Re: [PATCH v2] drm/mediatek: dsi: Add dsi per-frame lp code for mt8188

From: Shuijing Li (李水静)
Date: Tue Jul 09 2024 - 02:37:51 EST

Next message: Sungjong Seo: "RE: [PATCH] fat: print s_dev via fat_msg"
Previous message: Varadarajan Narayanan: "[PATCH v1] pci: qcom: Fix 'opp' variable usage"
Next in thread: CK Hu (胡俊光): "Re: [PATCH v2] drm/mediatek: dsi: Add dsi per-frame lp code for mt8188"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Dear CK,

1.
> Where do you define HSTX_BLLP_EN?
> DSI_TXRX_CTRL is written in mtk_dsi_txrx_control() and I guess that
> bllp_en is zero, right?
> If bllp_en is zero, drop the code related to bllp_en.
2.
> ps_wc = dsi->vm.hactive * dsi_buf_bpp;
3.
> This register is never written. Which setting would influence this
> value? Before dsi configuration finish, is this value stable?
>
==>Answer 1.2.3:
bllp_en/bllp_wc may be assigned values in the future.
ps_wc may change in the future when features are added(e.g. dsc).
Therefore, it is recommended to keep the original modifications to
increase compatibility.

4.
> Is this roundup the same as the one in mtk_dsi_config_vdo_timing()?
> If the same, merge these two.
==>Not same.

Based on the feedback you provided, I have made the necessary changes
and submitted version 3 for your consideration.

Thank you for your time and attention to this matter.

Best Regards,
Shuijing

On Wed, 2024-05-22 at 07:30 +0000, CK Hu (胡俊光) wrote:
> Hi, Shuijing:
>
> On Wed, 2024-04-24 at 17:16 +0800, Shuijing Li wrote:
> > Adding the per-frame lp function of mt8188, which can keep HFP in
> > HS and
> > reduce the time required for each line to enter and exit low power.
> > Per Frame LP:
> > |<----------One Active Frame-------->|
> > --______________________________________----___________________
> > ^HSA+HBP^^RGB^^HFP^^HSA+HBP^^RGB^^HFP^ ^HSA+HBP^^RGB^^HFP^
> >
> > Per Line LP:
> > |<---------------One Active Frame----------->|
> > --______________--______________--______________----______________
> > ^HSA+HBP^^RGB^ ^HSA+HBP^^RGB^ ^HSA+HBP^^RGB^ ^HSA+HBP^^RGB^
> >
> > Signed-off-by: Shuijing Li <shuijing.li@xxxxxxxxxxxx>
> > ---
> > Changes in v2:
> > Use bitfield macros and add new function for per prame lp and
> > improve
> > the format, per suggestion frome previous thread:
> >
https://urldefense.com/v3/__https://patchwork.kernel.org/project/linux-mediatek/patch/20240314094238.3315-1-shuijing.li@xxxxxxxxxxxx/__;!!CTRNKA9wMg0ARbw!iJrqnLwBi80AYshU8-sTvs0LiPH_URRrjTLsXFQLyUyAWZW0x5B0lCgIKTXl-wS9NmkTBcIdyncDMxDio1jk8w$
> >
> > ---
> > drivers/gpu/drm/mediatek/mtk_dsi.c | 143
> > +++++++++++++++++++++++++++++
> > 1 file changed, 143 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/mediatek/mtk_dsi.c
> > b/drivers/gpu/drm/mediatek/mtk_dsi.c
> > index 9501f4019199..75719b0535f7 100644
> > --- a/drivers/gpu/drm/mediatek/mtk_dsi.c
> > +++ b/drivers/gpu/drm/mediatek/mtk_dsi.c
> > @@ -68,6 +68,8 @@
> > #define EXT_TE_EDGE BIT(10)
> > #define MAX_RTN_SIZE GENMASK(15, 12)
> > #define HSTX_CKLP_EN BIT(16)
> > +#define HSTX_BLLP_EN_SHIFT 7
> > +#define HSTX_BLLP_EN_MASK 0x1
> >
> > #define DSI_PSCTRL 0x1c
> > #define DSI_PS_WC GENMASK(13, 0)
> > @@ -76,6 +78,8 @@
> > #define PACKED_PS_18BIT_RGB666 1
> > #define LOOSELY_PS_24BIT_RGB666 2
> > #define PACKED_PS_24BIT_RGB888 3
> > +#define PS_WC_SHIFT 0
> > +#define PS_WC_MASK 0x7fff
> >
> > #define DSI_VSA_NL 0x20
> > #define DSI_VBP_NL 0x24
> > @@ -88,12 +92,20 @@
> > #define DSI_HSA_WC 0x50
> > #define DSI_HBP_WC 0x54
> > #define DSI_HFP_WC 0x58
> > +#define HFP_HS_EN 31
> > +#define HFP_HS_VB_PS_WC_SHIFT 16
> > +
> > +#define DSI_BLLP_WC 0x5C
> > +#define BLLP_WC_SHIFT 0
> > +#define BLLP_WC_MASK 0xfff
> >
> > #define DSI_CMDQ_SIZE 0x60
> > #define CMDQ_SIZE 0x3f
> > #define CMDQ_SIZE_SEL BIT(15)
> >
> > #define DSI_HSTX_CKL_WC 0x64
> > +#define HSTX_CKL_WC_SHIFT 2
> > +#define HSTX_CKL_WC_MASK 0x3fff
> >
> > #define DSI_RX_DATA0 0x74
> > #define DSI_RX_DATA1 0x78
> > @@ -118,12 +130,22 @@
> > #define HS_PREP GENMASK(15, 8)
> > #define HS_ZERO GENMASK(23, 16)
> > #define HS_TRAIL GENMASK(31, 24)
> > +#define LPX_SHIFT 0
> > +#define LPX_MASK 0xff
> > +#define DA_HS_PREP_SHIFT 8
> > +#define DA_HS_PREP_MASK 0xff
> > +#define DA_HS_ZERO_SHIFT 16
> > +#define DA_HS_ZERO_MASK 0xff
> > +#define DA_HS_TRAIL_SHIFT 24
> > +#define DA_HS_TRAIL_MASK 0xff
> >
> > #define DSI_PHY_TIMECON1 0x114
> > #define TA_GO GENMASK(7, 0)
> > #define TA_SURE GENMASK(15, 8)
> > #define TA_GET GENMASK(23, 16)
> > #define DA_HS_EXIT GENMASK(31, 24)
> > +#define DA_HS_EXIT_SHIFT 24
> > +#define DA_HS_EXIT_MASK 0xff
> >
> > #define DSI_PHY_TIMECON2 0x118
> > #define CONT_DET GENMASK(7, 0)
> > @@ -154,6 +176,7 @@
> > #define DATA_1 GENMASK(31, 24)
> >
> > #define NS_TO_CYCLE(n, c) ((n) / (c) + (((n) % (c)) ? 1 : 0))
> > +#define REG_FIELD_VALUE(reg, field) (((reg) >> (field##_SHIFT)) &
> > (field##_MASK))
>
> Use function in bitfield.h instead of re-inventing it.
>
> >
> > #define MTK_DSI_HOST_IS_READ(type) \
> > ((type == MIPI_DSI_GENERIC_READ_REQUEST_0_PARAM) || \
> > @@ -187,6 +210,7 @@ struct mtk_dsi_driver_data {
> > bool has_shadow_ctl;
> > bool has_size_ctl;
> > bool cmdq_long_packet_ctl;
> > + bool support_per_frame_lp;
> > };
> >
> > struct mtk_dsi {
> > @@ -425,6 +449,121 @@ static void mtk_dsi_ps_control(struct mtk_dsi
> > *dsi, bool config_vact)
> > writel(ps_val, dsi->regs + DSI_PSCTRL);
> > }
> >
> > +static void mtk_dsi_config_vdo_timing_per_frame_lp(struct mtk_dsi
> > *dsi)
> > +{
> > + u32 horizontal_sync_active_byte;
> > + u32 horizontal_backporch_byte;
> > + u32 horizontal_frontporch_byte;
> > + u32 dsi_tmp_buf_bpp;
> > + unsigned int lpx, da_hs_exit, da_hs_prep, da_hs_trail;
> > + unsigned int da_hs_zero, ps_wc, hs_vb_ps_wc;
> > + u32 bllp_wc, bllp_en, v_active_roundup, hstx_cklp_wc;
> > + u32 hstx_cklp_wc_max, hstx_cklp_wc_min;
> > + struct videomode *vm = &dsi->vm;
> > +
> > + if (dsi->format == MIPI_DSI_FMT_RGB565)
> > + dsi_tmp_buf_bpp = 2;
> > + else
> > + dsi_tmp_buf_bpp = 3;
> > +
> > + da_hs_trail = REG_FIELD_VALUE(readl(dsi->regs +
> > DSI_PHY_TIMECON0),
> > + DA_HS_TRAIL);
>
> da_hs_trail = dsi->phy_timing.clk_hs_trail;
>
> > + bllp_en = REG_FIELD_VALUE(readl(dsi->regs + DSI_TXRX_CTRL),
> > + HSTX_BLLP_EN);
>
> Where do you define HSTX_BLLP_EN?
> DSI_TXRX_CTRL is written in mtk_dsi_txrx_control() and I guess that
> bllp_en is zero, right?
> If bllp_en is zero, drop the code related to bllp_en.
>
> > + if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO_SYNC_PULSE) {
> > + horizontal_sync_active_byte =
> > + vm->hsync_len * dsi_tmp_buf_bpp - 10;
> > + horizontal_backporch_byte =
> > + vm->hback_porch * dsi_tmp_buf_bpp - 10;
> > + horizontal_frontporch_byte =
> > + vm->hfront_porch * dsi_tmp_buf_bpp - 12;
> > +
> > + ps_wc = REG_FIELD_VALUE(readl(dsi->regs + DSI_PSCTRL),
> > PS_WC);
>
> ps_wc = dsi->vm.hactive * dsi_buf_bpp;
>
> > + v_active_roundup = (32 + horizontal_sync_active_byte +
> > + horizontal_backporch_byte + ps_wc +
> > + horizontal_frontporch_byte) % dsi->lanes;
> > + if (v_active_roundup)
> > + horizontal_backporch_byte =
> > horizontal_backporch_byte +
> > + dsi->lanes - v_active_roundup;
>
> Is this roundup the same as the one in mtk_dsi_config_vdo_timing()?
> If the same, merge these two.
>
> > + hstx_cklp_wc_min = (DIV_ROUND_UP((12 + 2 + 4 +
> > + horizontal_sync_active_byte), dsi->lanes) +
> > da_hs_trail + 1)
> > + * dsi->lanes / 6 - 1;
> > + hstx_cklp_wc_max = (DIV_ROUND_UP((20 + 6 + 4 +
> > + horizontal_sync_active_byte +
> > horizontal_backporch_byte +
> > + ps_wc), dsi->lanes) + da_hs_trail + 1) * dsi-
> > >lanes / 6 - 1;
> > + } else {
> > + horizontal_sync_active_byte = vm->hsync_len *
> > dsi_tmp_buf_bpp - 4;
> > +
> > + horizontal_backporch_byte = (vm->hback_porch + vm-
> > >hsync_len) *
> > + dsi_tmp_buf_bpp - 10;
> > + hstx_cklp_wc_min = (DIV_ROUND_UP(4, dsi->lanes) +
> > da_hs_trail + 1)
> > + * dsi->lanes / 6 - 1;
> > +
> > + if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO_BURST) {
> > + ps_wc = REG_FIELD_VALUE(readl(dsi->regs +
> > DSI_PSCTRL), PS_WC);
> > + bllp_wc = REG_FIELD_VALUE(readl(dsi->regs +
> > DSI_BLLP_WC), BLLP_WC);
>
> This register is never written. Which setting would influence this
> value? Before dsi configuration finish, is this value stable?
>
> > + horizontal_frontporch_byte = (vm->hfront_porch
> > *
> > + dsi_tmp_buf_bpp - 18);
> > +
> > + v_active_roundup = (28 +
> > horizontal_backporch_byte + ps_wc +
> > + horizontal_frontporch_byte + bllp_wc) %
> > dsi->lanes;
> > + if (v_active_roundup)
> > + horizontal_backporch_byte =
> > horizontal_backporch_byte +
> > + dsi->lanes - v_active_roundup;
> > + if (bllp_en) {
> > + hstx_cklp_wc_max = (DIV_ROUND_UP((16 +
> > 6 + 4 +
> > + horizontal_backporch_byte +
> > bllp_wc + ps_wc),
> > + dsi->lanes) + da_hs_trail + 1)
> > * dsi->lanes / 6 - 1;
> > + } else {
> > + hstx_cklp_wc_max = (DIV_ROUND_UP((12 +
> > 4 + 4 +
> > + horizontal_backporch_byte +
> > bllp_wc + ps_wc),
> > + dsi->lanes) + da_hs_trail + 1)
> > * dsi->lanes / 6 - 1;
> > + }
> > + } else {
> > + ps_wc = REG_FIELD_VALUE(readl(dsi->regs +
> > DSI_PSCTRL), PS_WC);
> > + horizontal_frontporch_byte = (vm->hfront_porch
> > *
> > + dsi_tmp_buf_bpp - 12);
> > +
> > + v_active_roundup = (22 +
> > horizontal_backporch_byte + ps_wc +
> > + horizontal_frontporch_byte) % dsi-
> > >lanes;
> > + if (v_active_roundup)
> > + horizontal_backporch_byte =
> > horizontal_backporch_byte +
> > + dsi->lanes - v_active_roundup;
> > +
> > + hstx_cklp_wc_max = (DIV_ROUND_UP((12 + 4 + 4 +
> > + horizontal_backporch_byte + ps_wc),
> > + dsi->lanes) + da_hs_trail + 1) * dsi-
> > >lanes / 6 - 1;
> > + }
> > + }
> > + hstx_cklp_wc = REG_FIELD_VALUE(ps_wc, HSTX_CKL_WC);
>
> hstx_cklp_wc = REG_FIELD_VALUE(dsi->vm.hactive * dsi_buf_bpp,
> HSTX_CKL_WC);
>
> But this looks a little weird.
>
> > + if (hstx_cklp_wc <= hstx_cklp_wc_min ||
> > + hstx_cklp_wc >= hstx_cklp_wc_max) {
> > + hstx_cklp_wc = (hstx_cklp_wc_max / 2) <<
> > HSTX_CKL_WC_SHIFT;
>
> You choose the new setting with a risk that has a warning, why not
> let hstx_cklp_wc = (hstx_cklp_wc_min + hstx_cklp_wc_max) / 2 ?
>
> > + writel(hstx_cklp_wc, dsi->regs + DSI_HSTX_CKL_WC);
> > + }
> > + hstx_cklp_wc = hstx_cklp_wc >> HSTX_CKL_WC_SHIFT;
> > + if (hstx_cklp_wc <= hstx_cklp_wc_min ||
> > + hstx_cklp_wc >= hstx_cklp_wc_max) {
> > + DRM_WARN("Wrong setting of hstx_ckl_wc\n");
> > + }
> > +
> > + lpx = REG_FIELD_VALUE(readl(dsi->regs + DSI_PHY_TIMECON0),
> > LPX);
>
> lpx = dsi->phy_timing.lpx;
>
> > + da_hs_exit = REG_FIELD_VALUE(readl(dsi->regs +
> > DSI_PHY_TIMECON1), DA_HS_EXIT);
>
> da_hs_exit = dsi->phy_timing.da_hs_exit;
>
> > + da_hs_prep = REG_FIELD_VALUE(readl(dsi->regs +
> > DSI_PHY_TIMECON0), DA_HS_PREP);
>
> da_hs_prep = dsi->phy_timing.da_hs_prepare;
>
> > + da_hs_zero = REG_FIELD_VALUE(readl(dsi->regs +
> > DSI_PHY_TIMECON0), DA_HS_ZERO);
>
> da_hs_zero = dsi->phy_timing.da_hs_zero;
>
> > + ps_wc = REG_FIELD_VALUE(readl(dsi->regs + DSI_PSCTRL), PS_WC);
>
> This value is assigned?
>
> > + hs_vb_ps_wc = ps_wc -
> > + (lpx + da_hs_exit + da_hs_prep + da_hs_zero + 2)
> > + * dsi->lanes;
> > + horizontal_frontporch_byte = (1 << HFP_HS_EN)
> > + | (hs_vb_ps_wc << HFP_HS_VB_PS_WC_SHIFT)
> > + | (horizontal_frontporch_byte);
> > +
> > + writel(horizontal_sync_active_byte, dsi->regs + DSI_HSA_WC);
> > + writel(horizontal_backporch_byte, dsi->regs + DSI_HBP_WC);
> > + writel(horizontal_frontporch_byte, dsi->regs + DSI_HFP_WC);
> > +}
> > +
> > static void mtk_dsi_config_vdo_timing(struct mtk_dsi *dsi)
> > {
> > u32 horizontal_sync_active_byte;
> > @@ -499,6 +638,9 @@ static void mtk_dsi_config_vdo_timing(struct
> > mtk_dsi *dsi)
> > writel(horizontal_backporch_byte, dsi->regs + DSI_HBP_WC);
> > writel(horizontal_frontporch_byte, dsi->regs + DSI_HFP_WC);
> >
>
> If dsi->driver_data->support_per_frame_lp is false, the calculation
> before this is redundant. It's worth to move previous calculation to
> a function and call that function when dsi->driver_data-
> > support_per_frame_lp is false.
> > + if (dsi->driver_data->support_per_frame_lp)
> > + mtk_dsi_config_vdo_timing_per_frame_lp(dsi);
> > +
> > mtk_dsi_ps_control(dsi, false);
> > }
> >
> > @@ -1193,6 +1335,7 @@ static const struct mtk_dsi_driver_data
> > mt8188_dsi_driver_data = {
> > .has_shadow_ctl = true,
> > .has_size_ctl = true,
> > .cmdq_long_packet_ctl = true,
> > + .support_per_frame_lp = true,
> > };
> >
> > static const struct of_device_id mtk_dsi_of_match[] = {

Next message: Sungjong Seo: "RE: [PATCH] fat: print s_dev via fat_msg"
Previous message: Varadarajan Narayanan: "[PATCH v1] pci: qcom: Fix 'opp' variable usage"
Next in thread: CK Hu (胡俊光): "Re: [PATCH v2] drm/mediatek: dsi: Add dsi per-frame lp code for mt8188"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]