Re: [RFC PATCH 2/2] drm: xlnx: driver for Xilinx DSI TX Subsystem
From: Hyun Kwon
Date: Mon May 04 2020 - 14:44:08 EST
Hi GVRao,
Thanks for the patch. Sorry for late reply.
On Mon, 2020-04-20 at 14:20:56 -0700, Venkateshwar Rao Gannavarapu wrote:
> The Xilinx MIPI DSI Tx Subsystem soft IP is used to display video
> data from AXI-4 stream interface.
>
> It supports upto 4 lanes, optional register interface for the DPHY,
I don't see the register interface for dphy support.
> multiple RGB color formats, command mode and video mode.
> This is a MIPI-DSI host driver and provides DSI bus for panels.
> This driver also helps to communicate with its panel using panel
> framework.
>
> Signed-off-by: Venkateshwar Rao Gannavarapu <venkateshwar.rao.gannavarapu@xxxxxxxxxx>
> ---
> drivers/gpu/drm/xlnx/Kconfig | 11 +
> drivers/gpu/drm/xlnx/Makefile | 2 +
> drivers/gpu/drm/xlnx/xlnx_dsi.c | 755 ++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 768 insertions(+)
> create mode 100644 drivers/gpu/drm/xlnx/xlnx_dsi.c
>
> diff --git a/drivers/gpu/drm/xlnx/Kconfig b/drivers/gpu/drm/xlnx/Kconfig
> index aa6cd88..73873cf 100644
> --- a/drivers/gpu/drm/xlnx/Kconfig
> +++ b/drivers/gpu/drm/xlnx/Kconfig
> @@ -11,3 +11,14 @@ config DRM_ZYNQMP_DPSUB
> This is a DRM/KMS driver for ZynqMP DisplayPort controller. Choose
> this option if you have a Xilinx ZynqMP SoC with DisplayPort
> subsystem.
> +
> +config DRM_XLNX_DSI
> + tristate "Xilinx DRM DSI Subsystem Driver"
> + select DRM_MIPI_DSI
> + select DRM_PANEL
> + select DRM_PANEL_SIMPLE
> + help
> + This enables support for Xilinx MIPI-DSI.
This sentence is not needed with below. Could you please rephrase the whole?
> + This is a DRM/KMS driver for Xilinx programmable DSI controller.
> + Choose this option if you have a Xilinx MIPI DSI-TX controller
> + subsytem.
These seem incorrectly indented.
> diff --git a/drivers/gpu/drm/xlnx/Makefile b/drivers/gpu/drm/xlnx/Makefile
> index 2b844c6..b7ee6ef 100644
> --- a/drivers/gpu/drm/xlnx/Makefile
> +++ b/drivers/gpu/drm/xlnx/Makefile
> @@ -1,2 +1,4 @@
> zynqmp-dpsub-objs += zynqmp_disp.o zynqmp_dpsub.o zynqmp_dp.o
> obj-$(CONFIG_DRM_ZYNQMP_DPSUB) += zynqmp-dpsub.o
> +
> +obj-$(CONFIG_DRM_XLNX_DSI) += xlnx_dsi.o
> diff --git a/drivers/gpu/drm/xlnx/xlnx_dsi.c b/drivers/gpu/drm/xlnx/xlnx_dsi.c
> new file mode 100644
> index 0000000..b8cae59
> --- /dev/null
> +++ b/drivers/gpu/drm/xlnx/xlnx_dsi.c
> @@ -0,0 +1,755 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Xilinx FPGA MIPI DSI Tx Controller driver
> + *
> + * Copyright (C) 2017 - 2019 Xilinx, Inc.
> + *
> + * Authors:
> + * - Saurabh Sengar <saurabhs@xxxxxxxxxx>
> + * - Venkateshwar Rao Gannavarapu <venkateshwar.rao.gannavarapu@xxxxxxxxxx>
> + */
> +
> +#include <drm/drm_atomic_helper.h>
> +#include <drm/drm_connector.h>
> +#include <drm/drm_crtc.h>
> +#include <drm/drm_crtc_helper.h>
> +#include <drm/drm_device.h>
> +#include <drm/drm_encoder.h>
> +#include <drm/drm_fourcc.h>
> +#include <drm/drm_gem_cma_helper.h>
> +#include <drm/drm_mipi_dsi.h>
> +#include <drm/drm_panel.h>
> +#include <drm/drm_probe_helper.h>
> +
> +#include <linux/clk.h>
> +#include <linux/component.h>
> +#include <linux/device.h>
> +#include <linux/iopoll.h>
> +#include <linux/of_device.h>
> +#include <linux/of_graph.h>
> +#include <linux/phy/phy.h>
> +
> +#include <video/mipi_display.h>
> +#include <video/videomode.h>
> +
> +/* DSI Tx IP registers */
> +#define XDSI_CCR 0x00
> +#define XDSI_CCR_COREENB BIT(0)
> +#define XDSI_CCR_SOFTRST BIT(1)
> +#define XDSI_CCR_CRREADY BIT(2)
> +#define XDSI_CCR_CMDMODE BIT(3)
> +#define XDSI_CCR_DFIFORST BIT(4)
> +#define XDSI_CCR_CMDFIFORST BIT(5)
> +#define XDSI_PCR 0x04
> +#define XDSI_PCR_VIDEOMODE(x) (((x) & 0x3) << 3)
> +#define XDSI_PCR_VIDEOMODE_MASK (0x3 << 3)
> +#define XDSI_PCR_VIDEOMODE_SHIFT 3
> +#define XDSI_PCR_BLLPTYPE(x) ((x) << 5)
> +#define XDSI_PCR_BLLPMODE(x) ((x) << 6)
> +#define XDSI_PCR_EOTPENABLE(x) ((x) << 13)
> +#define XDSI_GIER 0x20
> +#define XDSI_ISR 0x24
> +#define XDSI_IER 0x28
> +#define XDSI_STR 0x2C
> +#define XDSI_STR_RDY_SHPKT BIT(6)
> +#define XDSI_STR_RDY_LNGPKT BIT(7)
> +#define XDSI_STR_DFIFO_FULL BIT(8)
> +#define XDSI_STR_DFIFO_EMPTY BIT(9)
> +#define XDSI_STR_WAITFR_DATA BIT(10)
> +#define XDSI_STR_CMD_EXE_PGS BIT(11)
> +#define XDSI_STR_CCMD_PROC BIT(12)
> +#define XDSI_STR_LPKT_MASK (0x5 << 7)
> +#define XDSI_CMD 0x30
> +#define XDSI_CMD_QUEUE_PACKET(x) ((x) & GENMASK(23, 0))
> +#define XDSI_DFR 0x34
> +#define XDSI_TIME1 0x50
> +#define XDSI_TIME1_BLLP_BURST(x) ((x) & GENMASK(15, 0))
> +#define XDSI_TIME1_HSA(x) (((x) & GENMASK(15, 0)) << 16)
> +#define XDSI_TIME2 0x54
> +#define XDSI_TIME2_VACT(x) ((x) & GENMASK(15, 0))
> +#define XDSI_TIME2_HACT(x) (((x) & GENMASK(15, 0)) << 16)
> +#define XDSI_HACT_MULTIPLIER GENMASK(1, 0)
> +#define XDSI_TIME3 0x58
> +#define XDSI_TIME3_HFP(x) ((x) & GENMASK(15, 0))
> +#define XDSI_TIME3_HBP(x) (((x) & GENMASK(15, 0)) << 16)
> +#define XDSI_TIME4 0x5c
> +#define XDSI_TIME4_VFP(x) ((x) & GENMASK(7, 0))
> +#define XDSI_TIME4_VBP(x) (((x) & GENMASK(7, 0)) << 8)
> +#define XDSI_TIME4_VSA(x) (((x) & GENMASK(7, 0)) << 16)
> +#define XDSI_LTIME 0x60
> +#define XDSI_BLLP_TIME 0x64
> +/*
> + * XDSI_NUM_DATA_T represents number of data types in the
> + * enum mipi_dsi_pixel_format in the MIPI DSI part of DRM framework.
> + */
> +#define XDSI_NUM_DATA_T 4
> +#define XDSI_VIDEO_MODE_SYNC_PULSE 0x0
> +#define XDSI_VIDEO_MODE_SYNC_EVENT 0x1
> +#define XDSI_VIDEO_MODE_BURST 0x2
Please remove these unused 3 definitions.
> +
> +#define XDSI_DPHY_CLK_MIN 197000000000UL
> +#define XDSI_DPHY_CLK_MAX 203000000000UL
> +#define XDSI_DPHY_CLK_REQ 200000000000UL
> +
> +/* command timeout in usec */
> +#define XDSI_CMD_TIMEOUT_VAL (3000)
> +
> +/**
> + * struct xlnx_dsi - Xilinx DSI-TX core
> + * @encoder: DRM encoder structure
> + * @dsi_host: DSI host device
> + * @connector: DRM connector structure
> + * @panel_node: MIPI DSI device panel node
> + * @panel: DRM panel structure
> + * @dev: device structure
> + * @iomem: Base address of DSI subsystem
> + * @lanes: number of active data lanes supported by DSI controller
> + * @cmdmode: command mode support
> + * @mode_flags: DSI operation mode related flags
> + * @format: pixel format for video mode of DSI controller
> + * @vm: videomode data structure
> + * @mul_factor: multiplication factor for HACT timing parameter
> + * @video_aclk: Video clock
> + * @dphy_clk_200M: 200MHz DPHY clock and AXI Lite clock
> + */
> +struct xlnx_dsi {
> + struct drm_encoder encoder;
> + struct mipi_dsi_host dsi_host;
> + struct drm_connector connector;
> + struct device_node *panel_node;
> + struct drm_panel *panel;
> + struct device *dev;
> + void __iomem *iomem;
> + u32 lanes;
> + bool cmdmode;
> + u32 mode_flags;
> + enum mipi_dsi_pixel_format format;
> + struct videomode vm;
> + u32 mul_factor;
> + struct clk *video_aclk;
> + struct clk *dphy_clk_200M;
> +};
> +
> +#define host_to_dsi(host) container_of(host, struct xlnx_dsi, dsi_host)
> +#define connector_to_dsi(c) container_of(c, struct xlnx_dsi, connector)
> +#define encoder_to_dsi(e) container_of(e, struct xlnx_dsi, encoder)
> +
> +static inline void xlnx_dsi_writel(void __iomem *base, int offset, u32 val)
> +{
> + writel(val, base + offset);
> +}
> +
> +static inline u32 xlnx_dsi_readl(void __iomem *base, int offset)
> +{
> + return readl(base + offset);
> +}
> +
> +/**
> + * xlnx_dsi_set_display_mode - Configure DSI timing registers
> + * @dsi: xilinx DSI structure
> + *
> + * This function writes the timing parameters of DSI IP which are
> + * retrieved from panel timing values.
> + */
> +static void xlnx_dsi_set_display_mode(struct xlnx_dsi *dsi)
This function better directly inline to xlnx_dsi_atomic_mode_set(). Then,
the videomode is not really needed. Please remove it and use drm_display_mode.
> +{
> + struct videomode *vm = &dsi->vm;
> + u32 reg, video_mode;
> +
> + reg = xlnx_dsi_readl(dsi->iomem, XDSI_PCR);
> + video_mode = (reg & XDSI_PCR_VIDEOMODE_MASK) >>
> + XDSI_PCR_VIDEOMODE_SHIFT;
> +
> + /* configure the HSA value only if non_burst_sync_pluse video mode */
> + if (!video_mode &&
> + (dsi->mode_flags & MIPI_DSI_MODE_VIDEO_SYNC_PULSE)) {
> + reg = XDSI_TIME1_HSA(vm->hsync_len);
> + xlnx_dsi_writel(dsi->iomem, XDSI_TIME1, reg);
> + }
> +
> + reg = XDSI_TIME4_VFP(vm->vfront_porch) |
> + XDSI_TIME4_VBP(vm->vback_porch) |
> + XDSI_TIME4_VSA(vm->vsync_len);
> + xlnx_dsi_writel(dsi->iomem, XDSI_TIME4, reg);
> +
> + reg = XDSI_TIME3_HFP(vm->hfront_porch) |
> + XDSI_TIME3_HBP(vm->hback_porch);
> + xlnx_dsi_writel(dsi->iomem, XDSI_TIME3, reg);
> +
> + dev_dbg(dsi->dev, "mul factor for parsed datatype is = %d\n",
> + (dsi->mul_factor) / 100);
Remove unneeded ().
> + /*
> + * The HACT parameter received from panel timing values should be
> + * divisible by 4. The reason for this is, the word count given as
> + * input to DSI controller is HACT * mul_factor. The mul_factor is
> + * 3, 2.25, 2.25, 2 respectively for RGB888, RGB666_L, RGB666_P and
> + * RGB565.
> + * e.g. for RGB666_L color format and 1080p, the word count is
> + * 1920*2.25 = 4320 which is divisible by 4 and it is a valid input
> + * to DSI controller. Based on this 2.25 mul factor, we come up with
> + * the division factor of (XDSI_HACT_MULTIPLIER) as 4 for checking
> + */
> + if ((vm->hactive & XDSI_HACT_MULTIPLIER) != 0)
if (!(vm->hactive & XDSI_HACT_MULTIPLIER))?
> + dev_warn(dsi->dev, "Incorrect HACT will be programmed\n");
dev_warn() better come with more details. Probably it can mention that
hactive is not aligned along with its value.
> +
> + reg = XDSI_TIME2_HACT((vm->hactive) * (dsi->mul_factor) / 100) |
Ditto. No ().
> + XDSI_TIME2_VACT(vm->vactive);
> + xlnx_dsi_writel(dsi->iomem, XDSI_TIME2, reg);
> +
> + dev_dbg(dsi->dev, "LCD size = %dx%d\n", vm->hactive, vm->vactive);
> +}
> +
> +/**
> + * xlnx_dsi_set_display_enable - Enables the DSI Tx IP core enable
> + * register bit
> + * @dsi: Xilinx DSI structure
> + *
> + * This function takes the DSI strucure and enables the core enable bit
> + * of core configuration register.
> + */
> +static void xlnx_dsi_set_display_enable(struct xlnx_dsi *dsi)
> +{
> + u32 reg;
> +
> + reg = xlnx_dsi_readl(dsi->iomem, XDSI_CCR);
> + reg |= XDSI_CCR_COREENB;
> +
Nit. It can be less lines by removing this empty line and setting the reg
at declarartion an so on.
> + xlnx_dsi_writel(dsi->iomem, XDSI_CCR, reg);
> + dev_dbg(dsi->dev, "MIPI DSI Tx controller is enabled.\n");
> +}
> +
> +/**
> + * xlnx_dsi_set_display_disable - Disable the DSI Tx IP core enable
> + * register bit
> + * @dsi: Xilinx DSI structure
> + *
> + * This function takes the DSI strucure and disables the core enable bit
> + * of core configuration register.
> + */
> +static void xlnx_dsi_set_display_disable(struct xlnx_dsi *dsi)
> +{
> + u32 reg;
> +
> + reg = xlnx_dsi_readl(dsi->iomem, XDSI_CCR);
> + reg &= ~XDSI_CCR_COREENB;
> +
Ditto.
> + xlnx_dsi_writel(dsi->iomem, XDSI_CCR, reg);
> + dev_dbg(dsi->dev, "DSI Tx is disabled. reset regs to default values\n");
> +}
> +
> +/**
> + * xlnx_dsi_host_transfer - transfer command to panel
> + * @host: mipi dsi host structure
> + * @msg: mipi dsi msg with type, length and data
> + *
> + * This function is valid only in command mode.
> + * It checks the command fifo empty status and writes into
> + * data or cmd register and waits for the completion status.
> + *
> + * Return: number of bytes, on success and error number on failure
Nit. Extra spaces.
> + */
> +static ssize_t xlnx_dsi_host_transfer(struct mipi_dsi_host *host,
> + const struct mipi_dsi_msg *msg)
> +{
> + struct xlnx_dsi *dsi = host_to_dsi(host);
> + u32 data0, data1, cmd0, status, val;
> + const char *tx_buf = msg->tx_buf;
> +
> + if (!(xlnx_dsi_readl(dsi->iomem, XDSI_CCR) & (XDSI_CCR_COREENB |
> + XDSI_CCR_CMDMODE))) {
This passes when one of these 2 bits is set, and I don't think it's right.
> + dev_err(dsi->dev, "dsi command mode not enabled\n");
> + return -EINVAL;
> + }
> +
> + if (msg->type == MIPI_DSI_DCS_LONG_WRITE) {
> + status = readl_poll_timeout(dsi->iomem + XDSI_STR, val,
> + ((val & XDSI_STR_LPKT_MASK) ==
> + XDSI_STR_LPKT_MASK), 1,
Please remove unneeded (). Same for rest below.
> + XDSI_CMD_TIMEOUT_VAL);
It'd be nice if these delay / timeout values are described..
> + if (status) {
> + dev_err(dsi->dev, "long cmd fifo not empty!\n");
> + return -ETIMEDOUT;
Isn't this checking if the controller is ready for long packet transaction?
Then maybe -EBUSY is better. For timeout, the 'status' can propagate.
Same for below.
> + }
> + data0 = tx_buf[0] | (tx_buf[1] << 8) | (tx_buf[2] << 16) |
> + (tx_buf[3] << 24);
> + data1 = tx_buf[4] | (tx_buf[5] << 8);
> + cmd0 = msg->type | (MIPI_DSI_DCS_READ << 8);
> +
> + xlnx_dsi_writel(dsi->iomem, XDSI_DFR, data0);
> + xlnx_dsi_writel(dsi->iomem, XDSI_DFR, data1);
> + xlnx_dsi_writel(dsi->iomem, XDSI_CMD, cmd0);
> + } else {
I'm not familiar with DSI transaction handling, but is it safe to assume
all other than long write is short packet write (param, without param)?
> + data0 = tx_buf[0];
This line can be removed.
> + if (msg->type == MIPI_DSI_DCS_SHORT_WRITE_PARAM)
> + data0 = MIPI_DSI_DCS_SHORT_WRITE_PARAM |
> + (tx_buf[0] << 8) | (tx_buf[1] << 16);
> + else
> + data0 = MIPI_DSI_DCS_SHORT_WRITE | (tx_buf[0] << 8);
I'd put () for this if - else statments, but it may be just me.
Then, this can mvoe to right before xlnx_dsi_writel() below.
> +
> + status = readl_poll_timeout(dsi->iomem + XDSI_STR, val,
> + ((val & XDSI_STR_RDY_SHPKT) ==
> + XDSI_STR_RDY_SHPKT), 1,
> + XDSI_CMD_TIMEOUT_VAL);
> + if (status) {
> + dev_err(dsi->dev, "short cmd fifo not empty\n");
> + return -ETIMEDOUT;
> + }
> + xlnx_dsi_writel(dsi->iomem, XDSI_CMD, data0);
> + }
> +
> + status = readl_poll_timeout(dsi->iomem + XDSI_STR, val,
> + (!(val & XDSI_STR_CMD_EXE_PGS)), 1,
> + XDSI_CMD_TIMEOUT_VAL);
> + if (status) {
> + dev_err(dsi->dev, "cmd time out\n");
> + return -ETIMEDOUT;
> + }
> +
> + return msg->tx_len;
> +}
> +
> +static int xlnx_dsi_host_attach(struct mipi_dsi_host *host,
> + struct mipi_dsi_device *device)
> +{
> + u32 panel_lanes;
> + struct xlnx_dsi *dsi = host_to_dsi(host);
> +
> + panel_lanes = device->lanes;
I'd do this above,
u32 panel_lanes = device->lanes
> + dsi->mode_flags = device->mode_flags;
> + dsi->panel_node = device->dev.of_node;
> +
> + if (panel_lanes != dsi->lanes) {
> + dev_err(dsi->dev, "Mismatch of lanes. panel = %d, DSI = %d\n",
> + panel_lanes, dsi->lanes);
> + return -EINVAL;
> + }
> +
> + if (device->format != dsi->format) {
> + dev_err(dsi->dev, "Mismatch of format. panel = %d, DSI = %d\n",
> + device->format, dsi->format);
> + return -EINVAL;
> + }
> +
> + if (dsi->connector.dev)
> + drm_helper_hpd_irq_event(dsi->connector.dev);
> +
> + return 0;
> +}
> +
> +static int xlnx_dsi_host_detach(struct mipi_dsi_host *host,
> + struct mipi_dsi_device *device)
> +{
> + struct xlnx_dsi *dsi = host_to_dsi(host);
> +
> + if (dsi->panel) {
> + drm_panel_detach(dsi->panel);
> + dsi->panel = NULL;
> + }
> +
> + if (dsi->connector.dev)
> + drm_helper_hpd_irq_event(dsi->connector.dev);
> +
> + return 0;
> +}
> +
> +static const struct mipi_dsi_host_ops xlnx_dsi_ops = {
> + .attach = xlnx_dsi_host_attach,
> + .detach = xlnx_dsi_host_detach,
> + .transfer = xlnx_dsi_host_transfer,
> +};
> +
> +static int xlnx_dsi_connector_dpms(struct drm_connector *connector,
> + int mode)
> +{
> + struct xlnx_dsi *dsi = connector_to_dsi(connector);
> + int ret;
> +
> + dev_dbg(dsi->dev, "connector dpms state: %d\n", mode);
> +
> + switch (mode) {
> + case DRM_MODE_DPMS_ON:
> + ret = drm_panel_prepare(dsi->panel);
> + if (ret < 0) {
> + dev_err(dsi->dev, "DRM panel not found\n");
> + return ret;
> + }
> +
> + ret = drm_panel_enable(dsi->panel);
> + if (ret < 0) {
> + drm_panel_unprepare(dsi->panel);
> + dev_err(dsi->dev, "DRM panel not enabled\n");
> + return ret;
> + }
> + break;
> + default:
> + drm_panel_disable(dsi->panel);
> + drm_panel_unprepare(dsi->panel);
> + break;
> + }
> +
> + return drm_helper_connector_dpms(connector, mode);
> +}
> +
> +static enum drm_connector_status
> +xlnx_dsi_detect(struct drm_connector *connector, bool force)
> +{
> + struct xlnx_dsi *dsi = connector_to_dsi(connector);
> +
> + if (!dsi->panel) {
> + dsi->panel = of_drm_find_panel(dsi->panel_node);
> + if (dsi->panel) {
> + drm_panel_attach(dsi->panel, &dsi->connector);
Above are quite static, so can't it be done in host attach? Or do you see more
dynamic use cases?
> + if (dsi->cmdmode) {
> + xlnx_dsi_writel(dsi->iomem, XDSI_CCR,
> + XDSI_CCR_CMDMODE |
> + XDSI_CCR_COREENB);
> + drm_panel_prepare(dsi->panel);
Just to confirm, is this not needed for non-command mode? And how is this
aligned with same call in dpms(), meaning the panel will be prepared
twice in command mode.
> + xlnx_dsi_writel(dsi->iomem, XDSI_CCR, 0);
> + }
> + }
> + } else if (!dsi->panel_node) {
> + xlnx_dsi_connector_dpms(connector, DRM_MODE_DPMS_OFF);
> + drm_panel_detach(dsi->panel);
> + dsi->panel = NULL;
> + }
> +
> + if (dsi->panel)
> + return connector_status_connected;
> +
> + return connector_status_disconnected;
> +}
> +
> +static void xlnx_dsi_connector_destroy(struct drm_connector *connector)
> +{
> + drm_connector_unregister(connector);
> + drm_connector_cleanup(connector);
> + connector->dev = NULL;
> +}
> +
> +static const struct drm_connector_funcs xlnx_dsi_connector_funcs = {
> + .detect = xlnx_dsi_detect,
> + .fill_modes = drm_helper_probe_single_connector_modes,
> + .destroy = xlnx_dsi_connector_destroy,
> + .reset = drm_atomic_helper_connector_reset,
This better indent more consistently. Please do it through the entire file.
> + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
> + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
> + .dpms = xlnx_dsi_connector_dpms,
> +};
> +
> +static int xlnx_dsi_get_modes(struct drm_connector *connector)
> +{
> + struct xlnx_dsi *dsi = connector_to_dsi(connector);
> +
> + if (dsi->panel)
> + return drm_panel_get_modes(dsi->panel, connector);
> +
> + return 0;
> +}
> +
> +static struct drm_encoder *
> +xlnx_dsi_best_encoder(struct drm_connector *connector)
> +{
> + return &(connector_to_dsi(connector)->encoder);
> +}
> +
> +static const struct drm_connector_helper_funcs
> +xlnx_dsi_connector_helper_funcs = {
> + .get_modes = xlnx_dsi_get_modes,
> + .best_encoder = xlnx_dsi_best_encoder,
> +};
> +
> +static int xlnx_dsi_create_connector(struct drm_encoder *encoder)
> +{
> + struct xlnx_dsi *dsi = encoder_to_dsi(encoder);
> + struct drm_connector *connector = &dsi->connector;
> + struct drm_device *drm = encoder->dev;
> + int ret;
> +
> + connector->polled = DRM_CONNECTOR_POLL_HPD;
> +
> + ret = drm_connector_init(encoder->dev, connector,
> + &xlnx_dsi_connector_funcs,
> + DRM_MODE_CONNECTOR_DSI);
> + if (ret) {
> + dev_err(dsi->dev, "Failed to initialize connector with drm\n");
> + return ret;
> + }
> +
> + drm_connector_helper_add(connector, &xlnx_dsi_connector_helper_funcs);
> + drm_connector_attach_encoder(connector, encoder);
> + if (!drm->registered)
> + return 0;
> +
> + drm_connector_register(connector);
> +
> + return 0;
> +}
> +
> +/**
> + * xlnx_dsi_atomic_mode_set - derive the DSI timing parameters
> + *
> + * @encoder: pointer to Xilinx DRM encoder
> + * @crtc_state: Pointer to drm core crtc state
> + * @connector_state: DSI connector drm state
> + *
> + * This function derives the DSI IP timing parameters from the timing
> + * values given in the attached panel driver.
> + */
> +static void
> +xlnx_dsi_atomic_mode_set(struct drm_encoder *encoder,
> + struct drm_crtc_state *crtc_state,
> + struct drm_connector_state *connector_state)
> +{
> + struct xlnx_dsi *dsi = encoder_to_dsi(encoder);
> + struct videomode *vm = &dsi->vm;
> + struct drm_display_mode *m = &crtc_state->adjusted_mode;
> +
> + vm->hactive = m->hdisplay;
> + vm->vactive = m->vdisplay;
> + vm->vfront_porch = m->vsync_start - m->vdisplay;
> + vm->vback_porch = m->vtotal - m->vsync_end;
> + vm->vsync_len = m->vsync_end - m->vsync_start;
> + vm->hfront_porch = m->hsync_start - m->hdisplay;
> + vm->hback_porch = m->htotal - m->hsync_end;
> + vm->hsync_len = m->hsync_end - m->hsync_start;
> + xlnx_dsi_set_display_mode(dsi);
As commented above, this doesn't have to be re-directed through videomode.
Then the function can be inlined here.
> +}
> +
> +static void xlnx_dsi_disable(struct drm_encoder *encoder)
> +{
> + struct xlnx_dsi *dsi = encoder_to_dsi(encoder);
> +
> + xlnx_dsi_set_display_disable(dsi);
xlnx_dsi_set_display_disable(encoder_to_dsi(encoder));
Maybe even this wrapper can be removed?
> +}
> +
> +static void xlnx_dsi_enable(struct drm_encoder *encoder)
> +{
> + struct xlnx_dsi *dsi = encoder_to_dsi(encoder);
> +
> + xlnx_dsi_set_display_enable(dsi);
Ditto.
> +}
> +
> +static const struct drm_encoder_helper_funcs
> +xlnx_dsi_encoder_helper_funcs = {
> + .atomic_mode_set = xlnx_dsi_atomic_mode_set,
> + .enable = xlnx_dsi_enable,
> + .disable = xlnx_dsi_disable,
> +};
> +
> +static const struct drm_encoder_funcs xlnx_dsi_encoder_funcs = {
> + .destroy = drm_encoder_cleanup,
> +};
> +
> +static int xlnx_dsi_parse_dt(struct xlnx_dsi *dsi)
> +{
> + struct device *dev = dsi->dev;
> + struct device_node *node = dev->of_node;
> + int ret;
> + u32 datatype;
> + static const int xdsi_mul_fact[XDSI_NUM_DATA_T] = {300, 225, 225, 200};
> +
> + dsi->dphy_clk_200M = devm_clk_get(dev, "dphy_clk_200M");
> + if (IS_ERR(dsi->dphy_clk_200M)) {
> + ret = PTR_ERR(dsi->dphy_clk_200M);
> + dev_err(dev, "failed to get dphy_clk_200M %d\n", ret);
> + return ret;
> + }
> +
> + dsi->video_aclk = devm_clk_get(dev, "s_axis_aclk");
> + if (IS_ERR(dsi->video_aclk)) {
> + ret = PTR_ERR(dsi->video_aclk);
> + dev_err(dev, "failed to get video_clk %d\n", ret);
> + return ret;
> + }
> +
> + ret = of_property_read_u32(node, "xlnx,dsi-num-lanes", &dsi->lanes);
> + if (ret < 0) {
> + dev_err(dsi->dev, "missing xlnx,dsi-num-lanes property\n");
> + return ret;
> + }
> + if (dsi->lanes > 4 || dsi->lanes < 1) {
> + dev_err(dsi->dev, "%d lanes : invalid lanes\n", dsi->lanes);
> + return -EINVAL;
> + }
Nit. An empty line here would be helpful for readability.
> + ret = of_property_read_u32(node, "xlnx,dsi-data-type", &datatype);
> + if (ret < 0) {
> + dev_err(dsi->dev, "missing xlnx,dsi-data-type property\n");
> + return ret;
> + }
> + dsi->format = datatype;
> + if (datatype > MIPI_DSI_FMT_RGB565) {
> + dev_err(dsi->dev, "Invalid xlnx,dsi-data-type string\n");
> + return -EINVAL;
> + }
> +
> + /*
> + * multiplication factor used for HACT, based on data type.
> + *
> + * e.g. for RGB666_L datatype and 1920x1080 resolution,
> + * the Hact (WC) would be as follows -
> + * 1920 pixels * 18 bits per pixel / 8 bits per byte
> + * = 1920 pixels * 2.25 bytes per pixel = 4320 bytes.
> + *
> + * Data Type - Multiplication factor
> + * RGB888 - 3
> + * RGB666_L - 2.25
> +- * RGB666_P - 2.25
> + * RGB565 - 2
> + *
> + * Since the multiplication factor maybe a floating number,
> + * a 100x multiplication factor is used.
> + */
> + dsi->mul_factor = xdsi_mul_fact[datatype];
> +
> + dsi->cmdmode = of_property_read_bool(node, "xlnx,dsi-cmd-mode");
> +
> + dev_dbg(dsi->dev, "DSI controller num lanes = %d", dsi->lanes);
> + dev_dbg(dsi->dev, "DSI controller datatype = %d\n", datatype);
> + dev_dbg(dsi->dev, "DSI controller cmd mode = %d\n", dsi->cmdmode);
> +
> + return 0;
> +}
> +
> +static int xlnx_dsi_bind(struct device *dev, struct device *master,
> + void *data)
> +{
> + struct xlnx_dsi *dsi = dev_get_drvdata(dev);
> + struct drm_encoder *encoder = &dsi->encoder;
> + struct drm_device *drm_dev = data;
> + int ret;
> +
> + drm_encoder_init(drm_dev, encoder, &xlnx_dsi_encoder_funcs,
> + DRM_MODE_ENCODER_DSI, NULL);
> + drm_encoder_helper_add(encoder, &xlnx_dsi_encoder_helper_funcs);
> +
> + encoder->possible_crtcs = 1;
> +
> + ret = xlnx_dsi_create_connector(encoder);
> + if (ret) {
> + dev_err(dsi->dev, "fail creating connector, ret = %d\n", ret);
> + drm_encoder_cleanup(encoder);
> + return ret;
> + }
> +
> + ret = mipi_dsi_host_register(&dsi->dsi_host);
> + if (ret < 0) {
Just nit. if (ret) would do.
> + dev_err(dsi->dev, "failed to register DSI host: %d\n", ret);
> + xlnx_dsi_connector_destroy(&dsi->connector);
> + drm_encoder_cleanup(encoder);
> + return ret;
> + }
> +
> + return 0;
> +}
> +
> +static void xlnx_dsi_unbind(struct device *dev, struct device *master,
> + void *data)
> +{
> + struct xlnx_dsi *dsi = dev_get_drvdata(dev);
> +
> + xlnx_dsi_disable(&dsi->encoder);
> + mipi_dsi_host_unregister(&dsi->dsi_host);
> +}
> +
> +static const struct component_ops xlnx_dsi_component_ops = {
> + .bind = xlnx_dsi_bind,
> + .unbind = xlnx_dsi_unbind,
> +};
> +
> +static int xlnx_dsi_probe(struct platform_device *pdev)
> +{
> + struct device *dev = &pdev->dev;
> + struct resource *res;
> + struct xlnx_dsi *dsi;
> + int ret;
> + unsigned long rate;
> +
> + dsi = devm_kzalloc(dev, sizeof(*dsi), GFP_KERNEL);
> + if (!dsi)
> + return -ENOMEM;
> +
> + dsi->dsi_host.ops = &xlnx_dsi_ops;
> + dsi->dsi_host.dev = dev;
> + dsi->dev = dev;
> +
> + ret = xlnx_dsi_parse_dt(dsi);
> + if (ret)
> + return ret;
> +
> + res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> + dsi->iomem = devm_ioremap_resource(dev, res);
> + if (IS_ERR(dsi->iomem)) {
> + dev_err(dev, "failed to remap io region\n");
> + return PTR_ERR(dsi->iomem);
> + }
> + platform_set_drvdata(pdev, dsi);
> +
> + ret = clk_set_rate(dsi->dphy_clk_200M, XDSI_DPHY_CLK_REQ);
> + if (ret) {
> + dev_err(dev, "failed to set dphy clk rate %d\n", ret);
> + return ret;
> + }
> +
> + ret = clk_prepare_enable(dsi->dphy_clk_200M);
> + if (ret) {
> + dev_err(dev, "failed to enable dphy clk %d\n", ret);
> + return ret;
> + }
> +
> + rate = clk_get_rate(dsi->dphy_clk_200M);
> + if (rate < XDSI_DPHY_CLK_MIN && rate > XDSI_DPHY_CLK_MAX) {
A brief comment for this line wold be helpful.
> + dev_err(dev, "Error DPHY clock = %lu\n", rate);
> + ret = -EINVAL;
> + goto err_disable_dphy_clk;
> + }
> +
> + ret = clk_prepare_enable(dsi->video_aclk);
> + if (ret) {
> + dev_err(dev, "failed to enable video clk %d\n", ret);
> + goto err_disable_dphy_clk;
> + }
> +
> + ret = component_add(dev, &xlnx_dsi_component_ops);
> + if (ret < 0)
I prefer if (ret) where possible, but it may be just me. I let you decide.
> + goto err_disable_video_clk;
> +
> + return ret;
return 0;? but up to you.
> +
> +err_disable_video_clk:
> + clk_disable_unprepare(dsi->video_aclk);
> +err_disable_dphy_clk:
> + clk_disable_unprepare(dsi->dphy_clk_200M);
> + return ret;
> +}
At least, this initial one better come with a complete display pipeline,
ex crtc / plane are missing. This won't be functional without those.
I believe Laurent was planning to make the xlnx drm layer as a library,
so hopefully this can be a good starting point and come with drivers aligned
with it.
> +
> +static int xlnx_dsi_remove(struct platform_device *pdev)
> +{
> + struct xlnx_dsi *dsi = platform_get_drvdata(pdev);
> +
> + component_del(&pdev->dev, &xlnx_dsi_component_ops);
> + clk_disable_unprepare(dsi->video_aclk);
> + clk_disable_unprepare(dsi->dphy_clk_200M);
> +
> + return 0;
> +}
> +
> +static const struct of_device_id xlnx_dsi_of_match[] = {
> + { .compatible = "xlnx-dsi", },
This should change. I believe it should be something like, xlnx,<dsi ip name>-<version>
Thanks,
-hyun
> + { /* end of table */ },
> +};
> +MODULE_DEVICE_TABLE(of, xlnx_dsi_of_match);
> +
> +static struct platform_driver dsi_driver = {
> + .probe = xlnx_dsi_probe,
> + .remove = xlnx_dsi_remove,
> + .driver = {
> + .name = "xlnx-dsi",
> + .of_match_table = xlnx_dsi_of_match,
> + },
> +};
> +
> +module_platform_driver(dsi_driver);
> +
> +MODULE_AUTHOR("Xilinx, Inc.");
> +MODULE_DESCRIPTION("Xilinx FPGA MIPI DSI Tx Driver");
> +MODULE_LICENSE("GPL v2");
> --
> 2.7.4
>