in thread
From: Stephen Boyd @ 2018-10-29 18:43 UTC (permalink / raw)
To: Michael Turquette, Taniya Das
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu,
linux-arm-msm-owner
Quoting Taniya Das (2018-10-28 03:34:55)
> Hello Stephen,
>
> On 2018-10-19 16:04, Taniya Das wrote:
> > Hello Stephen,
> >
> > On 10/10/2018 2:04 AM, Stephen Boyd wrote:
> >> Quoting Taniya Das (2018-10-09 06:57:47)
> >>> diff --git a/drivers/clk/qcom/dispcc-sdm845.c
> >>> b/drivers/clk/qcom/dispcc-sdm845.c
> >>> index 0cc4909..6d3136a 100644
> >>> --- a/drivers/clk/qcom/dispcc-sdm845.c
> >>> +++ b/drivers/clk/qcom/dispcc-sdm845.c
> >>> @@ -128,6 +144,100 @@ enum {
> >>> },
> >>> };
> >>>
> >>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_aux_clk_src[] = {
> >>> + F(19200000, P_BI_TCXO, 1, 0, 0),
> >>> + { }
> >>> +};
> >>> +
> >>> +static struct clk_rcg2 disp_cc_mdss_dp_aux_clk_src = {
> >>> + .cmd_rcgr = 0x219c,
> >>> + .mnd_width = 0,
> >>> + .hid_width = 5,
> >>> + .parent_map = disp_cc_parent_map_2,
> >>> + .freq_tbl = ftbl_disp_cc_mdss_dp_aux_clk_src,
> >>> + .clkr.hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_aux_clk_src",
> >>> + .parent_names = disp_cc_parent_names_2,
> >>> + .num_parents = 2,
> >>> + .flags = CLK_SET_RATE_PARENT,
> >>> + .ops = &clk_rcg2_ops,
> >>> + },
> >>> +};
> >>> +
> >>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_crypto_clk_src[] =
> >>> {
> >>> + F(108000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
> >>> + F(180000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
> >>> + F(360000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
> >>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
> >>> + { }
> >>> +};
> >>> +
> >>> +static struct clk_rcg2 disp_cc_mdss_dp_crypto_clk_src = {
> >>> + .cmd_rcgr = 0x2154,
> >>> + .mnd_width = 0,
> >>> + .hid_width = 5,
> >>> + .parent_map = disp_cc_parent_map_1,
> >>> + .freq_tbl = ftbl_disp_cc_mdss_dp_crypto_clk_src,
> >>> + .clkr.hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_crypto_clk_src",
> >>> + .parent_names = disp_cc_parent_names_1,
> >>> + .num_parents = 4,
> >>> + .flags = CLK_GET_RATE_NOCACHE,
> >>
> >> Why?
> >>
> >>> + .ops = &clk_rcg2_ops,
> >>> + },
> >>> +};
> >>> +
> >>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
> >>> + F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> >>> + F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> >>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> >>> + F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> >>
> >> Are these in kHz? They really look like it and that's bad. Why do we
> >> need them at all? Just to make sure the display driver picks these
> >> exact
> >> frequencies? It seems like we could just pass whatever number comes in
> >> up to the parent and see what it can do.
> >>
> >
> > Let me check back the reason we had to make this change.
>
> We will need this flag since we reset/power-down the PLL every time we
> disconnect/connect the DP cable or during suspend/resume. Only with this
> flag, the calls to the PLL driver are properly called.
What does this mean? I wanted to know about the weird frequencies listed
above, and why it can't be done without a frequency table and direct
rates passed up to the parent.
>
> >
> >>> + { }
> >>> +};
> >>> +
> >>> +static struct clk_rcg2 disp_cc_mdss_dp_link_clk_src = {
> >>> + .cmd_rcgr = 0x2138,
> >>> + .mnd_width = 0,
> >>> + .hid_width = 5,
> >>> + .parent_map = disp_cc_parent_map_1,
> >>> + .freq_tbl = ftbl_disp_cc_mdss_dp_link_clk_src,
> >>> + .clkr.hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_link_clk_src",
> >>> + .parent_names = disp_cc_parent_names_1,
> >>> + .num_parents = 4,
> >>> + .flags = CLK_SET_RATE_PARENT,
> >>> + .ops = &clk_rcg2_ops,
> >>> + },
> >>> +};
> >>> +
> >>> +static struct clk_rcg2 disp_cc_mdss_dp_pixel1_clk_src = {
> >>> + .cmd_rcgr = 0x2184,
> >>> + .mnd_width = 16,
> >>> + .hid_width = 5,
> >>> + .parent_map = disp_cc_parent_map_1,
> >>> + .clkr.hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_pixel1_clk_src",
> >>> + .parent_names = disp_cc_parent_names_1,
> >>> + .num_parents = 4,
> >>> + .flags = CLK_SET_RATE_PARENT,
> >>> + .ops = &clk_dp_ops,
> >>> + },
> >>> +};
> >>> +
> >>> +static struct clk_rcg2 disp_cc_mdss_dp_pixel_clk_src = {
> >>> + .cmd_rcgr = 0x216c,
> >>> + .mnd_width = 16,
> >>> + .hid_width = 5,
> >>> + .parent_map = disp_cc_parent_map_1,
> >>> + .clkr.hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_pixel_clk_src",
> >>> + .parent_names = disp_cc_parent_names_1,
> >>> + .num_parents = 4,
> >>> + .flags = CLK_SET_RATE_PARENT,
> >>> + .ops = &clk_dp_ops,
> >>> + },
> >>> +};
> >>> +
> >>> static const struct freq_tbl ftbl_disp_cc_mdss_esc0_clk_src[] = {
> >>> F(19200000, P_BI_TCXO, 1, 0, 0),
> >>> { }
> >>> @@ -391,6 +501,115 @@ enum {
> >>> },
> >>> };
> >>>
> >>> +static struct clk_branch disp_cc_mdss_dp_aux_clk = {
> >>> + .halt_reg = 0x2054,
> >>> + .halt_check = BRANCH_HALT,
> >>> + .clkr = {
> >>> + .enable_reg = 0x2054,
> >>> + .enable_mask = BIT(0),
> >>> + .hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_aux_clk",
> >>> + .parent_names = (const char *[]){
> >>> + "disp_cc_mdss_dp_aux_clk_src",
> >>> + },
> >>> + .num_parents = 1,
> >>> + .flags = CLK_SET_RATE_PARENT,
> >>> + .ops = &clk_branch2_ops,
> >>> + },
> >>> + },
> >>> +};
> >>> +
> >>> +static struct clk_branch disp_cc_mdss_dp_crypto_clk = {
> >>> + .halt_reg = 0x2048,
> >>> + .halt_check = BRANCH_HALT,
> >>> + .clkr = {
> >>> + .enable_reg = 0x2048,
> >>> + .enable_mask = BIT(0),
> >>> + .hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_crypto_clk",
> >>> + .parent_names = (const char *[]){
> >>> + "disp_cc_mdss_dp_crypto_clk_src",
> >>> + },
> >>> + .num_parents = 1,
> >>> + .flags = CLK_SET_RATE_PARENT,
> >>> + .ops = &clk_branch2_ops,
> >>> + },
> >>> + },
> >>> +};
> >>> +
> >>> +static struct clk_branch disp_cc_mdss_dp_link_clk = {
> >>> + .halt_reg = 0x2040,
> >>> + .halt_check = BRANCH_HALT,
> >>> + .clkr = {
> >>> + .enable_reg = 0x2040,
> >>> + .enable_mask = BIT(0),
> >>> + .hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_link_clk",
> >>> + .parent_names = (const char *[]){
> >>> + "disp_cc_mdss_dp_link_clk_src",
> >>> + },
> >>> + .num_parents = 1,
> >>> + .flags = CLK_SET_RATE_PARENT,
> >>> + .ops = &clk_branch2_ops,
> >>> + },
> >>> + },
> >>> +};
> >>> +
> >>> +/* reset state of disp_cc_mdss_dp_link_div_clk_src divider is 0x3
> >>> (div 4) */
> >>
> >> Not sure what this comment is for. But it's interesting nonetheless.
> >>
> >>> +static struct clk_branch disp_cc_mdss_dp_link_intf_clk = {
> >>> + .halt_reg = 0x2044,
> >>> + .halt_check = BRANCH_HALT,
> >>> + .clkr = {
> >>> + .enable_reg = 0x2044,
> >>> + .enable_mask = BIT(0),
> >>> + .hw.init = &(struct clk_init_data){
> >>> + .name = "disp_cc_mdss_dp_link_intf_clk",
> >>> + .parent_names = (const char *[]){
> >>> + "disp_cc_mdss_dp_link_clk_src",
> >>> + },
> >>> + .num_parents = 1,
> >>> + .flags = CLK_GET_RATE_NOCACHE,
> >>
> >> Why?
> >>
> >
> > It was a requirement, but let me get back on this too.
> >
> I had a discussion with the Display Port teams and below is the requirement,
>
> This flag is required since we reset/power-down the PLL every time they
> disconnect/connect the DP cable or during suspend/resume.
> Only with this flag, the calls to the PLL driver properly.
Ok. So that explains the get rate nocache flag. Can you please add a
comment that explains that these clk registers here are lost across
suspend/resume of the display device? It really sounds like these
display clks are inside of the display power domain and thus they lose
their state across the display power domain power down. It would be
better if we could properly implement suspend/restore for these clk
registers across suspend/resume of the display device so that we don't
need this nocache flag and the display code can work together with the
clk code here to restore the frequency to the clk.
Is it really the case that the rcg here is always selecting a particular
PLL and doing a div-1? Because that is very simple then to just write
that setting again on genpd restore.
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-29 18:43 ` Stephen Boyd
@ 2018-10-30 6:01 ` Taniya Das
2018-10-30 16:33 ` Stephen Boyd
2019-02-02 0:05 ` chandanu
1 sibling, 1 reply; 16+ messages in thread
From: Taniya Das @ 2018-10-30 6:01 UTC (permalink / raw)
To: Stephen Boyd, Michael Turquette
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu,
linux-arm-msm-owner
On 10/30/2018 12:13 AM, Stephen Boyd wrote:
> Quoting Taniya Das (2018-10-28 03:34:55)
>> Hello Stephen,
>>
>> On 2018-10-19 16:04, Taniya Das wrote:
>>> Hello Stephen,
>>>
>>> On 10/10/2018 2:04 AM, Stephen Boyd wrote:
>>>> Quoting Taniya Das (2018-10-09 06:57:47)
>>>>> diff --git a/drivers/clk/qcom/dispcc-sdm845.c
>>>>> b/drivers/clk/qcom/dispcc-sdm845.c
>>>>> index 0cc4909..6d3136a 100644
>>>>> --- a/drivers/clk/qcom/dispcc-sdm845.c
>>>>> +++ b/drivers/clk/qcom/dispcc-sdm845.c
>>>>> @@ -128,6 +144,100 @@ enum {
>>>>> },
>>>>> };
>>>>>
>>>>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_aux_clk_src[] = {
>>>>> + F(19200000, P_BI_TCXO, 1, 0, 0),
>>>>> + { }
>>>>> +};
>>>>> +
>>>>> +static struct clk_rcg2 disp_cc_mdss_dp_aux_clk_src = {
>>>>> + .cmd_rcgr = 0x219c,
>>>>> + .mnd_width = 0,
>>>>> + .hid_width = 5,
>>>>> + .parent_map = disp_cc_parent_map_2,
>>>>> + .freq_tbl = ftbl_disp_cc_mdss_dp_aux_clk_src,
>>>>> + .clkr.hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_aux_clk_src",
>>>>> + .parent_names = disp_cc_parent_names_2,
>>>>> + .num_parents = 2,
>>>>> + .flags = CLK_SET_RATE_PARENT,
>>>>> + .ops = &clk_rcg2_ops,
>>>>> + },
>>>>> +};
>>>>> +
>>>>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_crypto_clk_src[] =
>>>>> {
>>>>> + F(108000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>>>>> + F(180000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>>>>> + F(360000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>>>>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 3, 0, 0),
>>>>> + { }
>>>>> +};
>>>>> +
>>>>> +static struct clk_rcg2 disp_cc_mdss_dp_crypto_clk_src = {
>>>>> + .cmd_rcgr = 0x2154,
>>>>> + .mnd_width = 0,
>>>>> + .hid_width = 5,
>>>>> + .parent_map = disp_cc_parent_map_1,
>>>>> + .freq_tbl = ftbl_disp_cc_mdss_dp_crypto_clk_src,
>>>>> + .clkr.hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_crypto_clk_src",
>>>>> + .parent_names = disp_cc_parent_names_1,
>>>>> + .num_parents = 4,
>>>>> + .flags = CLK_GET_RATE_NOCACHE,
>>>>
>>>> Why?
>>>>
>>>>> + .ops = &clk_rcg2_ops,
>>>>> + },
>>>>> +};
>>>>> +
>>>>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
>>>>> + F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>>>> + F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>>>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>>>> + F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>>>
>>>> Are these in kHz? They really look like it and that's bad. Why do we
>>>> need them at all? Just to make sure the display driver picks these
>>>> exact
>>>> frequencies? It seems like we could just pass whatever number comes in
>>>> up to the parent and see what it can do.
>>>>
>>>
>>> Let me check back the reason we had to make this change.
>>
>> We will need this flag since we reset/power-down the PLL every time we
>> disconnect/connect the DP cable or during suspend/resume. Only with this
>> flag, the calls to the PLL driver are properly called.
>
> What does this mean? I wanted to know about the weird frequencies listed
> above, and why it can't be done without a frequency table and direct
> rates passed up to the parent.
>
OOps, my bad :(.
We added these changes to handle higher clock rates. These rates when
greater than 4.3Ghz cannot be represented in 32bit variables. For DP, we
already have 5.4G and 8.1GHz freq for VCO clock. We will need these Khz
freq list in clock driver.
Let me check if they can do something like the byte/pixel clocks of
display.
>>
>>>
>>>>> + { }
>>>>> +};
>>>>> +
>>>>> +static struct clk_rcg2 disp_cc_mdss_dp_link_clk_src = {
>>>>> + .cmd_rcgr = 0x2138,
>>>>> + .mnd_width = 0,
>>>>> + .hid_width = 5,
>>>>> + .parent_map = disp_cc_parent_map_1,
>>>>> + .freq_tbl = ftbl_disp_cc_mdss_dp_link_clk_src,
>>>>> + .clkr.hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_link_clk_src",
>>>>> + .parent_names = disp_cc_parent_names_1,
>>>>> + .num_parents = 4,
>>>>> + .flags = CLK_SET_RATE_PARENT,
>>>>> + .ops = &clk_rcg2_ops,
>>>>> + },
>>>>> +};
>>>>> +
>>>>> +static struct clk_rcg2 disp_cc_mdss_dp_pixel1_clk_src = {
>>>>> + .cmd_rcgr = 0x2184,
>>>>> + .mnd_width = 16,
>>>>> + .hid_width = 5,
>>>>> + .parent_map = disp_cc_parent_map_1,
>>>>> + .clkr.hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_pixel1_clk_src",
>>>>> + .parent_names = disp_cc_parent_names_1,
>>>>> + .num_parents = 4,
>>>>> + .flags = CLK_SET_RATE_PARENT,
>>>>> + .ops = &clk_dp_ops,
>>>>> + },
>>>>> +};
>>>>> +
>>>>> +static struct clk_rcg2 disp_cc_mdss_dp_pixel_clk_src = {
>>>>> + .cmd_rcgr = 0x216c,
>>>>> + .mnd_width = 16,
>>>>> + .hid_width = 5,
>>>>> + .parent_map = disp_cc_parent_map_1,
>>>>> + .clkr.hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_pixel_clk_src",
>>>>> + .parent_names = disp_cc_parent_names_1,
>>>>> + .num_parents = 4,
>>>>> + .flags = CLK_SET_RATE_PARENT,
>>>>> + .ops = &clk_dp_ops,
>>>>> + },
>>>>> +};
>>>>> +
>>>>> static const struct freq_tbl ftbl_disp_cc_mdss_esc0_clk_src[] = {
>>>>> F(19200000, P_BI_TCXO, 1, 0, 0),
>>>>> { }
>>>>> @@ -391,6 +501,115 @@ enum {
>>>>> },
>>>>> };
>>>>>
>>>>> +static struct clk_branch disp_cc_mdss_dp_aux_clk = {
>>>>> + .halt_reg = 0x2054,
>>>>> + .halt_check = BRANCH_HALT,
>>>>> + .clkr = {
>>>>> + .enable_reg = 0x2054,
>>>>> + .enable_mask = BIT(0),
>>>>> + .hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_aux_clk",
>>>>> + .parent_names = (const char *[]){
>>>>> + "disp_cc_mdss_dp_aux_clk_src",
>>>>> + },
>>>>> + .num_parents = 1,
>>>>> + .flags = CLK_SET_RATE_PARENT,
>>>>> + .ops = &clk_branch2_ops,
>>>>> + },
>>>>> + },
>>>>> +};
>>>>> +
>>>>> +static struct clk_branch disp_cc_mdss_dp_crypto_clk = {
>>>>> + .halt_reg = 0x2048,
>>>>> + .halt_check = BRANCH_HALT,
>>>>> + .clkr = {
>>>>> + .enable_reg = 0x2048,
>>>>> + .enable_mask = BIT(0),
>>>>> + .hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_crypto_clk",
>>>>> + .parent_names = (const char *[]){
>>>>> + "disp_cc_mdss_dp_crypto_clk_src",
>>>>> + },
>>>>> + .num_parents = 1,
>>>>> + .flags = CLK_SET_RATE_PARENT,
>>>>> + .ops = &clk_branch2_ops,
>>>>> + },
>>>>> + },
>>>>> +};
>>>>> +
>>>>> +static struct clk_branch disp_cc_mdss_dp_link_clk = {
>>>>> + .halt_reg = 0x2040,
>>>>> + .halt_check = BRANCH_HALT,
>>>>> + .clkr = {
>>>>> + .enable_reg = 0x2040,
>>>>> + .enable_mask = BIT(0),
>>>>> + .hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_link_clk",
>>>>> + .parent_names = (const char *[]){
>>>>> + "disp_cc_mdss_dp_link_clk_src",
>>>>> + },
>>>>> + .num_parents = 1,
>>>>> + .flags = CLK_SET_RATE_PARENT,
>>>>> + .ops = &clk_branch2_ops,
>>>>> + },
>>>>> + },
>>>>> +};
>>>>> +
>>>>> +/* reset state of disp_cc_mdss_dp_link_div_clk_src divider is 0x3
>>>>> (div 4) */
>>>>
>>>> Not sure what this comment is for. But it's interesting nonetheless.
>>>>
>>>>> +static struct clk_branch disp_cc_mdss_dp_link_intf_clk = {
>>>>> + .halt_reg = 0x2044,
>>>>> + .halt_check = BRANCH_HALT,
>>>>> + .clkr = {
>>>>> + .enable_reg = 0x2044,
>>>>> + .enable_mask = BIT(0),
>>>>> + .hw.init = &(struct clk_init_data){
>>>>> + .name = "disp_cc_mdss_dp_link_intf_clk",
>>>>> + .parent_names = (const char *[]){
>>>>> + "disp_cc_mdss_dp_link_clk_src",
>>>>> + },
>>>>> + .num_parents = 1,
>>>>> + .flags = CLK_GET_RATE_NOCACHE,
>>>>
>>>> Why?
>>>>
>>>
>>> It was a requirement, but let me get back on this too.
>>>
>> I had a discussion with the Display Port teams and below is the requirement,
>>
>> This flag is required since we reset/power-down the PLL every time they
>> disconnect/connect the DP cable or during suspend/resume.
>> Only with this flag, the calls to the PLL driver properly.
>
> Ok. So that explains the get rate nocache flag. Can you please add a
> comment that explains that these clk registers here are lost across
> suspend/resume of the display device? It really sounds like these
> display clks are inside of the display power domain and thus they lose
> their state across the display power domain power down. It would be
> better if we could properly implement suspend/restore for these clk
> registers across suspend/resume of the display device so that we don't
> need this nocache flag and the display code can work together with the
> clk code here to restore the frequency to the clk.
>
> Is it really the case that the rcg here is always selecting a particular
> PLL and doing a div-1? Because that is very simple then to just write
> that setting again on genpd restore.
>
--
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation.
--
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-30 6:01 ` Taniya Das
@ 2018-10-30 16:33 ` Stephen Boyd
2018-11-01 5:02 ` Taniya Das
0 siblings, 1 reply; 16+ messages in thread
From: Stephen Boyd @ 2018-10-30 16:33 UTC (permalink / raw)
To: Michael Turquette, Taniya Das
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu,
linux-arm-msm-owner
Quoting Taniya Das (2018-10-29 23:01:44)
> On 10/30/2018 12:13 AM, Stephen Boyd wrote:
> > Quoting Taniya Das (2018-10-28 03:34:55)
> >> On 2018-10-19 16:04, Taniya Das wrote:
> >>> On 10/10/2018 2:04 AM, Stephen Boyd wrote:
> >>>> Quoting Taniya Das (2018-10-09 06:57:47)
> >>>>> diff --git a/drivers/clk/qcom/dispcc-sdm845.c
> >>>>> b/drivers/clk/qcom/dispcc-sdm845.c
> >>>>> index 0cc4909..6d3136a 100644
> >>>>> --- a/drivers/clk/qcom/dispcc-sdm845.c
> >>>>> +++ b/drivers/clk/qcom/dispcc-sdm845.c
> >>>>> + },
> >>>>> +};
> >>>>> +
> >>>>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
> >>>>> + F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> >>>>> + F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> >>>>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> >>>>> + F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
> >>>>
> >>>> Are these in kHz? They really look like it and that's bad. Why do we
> >>>> need them at all? Just to make sure the display driver picks these
> >>>> exact
> >>>> frequencies? It seems like we could just pass whatever number comes in
> >>>> up to the parent and see what it can do.
> >>>>
> >>>
> >>> Let me check back the reason we had to make this change.
> >>
> >> We will need this flag since we reset/power-down the PLL every time we
> >> disconnect/connect the DP cable or during suspend/resume. Only with this
> >> flag, the calls to the PLL driver are properly called.
> >
> > What does this mean? I wanted to know about the weird frequencies listed
> > above, and why it can't be done without a frequency table and direct
> > rates passed up to the parent.
> >
>
> OOps, my bad :(.
>
> We added these changes to handle higher clock rates. These rates when
> greater than 4.3Ghz cannot be represented in 32bit variables. For DP, we
> already have 5.4G and 8.1GHz freq for VCO clock. We will need these Khz
> freq list in clock driver.
> Let me check if they can do something like the byte/pixel clocks of
> display.
Well then we really should just throw away the freq table here and have
rcg ops that pass the frequency up to the display PLL. Also, those
numbers look like gigabits per second (Gbit/s) for the DP spec which
isn't exactly the same as a clk frequency. What frequency does the PLL
run at for these various DP link speeds?
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-30 16:33 ` Stephen Boyd
@ 2018-11-01 5:02 ` Taniya Das
2018-11-06 17:08 ` Stephen Boyd
0 siblings, 1 reply; 16+ messages in thread
From: Taniya Das @ 2018-11-01 5:02 UTC (permalink / raw)
To: Stephen Boyd, Michael Turquette, chandanu
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu,
linux-arm-msm-owner
+ Chandan from Display Port team,
On 10/30/2018 10:03 PM, Stephen Boyd wrote:
> Quoting Taniya Das (2018-10-29 23:01:44)
>> On 10/30/2018 12:13 AM, Stephen Boyd wrote:
>>> Quoting Taniya Das (2018-10-28 03:34:55)
>>>> On 2018-10-19 16:04, Taniya Das wrote:
>>>>> On 10/10/2018 2:04 AM, Stephen Boyd wrote:
>>>>>> Quoting Taniya Das (2018-10-09 06:57:47)
>>>>>>> diff --git a/drivers/clk/qcom/dispcc-sdm845.c
>>>>>>> b/drivers/clk/qcom/dispcc-sdm845.c
>>>>>>> index 0cc4909..6d3136a 100644
>>>>>>> --- a/drivers/clk/qcom/dispcc-sdm845.c
>>>>>>> +++ b/drivers/clk/qcom/dispcc-sdm845.c
>>>>>>> + },
>>>>>>> +};
>>>>>>> +
>>>>>>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
>>>>>>> + F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>>>>>> + F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>>>>>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>>>>>> + F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>>>>>>
>>>>>> Are these in kHz? They really look like it and that's bad. Why do we
>>>>>> need them at all? Just to make sure the display driver picks these
>>>>>> exact
>>>>>> frequencies? It seems like we could just pass whatever number comes in
>>>>>> up to the parent and see what it can do.
>>>>>>
>>>>>
>>>>> Let me check back the reason we had to make this change.
>>>>
>>>> We will need this flag since we reset/power-down the PLL every time we
>>>> disconnect/connect the DP cable or during suspend/resume. Only with this
>>>> flag, the calls to the PLL driver are properly called.
>>>
>>> What does this mean? I wanted to know about the weird frequencies listed
>>> above, and why it can't be done without a frequency table and direct
>>> rates passed up to the parent.
>>>
>>
>> OOps, my bad :(.
>>
>> We added these changes to handle higher clock rates. These rates when
>> greater than 4.3Ghz cannot be represented in 32bit variables. For DP, we
>> already have 5.4G and 8.1GHz freq for VCO clock. We will need these Khz
>> freq list in clock driver.
>> Let me check if they can do something like the byte/pixel clocks of
>> display.
>
> Well then we really should just throw away the freq table here and have
> rcg ops that pass the frequency up to the display PLL. Also, those
> numbers look like gigabits per second (Gbit/s) for the DP spec which
> isn't exactly the same as a clk frequency. What frequency does the PLL
> run at for these various DP link speeds?
>
Could you please help with the above query from Stephen?
--
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation.
--
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-11-01 5:02 ` Taniya Das
@ 2018-11-06 17:08 ` Stephen Boyd
2018-11-12 3:41 ` chandanu
0 siblings, 1 reply; 16+ messages in thread
From: Stephen Boyd @ 2018-11-06 17:08 UTC (permalink / raw)
To: Michael Turquette, Taniya Das, chandanu
Cc: Andy Gross, David Brown, Rajendra Nayak, linux-arm-msm,
linux-soc, linux-clk, linux-kernel, chandanu,
linux-arm-msm-owner
Quoting Taniya Das (2018-10-31 22:02:22)
> + Chandan from Display Port team,
>
> On 10/30/2018 10:03 PM, Stephen Boyd wrote:
> > Also, those
> > numbers look like gigabits per second (Gbit/s) for the DP spec which
> > isn't exactly the same as a clk frequency. What frequency does the PLL
> > run at for these various DP link speeds?
> >
> Could you please help with the above query from Stephen?
Can I safely assume that it matches the link rate shown on Wikipedia for
display port[1]? I.e.
RBR (Reduced Bit Rate): 1.62 Gbit/s bandwidth per lane (162 MHz link
symbol rate)
HBR (High Bit Rate): 2.70 Gbit/s bandwidth per lane (270 MHz link
symbol rate)
HBR2 (High Bit Rate 2): 5.40 Gbit/s bandwidth per lane (540 MHz link
symbol rate), introduced in DP 1.2
HBR3 (High Bit Rate 3): 8.10 Gbit/s bandwidth per lane (810 MHz link
symbol rate), introduced in DP 1.3
So then they're MHz but the table is written in kHz when it should be
written in Hz. Either way, the table can be removed and then we just
need to fix the DP PHY PLL code to accept Hz instead of kHz.
[1] https://en.wikipedia.org/wiki/DisplayPort#Main_link
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-11-06 17:08 ` Stephen Boyd
@ 2018-11-12 3:41 ` chandanu
0 siblings, 0 replies; 16+ messages in thread
From: chandanu @ 2018-11-12 3:41 UTC (permalink / raw)
To: Stephen Boyd
Cc: Michael Turquette, Taniya Das, Andy Gross, David Brown,
Rajendra Nayak, linux-arm-msm, linux-soc, linux-clk,
linux-kernel, linux-arm-msm-owner
On 2018-11-06 09:08, Stephen Boyd wrote:
> Quoting Taniya Das (2018-10-31 22:02:22)
>> + Chandan from Display Port team,
>>
>> On 10/30/2018 10:03 PM, Stephen Boyd wrote:
>> > Also, those
>> > numbers look like gigabits per second (Gbit/s) for the DP spec which
>> > isn't exactly the same as a clk frequency. What frequency does the PLL
>> > run at for these various DP link speeds?
>> >
>> Could you please help with the above query from Stephen?
Hello Stephen,
For DP link speed of 5.4Gbit/s, the PLL will be running at 10.8 Ghz. For
all
the other DP link speeds, the PLL will be running at 8.1 Ghz.
>
> Can I safely assume that it matches the link rate shown on Wikipedia
> for
> display port[1]? I.e.
>
> RBR (Reduced Bit Rate): 1.62 Gbit/s bandwidth per lane (162 MHz link
> symbol rate)
> HBR (High Bit Rate): 2.70 Gbit/s bandwidth per lane (270 MHz link
> symbol rate)
> HBR2 (High Bit Rate 2): 5.40 Gbit/s bandwidth per lane (540 MHz link
> symbol rate), introduced in DP 1.2
> HBR3 (High Bit Rate 3): 8.10 Gbit/s bandwidth per lane (810 MHz link
> symbol rate), introduced in DP 1.3
>
> So then they're MHz but the table is written in kHz when it should be
> written in Hz. Either way, the table can be removed and then we just
> need to fix the DP PHY PLL code to accept Hz instead of kHz.
>
> [1] https://en.wikipedia.org/wiki/DisplayPort#Main_link
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2018-10-29 18:43 ` Stephen Boyd
2018-10-30 6:01 ` Taniya Das
@ 2019-02-02 0:05 ` chandanu
2019-07-15 22:59 ` Stephen Boyd
1 sibling, 1 reply; 16+ messages in thread
From: chandanu @ 2019-02-02 0:05 UTC (permalink / raw)
To: Stephen Boyd
Cc: Michael Turquette, Taniya Das, Andy Gross, David Brown,
Rajendra Nayak, linux-arm-msm, linux-soc, linux-clk,
linux-kernel, linux-arm-msm-owner
Hello Stephen,
On 2018-10-29 11:43, Stephen Boyd wrote:
> Quoting Taniya Das (2018-10-28 03:34:55)
>> Hello Stephen,
>>
>> On 2018-10-19 16:04, Taniya Das wrote:
>> > Hello Stephen,
>> >
/snip
>> >>> +static struct clk_rcg2 disp_cc_mdss_dp_crypto_clk_src = {
>> >>> + .cmd_rcgr = 0x2154,
>> >>> + .mnd_width = 0,
>> >>> + .hid_width = 5,
>> >>> + .parent_map = disp_cc_parent_map_1,
>> >>> + .freq_tbl = ftbl_disp_cc_mdss_dp_crypto_clk_src,
>> >>> + .clkr.hw.init = &(struct clk_init_data){
>> >>> + .name = "disp_cc_mdss_dp_crypto_clk_src",
>> >>> + .parent_names = disp_cc_parent_names_1,
>> >>> + .num_parents = 4,
>> >>> + .flags = CLK_GET_RATE_NOCACHE,
>> >>
>> >> Why?
>> >>
>> >>> + .ops = &clk_rcg2_ops,
>> >>> + },
>> >>> +};
>> >>> +
>> >>> +static const struct freq_tbl ftbl_disp_cc_mdss_dp_link_clk_src[] = {
>> >>> + F(162000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>> >>> + F(270000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>> >>> + F(540000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>> >>> + F(810000, P_DP_PHY_PLL_LINK_CLK, 1, 0, 0),
>> >>
>> >> Are these in kHz? They really look like it and that's bad. Why do we
>> >> need them at all? Just to make sure the display driver picks these
>> >> exact
>> >> frequencies? It seems like we could just pass whatever number comes in
>> >> up to the parent and see what it can do.
>> >>
>> >
>> > Let me check back the reason we had to make this change.
>>
>> We will need this flag since we reset/power-down the PLL every time we
>> disconnect/connect the DP cable or during suspend/resume. Only with
>> this
>> flag, the calls to the PLL driver are properly called.
>
> What does this mean? I wanted to know about the weird frequencies
> listed
> above, and why it can't be done without a frequency table and direct
> rates passed up to the parent.
>
/snip
>> >>
>> >>> +static struct clk_branch disp_cc_mdss_dp_link_intf_clk = {
>> >>> + .halt_reg = 0x2044,
>> >>> + .halt_check = BRANCH_HALT,
>> >>> + .clkr = {
>> >>> + .enable_reg = 0x2044,
>> >>> + .enable_mask = BIT(0),
>> >>> + .hw.init = &(struct clk_init_data){
>> >>> + .name = "disp_cc_mdss_dp_link_intf_clk",
>> >>> + .parent_names = (const char *[]){
>> >>> + "disp_cc_mdss_dp_link_clk_src",
>> >>> + },
>> >>> + .num_parents = 1,
>> >>> + .flags = CLK_GET_RATE_NOCACHE,
>> >>
>> >> Why?
>> >>
>> >
>> > It was a requirement, but let me get back on this too.
>> >
>> I had a discussion with the Display Port teams and below is the
>> requirement,
>>
>> This flag is required since we reset/power-down the PLL every time
>> they
>> disconnect/connect the DP cable or during suspend/resume.
>> Only with this flag, the calls to the PLL driver properly.
>
> Ok. So that explains the get rate nocache flag. Can you please add a
> comment that explains that these clk registers here are lost across
> suspend/resume of the display device? It really sounds like these
> display clks are inside of the display power domain and thus they lose
> their state across the display power domain power down. It would be
> better if we could properly implement suspend/restore for these clk
> registers across suspend/resume of the display device so that we don't
> need this nocache flag and the display code can work together with the
> clk code here to restore the frequency to the clk.
We already handle the suspend/restore for these clk registers
in Dp PLL domain. Without the "NOCACHE_FLAG", and if we are requesting
the same clock rate
for any of the clocks, the set_rate call never reaches the DP PLL Ops.
I am not clear on what you are suggesting for removing the
"NOCACHE_FLAG" for
the DisplayPort clocks. Are you suggesting design changes in DP PLL
driver or in dispcc-driver?
Can you please provide more details?
thanks
Chandan
>
> Is it really the case that the rcg here is always selecting a
> particular
> PLL and doing a div-1? Because that is very simple then to just write
> that setting again on genpd restore.
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks
2019-02-02 0:05 ` chandanu
@ 2019-07-15 22:59 ` Stephen Boyd
0 siblings, 0 replies; 16+ messages in thread
From: Stephen Boyd @ 2019-07-15 22:59 UTC (permalink / raw)
To: chandanu
Cc: Michael Turquette, Taniya Das, Andy Gross, David Brown,
Rajendra Nayak, linux-arm-msm, linux-soc, linux-clk,
linux-kernel, linux-arm-msm-owner
Quoting chandanu@codeaurora.org (2019-02-01 16:05:55)
> On 2018-10-29 11:43, Stephen Boyd wrote:
> > Quoting Taniya Das (2018-10-28 03:34:55)
> >> On 2018-10-19 16:04, Taniya Das wrote:
>
> >> >>
> >> >>> +static struct clk_branch disp_cc_mdss_dp_link_intf_clk = {
> >> >>> + .halt_reg = 0x2044,
> >> >>> + .halt_check = BRANCH_HALT,
> >> >>> + .clkr = {
> >> >>> + .enable_reg = 0x2044,
> >> >>> + .enable_mask = BIT(0),
> >> >>> + .hw.init = &(struct clk_init_data){
> >> >>> + .name = "disp_cc_mdss_dp_link_intf_clk",
> >> >>> + .parent_names = (const char *[]){
> >> >>> + "disp_cc_mdss_dp_link_clk_src",
> >> >>> + },
> >> >>> + .num_parents = 1,
> >> >>> + .flags = CLK_GET_RATE_NOCACHE,
> >> >>
> >> >> Why?
> >> >>
> >> >
> >> > It was a requirement, but let me get back on this too.
> >> >
> >> I had a discussion with the Display Port teams and below is the
> >> requirement,
> >>
> >> This flag is required since we reset/power-down the PLL every time
> >> they
> >> disconnect/connect the DP cable or during suspend/resume.
> >> Only with this flag, the calls to the PLL driver properly.
> >
> > Ok. So that explains the get rate nocache flag. Can you please add a
> > comment that explains that these clk registers here are lost across
> > suspend/resume of the display device? It really sounds like these
> > display clks are inside of the display power domain and thus they lose
> > their state across the display power domain power down. It would be
> > better if we could properly implement suspend/restore for these clk
> > registers across suspend/resume of the display device so that we don't
> > need this nocache flag and the display code can work together with the
> > clk code here to restore the frequency to the clk.
>
This patch came again and it didn't have any comments to this effect in
the code around the flag.
>
> We already handle the suspend/restore for these clk registers
> in Dp PLL domain. Without the "NOCACHE_FLAG", and if we are requesting
> the same clock rate
> for any of the clocks, the set_rate call never reaches the DP PLL Ops.
So do you restore the frequency of the PLL manually? Or that is done by
calling clk_set_rate() on the leaf clk again?
>
> I am not clear on what you are suggesting for removing the
> "NOCACHE_FLAG" for
> the DisplayPort clocks. Are you suggesting design changes in DP PLL
> driver or in dispcc-driver?
> Can you please provide more details?
>
I'm suggesting that the clk framework needs to be told that the PLL has
lost the rate and thus should do a save/restore of the registers so that
the clk framework can be back in sync with the clk hardware. Maybe it's
as simple as calling clk_set_rate(&pll, XO_RATE) or clk_set_parent(&pll,
&xo_clk), so that we can recalc the rate down the tree and fix up the
child clk frequencies. Or, maybe we need to add some sort of mechanism
to the clk framework so it can be told that the frequency here has
changed. Or we need to add a hook in the power domain for the DP PLL to
tell the clk framework that the clk has changed rate and thus should
recalc down to the children. Something like this, instead of an obtuse
flag that tells us very little about what's going on.
^ permalink raw reply [flat|nested] 16+ messages in thread
end of thread, other threads:[~2019-07-15 22:59 UTC | newest]
Thread overview: 16+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-10-09 13:57 [PATCH v1 0/2] Add support for display port clocks and clock ops Taniya Das
2018-10-09 13:57 ` [PATCH v1 1/2] clk: qcom: rcg2: Add support for display port " Taniya Das
2018-10-09 20:46 ` Stephen Boyd
2018-10-19 10:31 ` Taniya Das
2018-10-09 13:57 ` [PATCH v1 2/2] clk: qcom : dispcc: Add support for display port clocks Taniya Das
2018-10-09 20:34 ` Stephen Boyd
2018-10-19 10:34 ` Taniya Das
2018-10-28 10:34 ` Taniya Das
2018-10-29 18:43 ` Stephen Boyd
2018-10-30 6:01 ` Taniya Das
2018-10-30 16:33 ` Stephen Boyd
2018-11-01 5:02 ` Taniya Das
2018-11-06 17:08 ` Stephen Boyd
2018-11-12 3:41 ` chandanu
2019-02-02 0:05 ` chandanu
2019-07-15 22:59 ` Stephen Boyd
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).