Re: [PATCH v6 07/11] drm/hisilicon: Add designware dsi encoder driver

[Archit Taneja <architt@codeaurora.org>]

On 3/1/2016 4:03 PM, Xinliang Liu wrote:
> Hi,
>
> On 1 March 2016 at 02:49, Archit Taneja <architt@codeaurora.org> wrote:
>>
>>
>> On 2/26/2016 2:10 PM, Xinliang Liu wrote:
>>>
>>> Add DesignWare MIPI DSI Host Controller v1.02 encoder driver
>>> for hi6220 SoC.
>>>
>>> v6:
>>> - Change "pclk_dsi" to "pclk".
>>> v5: None.
>>> v4: None.
>>> v3:
>>> - Rename file name to dw_drm_dsi.c
>>> - Make encoder type as DRM_MODE_ENCODER_DSI.
>>> - A few cleanup.
>>> v2:
>>> - Remove abtraction layer.
>>>
>>> Signed-off-by: Xinliang Liu <xinliang.liu@linaro.org>
>>> Signed-off-by: Xinwei Kong <kong.kongxinwei@hisilicon.com>
>>> Signed-off-by: Andy Green <andy.green@linaro.org>
>>> ---
>>>    drivers/gpu/drm/hisilicon/kirin/Kconfig      |   1 +
>>>    drivers/gpu/drm/hisilicon/kirin/Makefile     |   3 +-
>>>    drivers/gpu/drm/hisilicon/kirin/dw_drm_dsi.c | 743
>>> +++++++++++++++++++++++++++
>>>    drivers/gpu/drm/hisilicon/kirin/dw_dsi_reg.h |  83 +++
>>>    4 files changed, 829 insertions(+), 1 deletion(-)
>>>    create mode 100644 drivers/gpu/drm/hisilicon/kirin/dw_drm_dsi.c
>>>    create mode 100644 drivers/gpu/drm/hisilicon/kirin/dw_dsi_reg.h
>>>
>>> diff --git a/drivers/gpu/drm/hisilicon/kirin/Kconfig
>>> b/drivers/gpu/drm/hisilicon/kirin/Kconfig
>>> index 3ac4b8edeac1..de0d454c5c13 100644
>>> --- a/drivers/gpu/drm/hisilicon/kirin/Kconfig
>>> +++ b/drivers/gpu/drm/hisilicon/kirin/Kconfig
>>> @@ -4,6 +4,7 @@ config DRM_HISI_KIRIN
>>>          select DRM_KMS_HELPER
>>>          select DRM_GEM_CMA_HELPER
>>>          select DRM_KMS_CMA_HELPER
>>> +       select DRM_MIPI_DSI
>>>          help
>>>            Choose this option if you have a hisilicon Kirin
>>> chipsets(hi6220).
>>>            If M is selected the module will be called kirin-drm.
>>> diff --git a/drivers/gpu/drm/hisilicon/kirin/Makefile
>>> b/drivers/gpu/drm/hisilicon/kirin/Makefile
>>> index 2a61ab006ddb..5dcd0d4328b6 100644
>>> --- a/drivers/gpu/drm/hisilicon/kirin/Makefile
>>> +++ b/drivers/gpu/drm/hisilicon/kirin/Makefile
>>> @@ -1,4 +1,5 @@
>>>    kirin-drm-y := kirin_drm_drv.o \
>>> -              kirin_drm_ade.o
>>> +              kirin_drm_ade.o \
>>> +              dw_drm_dsi.o
>>>
>>>    obj-$(CONFIG_DRM_HISI_KIRIN) += kirin-drm.o
>>> diff --git a/drivers/gpu/drm/hisilicon/kirin/dw_drm_dsi.c
>>> b/drivers/gpu/drm/hisilicon/kirin/dw_drm_dsi.c
>>> new file mode 100644
>>> index 000000000000..8329148cc89d
>>> --- /dev/null
>>> +++ b/drivers/gpu/drm/hisilicon/kirin/dw_drm_dsi.c
>>> @@ -0,0 +1,743 @@
>>> +/*
>>> + * DesignWare MIPI DSI Host Controller v1.02 driver
>>> + *
>>> + * Copyright (c) 2016 Linaro Limited.
>>> + * Copyright (c) 2014-2016 Hisilicon Limited.
>>> + *
>>> + * Author:
>>> + *     Xinliang Liu <z.liuxinliang@hisilicon.com>
>>> + *     Xinliang Liu <xinliang.liu@linaro.org>
>>> + *     Xinwei Kong <kong.kongxinwei@hisilicon.com>
>>> + *
>>> + * This program is free software; you can redistribute it and/or modify
>>> + * it under the terms of the GNU General Public License version 2 as
>>> + * published by the Free Software Foundation.
>>> + *
>>> + */
>>> +
>>> +#include <linux/clk.h>
>>> +#include <linux/component.h>
>>> +#include <linux/of_graph.h>
>>> +
>>> +#include <drm/drm_of.h>
>>> +#include <drm/drm_crtc_helper.h>
>>> +#include <drm/drm_mipi_dsi.h>
>>> +#include <drm/drm_encoder_slave.h>
>>> +#include <drm/drm_atomic_helper.h>
>>> +
>>> +#include "dw_dsi_reg.h"
>>> +
>>> +#define MAX_TX_ESC_CLK            (10)
>>> +#define ROUND(x, y) ((x) / (y) + ((x) % (y) * 10 / (y) >= 5 ? 1 : 0))
>>> +#define PHY_REF_CLK_RATE   19200000
>>> +#define PHY_REF_CLK_PERIOD_PS (1000000000 / (PHY_REF_CLK_RATE / 1000))
>>> +
>>> +#define encoder_to_dsi(encoder) \
>>> +       container_of(encoder, struct dw_dsi, encoder)
>>> +#define host_to_dsi(host) \
>>> +       container_of(host, struct dw_dsi, host)
>>> +
>>> +struct mipi_phy_params {
>>> +       u32 clk_t_lpx;
>>> +       u32 clk_t_hs_prepare;
>>> +       u32 clk_t_hs_zero;
>>> +       u32 clk_t_hs_trial;
>>> +       u32 clk_t_wakeup;
>>> +       u32 data_t_lpx;
>>> +       u32 data_t_hs_prepare;
>>> +       u32 data_t_hs_zero;
>>> +       u32 data_t_hs_trial;
>>> +       u32 data_t_ta_go;
>>> +       u32 data_t_ta_get;
>>> +       u32 data_t_wakeup;
>>> +       u32 hstx_ckg_sel;
>>> +       u32 pll_fbd_div5f;
>>> +       u32 pll_fbd_div1f;
>>> +       u32 pll_fbd_2p;
>>> +       u32 pll_enbwt;
>>> +       u32 pll_fbd_p;
>>> +       u32 pll_fbd_s;
>>> +       u32 pll_pre_div1p;
>>> +       u32 pll_pre_p;
>>> +       u32 pll_vco_750M;
>>> +       u32 pll_lpf_rs;
>>> +       u32 pll_lpf_cs;
>>> +       u32 clklp2hs_time;
>>> +       u32 clkhs2lp_time;
>>> +       u32 lp2hs_time;
>>> +       u32 hs2lp_time;
>>> +       u32 clk_to_data_delay;
>>> +       u32 data_to_clk_delay;
>>> +       u32 lane_byte_clk_kHz;
>>> +       u32 clk_division;
>>> +};
>>> +
>>> +struct dsi_hw_ctx {
>>> +       void __iomem *base;
>>> +       struct clk *pclk;
>>> +};
>>> +
>>> +struct dw_dsi {
>>> +       struct drm_encoder encoder;
>>> +       struct drm_display_mode cur_mode;
>>> +       struct dsi_hw_ctx *ctx;
>>> +       struct mipi_phy_params phy;
>>> +
>>> +       u32 lanes;
>>> +       enum mipi_dsi_pixel_format format;
>>> +       unsigned long mode_flags;
>>> +       bool enable;
>>> +};
>>> +
>>> +struct dsi_data {
>>> +       struct dw_dsi dsi;
>>> +       struct dsi_hw_ctx ctx;
>>> +};
>>> +
>>> +struct dsi_phy_range {
>>> +       u32 min_range_kHz;
>>> +       u32 max_range_kHz;
>>> +       u32 pll_vco_750M;
>>> +       u32 hstx_ckg_sel;
>>> +};
>>> +
>>> +static const struct dsi_phy_range dphy_range_info[] = {
>>> +       {   46875,    62500,   1,    7 },
>>> +       {   62500,    93750,   0,    7 },
>>> +       {   93750,   125000,   1,    6 },
>>> +       {  125000,   187500,   0,    6 },
>>> +       {  187500,   250000,   1,    5 },
>>> +       {  250000,   375000,   0,    5 },
>>> +       {  375000,   500000,   1,    4 },
>>> +       {  500000,   750000,   0,    4 },
>>> +       {  750000,  1000000,   1,    0 },
>>> +       { 1000000,  1500000,   0,    0 }
>>> +};
>>> +
>>> +static void dsi_get_phy_params(u32 phy_freq_kHz,
>>> +                              struct mipi_phy_params *phy)
>>> +{
>>> +       u32 ui = 0;
>>> +       u32 cfg_clk_ps = PHY_REF_CLK_PERIOD_PS;
>>> +       u32 i = 0;
>>> +       u32 q_pll = 1;
>>> +       u32 m_pll = 0;
>>> +       u32 n_pll = 0;
>>> +       u32 r_pll = 1;
>>> +       u32 m_n = 0;
>>> +       u32 m_n_int = 0;
>>> +       u64 f_kHz;
>>> +       u64 temp;
>>> +       u64 tmp_kHz = phy_freq_kHz;
>>> +
>>> +       do {
>>> +               f_kHz = tmp_kHz;
>>> +
>>> +               /* Find the PLL clock range from the table */
>>> +               for (i = 0; i < ARRAY_SIZE(dphy_range_info); i++)
>>> +                       if (f_kHz >= dphy_range_info[i].min_range_kHz &&
>>> +                           f_kHz <= dphy_range_info[i].max_range_kHz)
>>> +                               break;
>>> +
>>> +               if (i == ARRAY_SIZE(dphy_range_info)) {
>>> +                       DRM_ERROR("%lldkHz out of range\n", f_kHz);
>>> +                       return;
>>> +               }
>>> +
>>> +               phy->pll_vco_750M = dphy_range_info[i].pll_vco_750M;
>>> +               phy->hstx_ckg_sel = dphy_range_info[i].hstx_ckg_sel;
>>> +
>>> +               if (phy->hstx_ckg_sel <= 7 &&
>>> +                   phy->hstx_ckg_sel >= 4)
>>> +                       q_pll = 0x10 >> (7 - phy->hstx_ckg_sel);
>>> +
>>> +               temp = f_kHz * (u64)q_pll * (u64)cfg_clk_ps;
>>> +               m_n_int = temp / (u64)1000000000;
>>> +               m_n = (temp % (u64)1000000000) / (u64)100000000;
>>> +
>>> +               if (m_n_int % 2 == 0) {
>>> +                       if (m_n * 6 >= 50) {
>>> +                               n_pll = 2;
>>> +                               m_pll = (m_n_int + 1) * n_pll;
>>> +                       } else if (m_n * 6 >= 30) {
>>> +                               n_pll = 3;
>>> +                               m_pll = m_n_int * n_pll + 2;
>>> +                       } else {
>>> +                               n_pll = 1;
>>> +                               m_pll = m_n_int * n_pll;
>>> +                       }
>>> +               } else {
>>> +                       if (m_n * 6 >= 50) {
>>> +                               n_pll = 1;
>>> +                               m_pll = (m_n_int + 1) * n_pll;
>>> +                       } else if (m_n * 6 >= 30) {
>>> +                               n_pll = 1;
>>> +                               m_pll = (m_n_int + 1) * n_pll;
>>> +                       } else if (m_n * 6 >= 10) {
>>> +                               n_pll = 3;
>>> +                               m_pll = m_n_int * n_pll + 1;
>>> +                       } else {
>>> +                               n_pll = 2;
>>> +                               m_pll = m_n_int * n_pll;
>>> +                       }
>>> +               }
>>> +
>>> +               if (n_pll == 1) {
>>> +                       phy->pll_fbd_p = 0;
>>> +                       phy->pll_pre_div1p = 1;
>>> +               } else {
>>> +                       phy->pll_fbd_p = n_pll;
>>> +                       phy->pll_pre_div1p = 0;
>>> +               }
>>> +
>>> +               if (phy->pll_fbd_2p <= 7 && phy->pll_fbd_2p >= 4)
>>> +                       r_pll = 0x10 >> (7 - phy->pll_fbd_2p);
>>> +
>>> +               if (m_pll == 2) {
>>> +                       phy->pll_pre_p = 0;
>>> +                       phy->pll_fbd_s = 0;
>>> +                       phy->pll_fbd_div1f = 0;
>>> +                       phy->pll_fbd_div5f = 1;
>>> +               } else if (m_pll >= 2 * 2 * r_pll && m_pll <= 2 * 4 *
>>> r_pll) {
>>> +                       phy->pll_pre_p = m_pll / (2 * r_pll);
>>> +                       phy->pll_fbd_s = 0;
>>> +                       phy->pll_fbd_div1f = 1;
>>> +                       phy->pll_fbd_div5f = 0;
>>> +               } else if (m_pll >= 2 * 5 * r_pll && m_pll <= 2 * 150 *
>>> r_pll) {
>>> +                       if (((m_pll / (2 * r_pll)) % 2) == 0) {
>>> +                               phy->pll_pre_p =
>>> +                                       (m_pll / (2 * r_pll)) / 2 - 1;
>>> +                               phy->pll_fbd_s =
>>> +                                       (m_pll / (2 * r_pll)) % 2 + 2;
>>> +                       } else {
>>> +                               phy->pll_pre_p =
>>> +                                       (m_pll / (2 * r_pll)) / 2;
>>> +                               phy->pll_fbd_s =
>>> +                                       (m_pll / (2 * r_pll)) % 2;
>>> +                       }
>>> +                       phy->pll_fbd_div1f = 0;
>>> +                       phy->pll_fbd_div5f = 0;
>>> +               } else {
>>> +                       phy->pll_pre_p = 0;
>>> +                       phy->pll_fbd_s = 0;
>>> +                       phy->pll_fbd_div1f = 0;
>>> +                       phy->pll_fbd_div5f = 1;
>>> +               }
>>> +
>>> +               f_kHz = (u64)1000000000 * (u64)m_pll /
>>> +                       ((u64)cfg_clk_ps * (u64)n_pll * (u64)q_pll);
>>> +
>>> +               if (f_kHz >= phy_freq_kHz)
>>> +                       break;
>>> +
>>> +               tmp_kHz += 10;
>>> +
>>> +       } while (1);
>>> +
>>> +       ui = 1000000 / f_kHz;
>>> +
>>> +       phy->clk_t_lpx = ROUND(50, 8 * ui);
>>> +       phy->clk_t_hs_prepare = ROUND(133, 16 * ui) - 1;
>>> +
>>> +       phy->clk_t_hs_zero = ROUND(262, 8 * ui);
>>> +       phy->clk_t_hs_trial = 2 * (ROUND(60, 8 * ui) - 1);
>>> +       phy->clk_t_wakeup = ROUND(1000000, (cfg_clk_ps / 1000) - 1);
>>> +       if (phy->clk_t_wakeup > 0xff)
>>> +               phy->clk_t_wakeup = 0xff;
>>> +       phy->data_t_wakeup = phy->clk_t_wakeup;
>>> +       phy->data_t_lpx = phy->clk_t_lpx;
>>> +       phy->data_t_hs_prepare = ROUND(125 + 10 * ui, 16 * ui) - 1;
>>> +       phy->data_t_hs_zero = ROUND(105 + 6 * ui, 8 * ui);
>>> +       phy->data_t_hs_trial = 2 * (ROUND(60 + 4 * ui, 8 * ui) - 1);
>>> +       phy->data_t_ta_go = 3;
>>> +       phy->data_t_ta_get = 4;
>>> +
>>> +       phy->pll_enbwt = 1;
>>> +       phy->clklp2hs_time = ROUND(407, 8 * ui) + 12;
>>> +       phy->clkhs2lp_time = ROUND(105 + 12 * ui, 8 * ui);
>>> +       phy->lp2hs_time = ROUND(240 + 12 * ui, 8 * ui) + 1;
>>> +       phy->hs2lp_time = phy->clkhs2lp_time;
>>> +       phy->clk_to_data_delay = 1 + phy->clklp2hs_time;
>>> +       phy->data_to_clk_delay = ROUND(60 + 52 * ui, 8 * ui) +
>>> +                               phy->clkhs2lp_time;
>>> +
>>> +       phy->lane_byte_clk_kHz = f_kHz / 8;
>>> +       phy->clk_division = phy->lane_byte_clk_kHz / MAX_TX_ESC_CLK;
>>> +       if (phy->lane_byte_clk_kHz % MAX_TX_ESC_CLK)
>>> +               phy->clk_division++;
>>> +}
>>> +
>>> +static u32 dsi_get_dpi_color_coding(enum mipi_dsi_pixel_format format)
>>> +{
>>> +       u32 val;
>>> +
>>> +       /* TODO: only support RGB888 now, to support more */
>>> +       switch (format) {
>>> +       case MIPI_DSI_FMT_RGB888:
>>> +               val = DSI_24BITS_1;
>>> +               break;
>>> +       default:
>>> +               val = DSI_24BITS_1;
>>> +               break;
>>> +       }
>>> +
>>> +       return val;
>>> +}
>>> +
>>> +/*
>>> + * dsi phy reg write function
>>> + */
>>> +static void dsi_phy_tst_set(void __iomem *base, u32 reg, u32 val)
>>> +{
>>> +       writel(reg, base + PHY_TST_CTRL1);
>>> +       /* reg addr written at first */
>>> +       wmb();
>>> +       writel(0x02, base + PHY_TST_CTRL0);
>>> +       /* cmd1 sent for write */
>>> +       wmb();
>>> +       writel(0x00, base + PHY_TST_CTRL0);
>>> +       /* cmd2 sent for write */
>>> +       wmb();
>>> +       writel(val, base + PHY_TST_CTRL1);
>>> +       /* Then write data */
>>> +       wmb();
>>> +       writel(0x02, base + PHY_TST_CTRL0);
>>> +       /* cmd2 sent for write */
>>> +       wmb();
>>> +       writel(0x00, base + PHY_TST_CTRL0);
>>
>>
>> You can use iowrite32 here to prevent calling wmb before each
>> write.
>>
>
> This chip is arm64. I find that for arm64,  it just wrap iowrite32 is
> the same to writel. Am I wrong?

You're right. So, you can keep writel, but drop remove the wmb() calls,
since writel already calls wmb() internally.

Archit

>
>>
>>> +}
>>> +
>>> +static void dsi_set_mipi_phy(void __iomem *base,
>>> +                            struct mipi_phy_params *phy,
>>> +                            u32 lanes)
>>> +{
>>> +       u32 delay_count;
>>> +       u32 val;
>>> +       u32 i;
>>> +
>>> +       /* set lanes value */
>>> +       val = (lanes - 1) | (PHY_STOP_WAIT_TIME << 8);
>>> +       writel(val, base + PHY_IF_CFG);
>>> +
>>> +       /* set phy clk division */
>>> +       val = readl(base + CLKMGR_CFG) | phy->clk_division;
>>> +       writel(val, base + CLKMGR_CFG);
>>> +
>>> +       /* clean up phy set param */
>>> +       writel(0, base + PHY_RSTZ);
>>> +       writel(0, base + PHY_TST_CTRL0);
>>> +       writel(1, base + PHY_TST_CTRL0);
>>> +       writel(0, base + PHY_TST_CTRL0);
>>> +
>>> +       /* clock lane Timing control - TLPX */
>>> +       dsi_phy_tst_set(base, 0x10010, phy->clk_t_lpx);
>>> +
>>> +       /* clock lane Timing control - THS-PREPARE */
>>> +       dsi_phy_tst_set(base, 0x10011, phy->clk_t_hs_prepare);
>>> +
>>> +       /* clock lane Timing control - THS-ZERO */
>>> +       dsi_phy_tst_set(base, 0x10012, phy->clk_t_hs_zero);
>>> +
>>> +       /* clock lane Timing control - THS-TRAIL */
>>> +       dsi_phy_tst_set(base, 0x10013, phy->clk_t_hs_trial);
>>> +
>>> +       /* clock lane Timing control - TWAKEUP */
>>> +       dsi_phy_tst_set(base, 0x10014, phy->clk_t_wakeup);
>>> +
>>> +       /* data lane */
>>> +       for (i = 0; i < lanes; i++) {
>>> +               /* Timing control - TLPX */
>>> +               dsi_phy_tst_set(base, 0x10020 + (i << 4),
>>> phy->data_t_lpx);
>>
>>
>> Some macros here for these PHY registers would be nice.
>
> Agree, will fix in next version.
>
>>
>>> +
>>> +               /* Timing control - THS-PREPARE */
>>> +               dsi_phy_tst_set(base, 0x10021 + (i << 4),
>>> +                               phy->data_t_hs_prepare);
>>> +
>>> +               /* Timing control - THS-ZERO */
>>> +               dsi_phy_tst_set(base, 0x10022 + (i << 4),
>>> phy->data_t_hs_zero);
>>> +
>>> +               /* Timing control - THS-TRAIL */
>>> +               dsi_phy_tst_set(base, 0x10023 + (i << 4),
>>> phy->data_t_hs_trial);
>>> +
>>> +               /* Timing control - TTA-GO */
>>> +               dsi_phy_tst_set(base, 0x10024 + (i << 4),
>>> phy->data_t_ta_go);
>>> +
>>> +               /* Timing control - TTA-GET */
>>> +               dsi_phy_tst_set(base, 0x10025 + (i << 4),
>>> phy->data_t_ta_get);
>>> +
>>> +               /* Timing control - TWAKEUP */
>>> +               dsi_phy_tst_set(base, 0x10026 + (i << 4),
>>> phy->data_t_wakeup);
>>
>>
>> Something like the macro below would be more readable here:
>>
>> #define PHY_TWAKEUP(i) (0x10026 + (i << 4))
>
> Agree, will fix in next version.
>
> Thanks,
> -xinliang
>
>>
>> Looks good otherwise.
>>
>> Archit
>>
>> --
>> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
>> a Linux Foundation Collaborative Project

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project