Hi,
On 10/17/2016 10:03 PM, Marek Vasut wrote:
> Add driver for the ITE IT6251 LVDS-to-eDP bridge.
>
> Signed-off-by: Marek Vasut <marex at denx.de>
> Cc: Daniel Vetter <daniel.vetter at ffwll.ch>
> Cc: Sean Cross <xobs at kosagi.com>
> ---
> drivers/gpu/drm/bridge/Kconfig | 9 +
> drivers/gpu/drm/bridge/Makefile | 1 +
> drivers/gpu/drm/bridge/ite-it6251.c | 606
> ++++++++++++++++++++++++++++++++++++
> 3 files changed, 616 insertions(+)
> create mode 100644 drivers/gpu/drm/bridge/ite-it6251.c
>
> diff --git a/drivers/gpu/drm/bridge/Kconfig b/drivers/gpu/drm/bridge/Kconfig
> index 10e12e7..e9c96b9 100644
> --- a/drivers/gpu/drm/bridge/Kconfig
> +++ b/drivers/gpu/drm/bridge/Kconfig
> @@ -39,6 +39,15 @@ config DRM_DW_HDMI_AHB_AUDIO
> Designware HDMI block. This is used in conjunction with
> the i.MX6 HDMI driver.
>
> +config DRM_ITE_IT6251
> + tristate "ITE IT6251 LVDS/eDP bridge"
> + depends on OF
> + select DRM_KMS_HELPER
> + select DRM_PANEL
> + select REGMAP_I2C
> + ---help---
> + ITE IT6251 LVDS-eDP bridge chip driver.
> +
> config DRM_NXP_PTN3460
> tristate "NXP PTN3460 DP/LVDS bridge"
> depends on OF
> diff --git a/drivers/gpu/drm/bridge/Makefile b/drivers/gpu/drm/bridge/Makefile
> index cdf3a3c..736dba7 100644
> --- a/drivers/gpu/drm/bridge/Makefile
> +++ b/drivers/gpu/drm/bridge/Makefile
> @@ -4,6 +4,7 @@ obj-$(CONFIG_DRM_ANALOGIX_ANX78XX) += analogix-anx78xx.o
> obj-$(CONFIG_DRM_DUMB_VGA_DAC) += dumb-vga-dac.o
> obj-$(CONFIG_DRM_DW_HDMI) += dw-hdmi.o
> obj-$(CONFIG_DRM_DW_HDMI_AHB_AUDIO) += dw-hdmi-ahb-audio.o
> +obj-$(CONFIG_DRM_ITE_IT6251) += ite-it6251.o
> obj-$(CONFIG_DRM_NXP_PTN3460) += nxp-ptn3460.o
> obj-$(CONFIG_DRM_PARADE_PS8622) += parade-ps8622.o
> obj-$(CONFIG_DRM_SII902X) += sii902x.o
> diff --git a/drivers/gpu/drm/bridge/ite-it6251.c
> b/drivers/gpu/drm/bridge/ite-it6251.c
> new file mode 100644
> index 0000000..a19bb4d
> --- /dev/null
> +++ b/drivers/gpu/drm/bridge/ite-it6251.c
> @@ -0,0 +1,606 @@
> +/*
We're trying to mention the bridge name, and what encoding it does for all
bridge
drivers. Could you describe it here too? Thanks.
> + * Copyright (C) 2014 Sean Cross <xobs at kosagi.com>
> + *
> + * Rework for mainline: Marek Vasut <marex at denx.de>
> + *
> + * This software is licensed under the terms of the GNU General Public
> + * License version 2, as published by the Free Software Foundation, and
> + * may be copied, distributed, and modified under those terms.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + */
> +
> +#include <linux/delay.h>
> +#include <linux/i2c.h>
> +#include <linux/module.h>
> +#include <linux/of_gpio.h>
This shouldn't be needed.
> +#include <linux/of_graph.h>
> +#include <linux/of_platform.h>
This too.
> +#include <linux/regmap.h>
> +#include <linux/regulator/consumer.h>
> +#include <drm/drm_panel.h>
> +
> +#include "drmP.h"
> +#include "drm_crtc.h"
> +#include "drm_crtc_helper.h"
> +#include "drm_atomic_helper.h"
> +
> +struct it6251_bridge {
> + struct i2c_client *client;
> + struct i2c_client *lvds_client;
> + struct regmap *regmap;
> + struct regmap *lvds_regmap;
> + struct regulator *regulator;
> +
> + struct drm_connector connector;
> + struct drm_bridge bridge;
> + struct drm_panel *panel;
> +};
> +
> +/* Register definitions */
> +#define IT6251_VENDOR_ID_LOW 0x00
> +#define IT6251_VENDOR_ID_HIGH 0x01
> +#define IT6251_DEVICE_ID_LOW 0x02
> +#define IT6251_DEVICE_ID_HIGH 0x03
> +#define IT6251_SYSTEM_STATUS 0x0d
> +#define IT6251_SYSTEM_STATUS_RINTSTATUS BIT(0)
> +#define IT6251_SYSTEM_STATUS_RHPDSTATUS BIT(1)
> +#define IT6251_SYSTEM_STATUS_RVIDEOSTABLE BIT(2)
> +#define IT6251_SYSTEM_STATUS_RPLL_IOLOCK BIT(3)
> +#define IT6251_SYSTEM_STATUS_RPLL_XPLOCK BIT(4)
> +#define IT6251_SYSTEM_STATUS_RPLL_SPLOCK BIT(5)
> +#define IT6251_SYSTEM_STATUS_RAUXFREQ_LOCK BIT(6)
> +#define IT6251_REF_STATE 0x0e
> +#define IT6251_REF_STATE_MAIN_LINK_DISABLED BIT(0)
> +#define IT6251_REF_STATE_AUX_CHANNEL_READ BIT(1)
> +#define IT6251_REF_STATE_CR_PATTERN BIT(2)
> +#define IT6251_REF_STATE_EQ_PATTERN BIT(3)
> +#define IT6251_REF_STATE_NORMAL_OPERATION BIT(4)
> +#define IT6251_REF_STATE_MUTED BIT(5)
> +#define IT6251_RPCLK_CNT_LOW 0x13
> +#define IT6251_RPCLK_CNT_HIGH 0x14
> +#define IT6251_RPC_REQ 0x2b
> +#define IT6251_RPC_REQ_RPC_FIFOFULL BIT(6)
> +#define IT6251_RPC_REQ_RPC_FIFOEMPTY BIT(7)
> +#define IT6251_PCLK_CNT_LOW 0x57
> +#define IT6251_PCLK_CNT_HIGH 0x58
> +#define IT6251_DPHDEW_LOW 0xa5
> +#define IT6251_DPHDEW_HIGH 0xa6
> +#define IT6251_DPVDEW_LOW 0xaf
> +#define IT6251_DPVDEW_HIGH 0xb0
> +#define IT6251_LVDS_PORT_ADDR 0xfd
> +#define IT6251_LVDS_PORT_CTRL 0xfe
> +#define IT6251_LVDS_PORT_CTRL_EN BIT(0)
> +
> +/*
> + * Register programming sequences.
> + * NOTE: There is a lot of registers here which are completely undocumented
> + * and/or their meaning is not clear from the little documentation
> + * that is available for this chip. These values below just seem to
> + * work well enough.
Are these configs valid for different pixel clock rates?
> + */
> +static const struct reg_sequence it6251_lvds_rx_sequence[] = {
> + { 0x05, 0x00 },
> +
> + { 0x3b, 0x42 }, /* reset LVDSRX PLL */
> + { 0x3b, 0x43 },
> +
> + { 0x3c, 0x08 }, /* something with SSC PLL */
> + { 0x0b, 0x88 }, /* don't swap links, writing reserved regs */
> +
> + { 0x2c, 0x01 }, /* JEIDA, 8-bit depth 0x11, original 0x42 */
> + { 0x32, 0x04 }, /* "reserved" */
> + { 0x35, 0xe0 }, /* "reserved" */
> + { 0x2b, 0x24 }, /* "reserved" + clock delay */
> +
> + { 0x05, 0x02 }, /* reset LVDSRX pix clock */
> + { 0x05, 0x00 },
> +};
> +
> +static const struct reg_sequence it6251_edp_tx_sequence[] = {
> + /* two lane mode, normal operation, no swapping, no downspread */
> + { 0x16, 0x02 },
> + { 0x23, 0x40 }, /* some AUX channel EDID magic */
> + { 0x5c, 0xf3 }, /* power down lanes 3-0 */
> + { 0x5f, 0x06 }, /* enable DP scrambling, change EQ CR phase */
> + { 0x60, 0x02 }, /* color mode RGB, pclk/2 */
> + { 0x61, 0x04 }, /* dual pixel input mode, no EO swap, no RGB swap */
> + { 0x62, 0x01 }, /* M444B24 video format */
> +
> + /* vesa range / not interlace / vsync high / hsync high */
> + { 0xa0, 0x0F },
> +
> + { 0xc9, 0xf5 }, /* hpd event timer set to 1.6-ish ms */
> +
> + { 0xca, 0x4d }, /* more reserved magic */
> + { 0xcb, 0x37 },
> +
> + /* enhanced framing mode, auto video fifo reset, video mute disable */
> + { 0xd3, 0x03 },
> + { 0xd4, 0x45 }, /* "vidstmp" and some reserved stuff */
> +
> + { 0xe7, 0xa0 }, /* queue number -- reserved */
> + { 0xe8, 0x33 }, /* info frame packets and reserved */
> + { 0xec, 0x00 }, /* more AVI stuff */
> +
> + { 0x23, 0x42 }, /* select PC master reg for aux channel? */
> +
> + { 0x24, 0x00 }, /* send PC request commands */
> + { 0x25, 0x00 },
> + { 0x26, 0x00 },
> +
> + { 0x2b, 0x00 }, /* native aux read */
> + { 0x23, 0x40 }, /* back to internal */
> +
> + { 0x19, 0xff }, /* voltage swing level 3 */
> + { 0x1a, 0xff }, /* pre-emphasis level 3 */
> +
> + { 0x17, 0x01 }, /* start link training */
Have you tried seeing if these map to the drm_dp helpers? We should ideally
set up an aux channel to configure some of the DPCD params on the eDP panel.
You could have a look at the TC358767 driver for reference.
> +};
> +
> +static struct it6251_bridge *bridge_to_it6251(struct drm_bridge *bridge)
> +{
> + return container_of(bridge, struct it6251_bridge, bridge);
> +}
> +
> +static struct it6251_bridge *conn_to_it6251(struct drm_connector *connector)
> +{
> + return container_of(connector, struct it6251_bridge, connector);
> +}
> +
> +static int it6251_is_stable(struct it6251_bridge *it6251)
> +{
> + struct drm_display_mode *mode;
> + unsigned int status, rpclkcnt, clkcnt, refstate, rpcreq;
> + u8 regs[2];
> + u16 hactive;
> + u16 vactive;
> + int ret;
> +
> + ret = regmap_read(it6251->regmap, IT6251_SYSTEM_STATUS, &status);
> + if (ret)
> + return ret;
> + dev_dbg(&it6251->client->dev, "System status: 0x%02x\n", status);
> +
> + if (!(status & IT6251_SYSTEM_STATUS_RVIDEOSTABLE))
> + return -EINVAL;
> +
> + ret = regmap_bulk_read(it6251->regmap, IT6251_RPCLK_CNT_LOW, regs, 2);
> + if (ret)
> + return ret;
> + rpclkcnt = (regs[0] & 0xff) | ((regs[1] & 0x0f) << 8);
> + dev_dbg(&it6251->client->dev, "RPCLKCnt: %d\n", rpclkcnt);
> +
> + ret = regmap_bulk_read(it6251->lvds_regmap, IT6251_PCLK_CNT_LOW,
> + regs, 2);
> + if (ret)
> + return ret;
> + clkcnt = (regs[0] & 0xff) | ((regs[1] & 0x0f) << 8);
> + dev_dbg(&it6251->client->dev, "Clock: 0x%02x\n", clkcnt);
> +
> + ret = regmap_read(it6251->lvds_regmap, IT6251_REF_STATE, &refstate);
> + if (ret)
> + return ret;
> + dev_dbg(&it6251->client->dev, "Ref Link State: 0x%02x\n", refstate);
> +
> + ret = regmap_read(it6251->lvds_regmap, IT6251_RPC_REQ, &rpcreq);
> + if (ret)
> + return ret;
> + dev_dbg(&it6251->client->dev, "RPC Req: 0x%02x\n", rpcreq);
> +
> + ret = regmap_bulk_read(it6251->regmap, IT6251_DPHDEW_LOW, regs, 2);
> + if (ret)
> + return ret;
> + hactive = (regs[0] & 0xff) | ((regs[1] & 0x1f) << 8);
> + dev_dbg(&it6251->client->dev, "hactive: %d\n", hactive);
> +
> + ret = regmap_bulk_read(it6251->regmap, IT6251_DPVDEW_LOW, regs, 2);
> + if (ret)
> + return ret;
> + vactive = (regs[0] & 0xff) | ((regs[1] & 0x0f) << 8);
> + dev_dbg(&it6251->client->dev, "vactive: %d\n", vactive);
> +
> + if ((refstate & 0x1f) != 0)
> + return -EINVAL;
> +
> + if (rpcreq & IT6251_RPC_REQ_RPC_FIFOFULL) {
> + dev_err(&it6251->client->dev,
> + "RPC fifofull is set, might be an error\n");
> + return -EINVAL;
> + }
> +
> + /* If video is muted, that's a failure */
> + if (refstate & IT6251_REF_STATE_MUTED)
> + return -EINVAL;
> +
> + list_for_each_entry(mode, &it6251->panel->connector->modes, head)
> + if ((mode->hdisplay == hactive) && (mode->vdisplay == vactive))
> + return 0;
Shouldn't the hactive and vactive registers be something that's configured
by the driver? It seems wrong to read off the registers and compare it
with a mode that we're trying to set in the first place.
Can you consider using the bridge mode_set and mode_fixup ops to set the
modes?
> +
> + dev_info(&it6251->client->dev, "no mode match found\n");
> + return -EINVAL;
> +}
> +
> +static int it6251_init(struct it6251_bridge *it6251)
> +{
> + const struct reg_sequence it6251_reset_reg_sequence[] = {
> + { 0x05, 0x00 },
> + { IT6251_LVDS_PORT_ADDR, it6251->lvds_client->addr << 1 },
> + { IT6251_LVDS_PORT_CTRL, IT6251_LVDS_PORT_CTRL_EN },
> + };
> +
> + int ret, stable_delays;
> + unsigned int reg;
> +
> + /*
> + * Reset DisplayPort half. Setting bit 2 causes IT6251 to not
> + * respond over i2c, which is considered "normal". This write
> + * will report failure, but will actually succeed.
> + */
> + regmap_write(it6251->regmap, 0x05, 0xff);
> +
> + /* Un-reset DisplayPort half and configure LVDS receiver. */
> + ret = regmap_multi_reg_write(it6251->regmap, it6251_reset_reg_sequence,
> + ARRAY_SIZE(it6251_reset_reg_sequence));
> + if (ret) {
> + dev_err(&it6251->client->dev, "cannot setup eDP half\n");
> + return ret;
> + }
> +
> + /* LVDS RX */
> + regmap_write(it6251->lvds_regmap, 0x05, 0xff);
> + ret = regmap_multi_reg_write(it6251->lvds_regmap,
> + it6251_lvds_rx_sequence,
> + ARRAY_SIZE(it6251_lvds_rx_sequence));
> + if (ret) {
> + dev_err(&it6251->lvds_client->dev, "cannot setup LVDS RX\n");
> + return ret;
> + }
> +
> + /* eDP TX */
> + ret = regmap_multi_reg_write(it6251->regmap,
> + it6251_edp_tx_sequence,
> + ARRAY_SIZE(it6251_edp_tx_sequence));
> + if (ret) {
> + dev_err(&it6251->client->dev, "cannot setup eDP TX\n");
> + return ret;
> + }
> +
> + for (stable_delays = 0; stable_delays < 100; stable_delays++) {
> + ret = regmap_read(it6251->regmap, 0x0e, ®);
> + if (ret || ((reg & 0x1f) != 0x10)) {
> + mdelay(2);
We aren't in an atomic context here, we could use usleep_range here.
> + continue;
> + }
> +
> + ret = regmap_read(it6251->regmap, IT6251_SYSTEM_STATUS, ®);
> + if (ret || !(reg & IT6251_SYSTEM_STATUS_RVIDEOSTABLE)) {
> + mdelay(2);
> + continue;
> + }
> +
> + break;
> + }
> +
> + /*
> + * If we couldn't stabilize, requeue and try again, because it means
> + * that the LVDS channel isn't stable yet.
> + */
> + ret = it6251_is_stable(it6251);
> + if (ret)
> + dev_err(&it6251->client->dev, "bridge is not stable\n");
> +
> + return ret;
> +}
> +
> +static int it6251_power_down(struct it6251_bridge *it6251)
> +{
> + struct device *dev = &it6251->client->dev;
> + int ret = 0;
> +
> + if (regulator_is_enabled(it6251->regulator)) {
> + ret = regulator_disable(it6251->regulator);
> + if (ret)
> + dev_err(dev, "unable to disable regulator\n");
> + }
> +
> + return ret;
> +}
> +
> +static int it6251_power_up(struct it6251_bridge *it6251)
> +{
> + struct i2c_client *client = it6251->client;
> + u8 regs[4];
> + int i, ret;
> +
> + ret = regulator_enable(it6251->regulator);
> + if (ret) {
> + dev_err(&client->dev, "unable to enable regulator\n");
> + return ret;
> + }
> +
> + /* Sometimes it seems like multiple tries are needed */
> + for (i = 0; i < 5; i++) {
> + ret = regmap_bulk_read(it6251->regmap, IT6251_VENDOR_ID_LOW,
> + regs, 4);
> + if (!ret && regs[0] && regs[1] && regs[2] && regs[3]) {
> + dev_info(&client->dev, "found ITE6251 [%04x:%04x]\n",
> + (regs[1] << 8) | regs[0],
> + (regs[3] << 8) | regs[2]);
> + return 0;
> + }
> +
> + usleep_range(100000, 200000);
Was this intended to be 100ms to 200ms? If so, msleep should be used.
> + }
> +
> + dev_err(&client->dev, "unable to read product id\n");
> + it6251_power_down(it6251);
> + return -EINVAL;
> +}
> +
> +/* I2C driver functions */
> +static void it6251_pre_enable(struct drm_bridge *bridge)
> +{
> + struct it6251_bridge *it6251 = bridge_to_it6251(bridge);
> +
> + if (drm_panel_prepare(it6251->panel)) {
> + DRM_ERROR("failed to prepare panel\n");
> + return;
> + }
> +
> + it6251_power_up(it6251);
> +}
> +
> +static void it6251_enable(struct drm_bridge *bridge)
> +{
> + struct it6251_bridge *it6251 = bridge_to_it6251(bridge);
> + int tries, ret;
> +
> + if (drm_panel_enable(it6251->panel)) {
> + DRM_ERROR("failed to enable panel\n");
> + return;
> + }
> +
> + for (tries = 0; tries < 5; tries++) {
> + ret = it6251_init(it6251);
> + if (!ret)
> + return;
> +
> + /* If the init failed, restart the chip */
> + it6251_power_down(it6251);
> + it6251_power_up(it6251);
> + }
> +}
> +
> +static void it6251_disable(struct drm_bridge *bridge)
> +{
> + struct it6251_bridge *it6251 = bridge_to_it6251(bridge);
> +
> + if (drm_panel_disable(it6251->panel))
> + DRM_ERROR("failed to disable panel\n");
> +}
> +
> +static void it6251_post_disable(struct drm_bridge *bridge)
> +{
> + struct it6251_bridge *it6251 = bridge_to_it6251(bridge);
> +
> + if (drm_panel_unprepare(it6251->panel))
> + DRM_ERROR("failed to unprepare panel\n");
> +
> + it6251_power_down(it6251);
> +}
> +
> +static int it6251_get_modes(struct drm_connector *connector)
> +{
> + struct it6251_bridge *it6251 = conn_to_it6251(connector);
> +
> + return drm_panel_get_modes(it6251->panel);
> +}
> +
> +static const struct drm_connector_helper_funcs it6251_connector_helper_funcs
> = {
> + .get_modes = it6251_get_modes,
> +};
> +
> +static enum drm_connector_status it6251_detect(struct drm_connector
> *connector,
> + bool force)
> +{
> + return connector_status_connected;
> +}
> +
> +static const struct drm_connector_funcs it6251_connector_funcs = {
> + .dpms = drm_atomic_helper_connector_dpms,
> + .fill_modes = drm_helper_probe_single_connector_modes,
> + .detect = it6251_detect,
> + .destroy = drm_connector_cleanup,
> + .reset = drm_atomic_helper_connector_reset,
> + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
> + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
> +};
> +
> +static int it6251_attach(struct drm_bridge *bridge)
> +{
> + struct it6251_bridge *it6251 = bridge_to_it6251(bridge);
> + int ret;
> +
> + if (!bridge->encoder) {
> + DRM_ERROR("Parent encoder object not found");
> + return -ENODEV;
> + }
> +
> + it6251->connector.polled = DRM_CONNECTOR_POLL_HPD;
> + ret = drm_connector_init(bridge->dev, &it6251->connector,
> + &it6251_connector_funcs,
> + DRM_MODE_CONNECTOR_eDP);
> + if (ret) {
> + DRM_ERROR("Failed to initialize connector with drm\n");
> + return ret;
> + }
> +
> + drm_atomic_helper_connector_reset(&it6251->connector);
> + drm_connector_helper_add(&it6251->connector,
> + &it6251_connector_helper_funcs);
> + drm_mode_connector_attach_encoder(&it6251->connector, bridge->encoder);
> +
> + if (it6251->panel)
> + drm_panel_attach(it6251->panel, &it6251->connector);
> +
> + drm_helper_hpd_irq_event(it6251->connector.dev);
The connector isn't registered at this point. So there isn't much benefit
of calling this here.
> +
> + return 0;
> +}
> +
> +static const struct drm_bridge_funcs it6251_bridge_funcs = {
> + .pre_enable = it6251_pre_enable,
> + .enable = it6251_enable,
> + .disable = it6251_disable,
> + .post_disable = it6251_post_disable,
> + .attach = it6251_attach,
> +};
> +
> +static const struct regmap_config it6251_regmap_config = {
> + .reg_bits = 8,
> + .val_bits = 8,
> + .max_register = 0xff,
> + .cache_type = REGCACHE_NONE,
> +};
> +
> +static int
> +it6251_probe(struct i2c_client *client, const struct i2c_device_id *id)
> +{
> + struct device *dev = &client->dev;
> + struct it6251_bridge *it6251;
> + struct device_node *endpoint, *panel_node;
> + int ret;
> +
> + it6251 = devm_kzalloc(dev, sizeof(*it6251), GFP_KERNEL);
> + if (!it6251)
> + return -ENOMEM;
> +
> + endpoint = of_graph_get_next_endpoint(dev->of_node, NULL);
There is an of_node_put() missing for this endpoint.
Thanks,
Archit
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
