From mboxrd@z Thu Jan  1 00:00:00 1970
From: Thierry Reding <thierry.reding@gmail.com>
Subject: Re: [PATCH v4 1/2] drm/panel: Add support for Truly NT35597 panel
Date: Tue, 10 Jul 2018 11:44:44 +0200
Message-ID: <20180710094444.GB1504@ulmo>
References: <1527294472-4451-1-git-send-email-abhinavk@codeaurora.org>
 <20180530191247.GO3373@art_vandelay>
 <0beaa3a7765dbf9bfffdee59235f180d@codeaurora.org>
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="===============1143224104=="
Return-path: <dri-devel-bounces@lists.freedesktop.org>
In-Reply-To: <0beaa3a7765dbf9bfffdee59235f180d@codeaurora.org>
List-Unsubscribe: <https://lists.freedesktop.org/mailman/options/dri-devel>,
 <mailto:dri-devel-request@lists.freedesktop.org?subject=unsubscribe>
List-Archive: <https://lists.freedesktop.org/archives/dri-devel>
List-Post: <mailto:dri-devel@lists.freedesktop.org>
List-Help: <mailto:dri-devel-request@lists.freedesktop.org?subject=help>
List-Subscribe: <https://lists.freedesktop.org/mailman/listinfo/dri-devel>,
 <mailto:dri-devel-request@lists.freedesktop.org?subject=subscribe>
Errors-To: dri-devel-bounces@lists.freedesktop.org
Sender: "dri-devel" <dri-devel-bounces@lists.freedesktop.org>
To: abhinavk@codeaurora.org
Cc: linux-arm-msm@vger.kernel.org, dri-devel@lists.freedesktop.org, hoegsberg@google.com, chandanu@codeaurora.org
List-Id: linux-arm-msm@vger.kernel.org


--===============1143224104==
Content-Type: multipart/signed; micalg=pgp-sha256;
	protocol="application/pgp-signature"; boundary="2B/JsCI69OhZNC5r"
Content-Disposition: inline


--2B/JsCI69OhZNC5r
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, May 30, 2018 at 03:37:26PM -0700, abhinavk@codeaurora.org wrote:
> Hi Sean
>=20
> Thanks for your review.
>=20
> Some responses below. Please help to check.
>=20
> Thanks
>=20
> Abhinav

Looks like this got stuck in review. Some comments below, which I hope
will match Sean's opinion.

> On 2018-05-30 12:12, Sean Paul wrote:
> > On Fri, May 25, 2018 at 05:27:51PM -0700, Abhinav Kumar wrote:
> > > Add support for Truly NT35597 panel used
> > > in MSM reference platforms.
> > >=20
> > > This panel supports both single DSI and dual DSI
> > > modes.
> > >=20
> > > However, this patch series adds support only for
> > > dual DSI mode.
> > >=20
> > > Changes in v4:
> > > - Fix the license identifier
> > > - Fix formatting issues for the regulator loads
> > > - Fix error messages and return code
> > >=20
> > > Signed-off-by: Archit Taneja <architt@codeaurora.org>
> > > Signed-off-by: Abhinav Kumar <abhinavk@codeaurora.org>
> > > ---
> > >  drivers/gpu/drm/panel/Kconfig               |   8 +
> > >  drivers/gpu/drm/panel/Makefile              |   1 +
> > >  drivers/gpu/drm/panel/panel-truly-nt35597.c | 576
> > > ++++++++++++++++++++++++++++
> > >  3 files changed, 585 insertions(+)
> > >  create mode 100644 drivers/gpu/drm/panel/panel-truly-nt35597.c
> > >=20
> > > diff --git a/drivers/gpu/drm/panel/Kconfig
> > > b/drivers/gpu/drm/panel/Kconfig
> > > index 25682ff..2fcd9b1 100644
> > > --- a/drivers/gpu/drm/panel/Kconfig
> > > +++ b/drivers/gpu/drm/panel/Kconfig
> > > @@ -177,4 +177,12 @@ config DRM_PANEL_SITRONIX_ST7789V
> > >  	  Say Y here if you want to enable support for the Sitronix
> > >  	  ST7789V controller for 240x320 LCD panels
> > >=20
> > > +config DRM_PANEL_TRULY_NT35597_WQXGA
> > > +	tristate "Truly WQXGA"
> > > +	depends on OF
> > > +	depends on DRM_MIPI_DSI
> > > +	select VIDEOMODE_HELPERS
> > > +	help
> > > +	  Say Y here if you want to enable support for Truly NT35597 WQXGA
> > > Dual DSI
> > > +	  Video Mode panel
> > >  endmenu
> > > diff --git a/drivers/gpu/drm/panel/Makefile
> > > b/drivers/gpu/drm/panel/Makefile
> > > index f26efc1..056ea93 100644
> > > --- a/drivers/gpu/drm/panel/Makefile
> > > +++ b/drivers/gpu/drm/panel/Makefile
> > > @@ -18,3 +18,4 @@ obj-$(CONFIG_DRM_PANEL_SEIKO_43WVF1G) +=3D
> > > panel-seiko-43wvf1g.o
> > >  obj-$(CONFIG_DRM_PANEL_SHARP_LQ101R1SX01) +=3D
> > > panel-sharp-lq101r1sx01.o
> > >  obj-$(CONFIG_DRM_PANEL_SHARP_LS043T1LE01) +=3D
> > > panel-sharp-ls043t1le01.o
> > >  obj-$(CONFIG_DRM_PANEL_SITRONIX_ST7789V) +=3D panel-sitronix-st7789v=
=2Eo
> > > +obj-$(CONFIG_DRM_PANEL_TRULY_NT35597_WQXGA) +=3D panel-truly-nt35597=
=2Eo
> > > diff --git a/drivers/gpu/drm/panel/panel-truly-nt35597.c
> > > b/drivers/gpu/drm/panel/panel-truly-nt35597.c
> > > new file mode 100644
> > > index 0000000..a57cbf0
> > > --- /dev/null
> > > +++ b/drivers/gpu/drm/panel/panel-truly-nt35597.c
> > > @@ -0,0 +1,576 @@
> > > +// SPDX-License-Identifier: GPL-2.0
> > > +/*
> > > + * Copyright (c) 2018, The Linux Foundation. All rights reserved.
> > > + */
> > > +
> > > +#include <linux/gpio/consumer.h>
> > > +#include <linux/of_graph.h>
> > > +#include <linux/regulator/consumer.h>
> > > +#include <linux/pinctrl/consumer.h>
> > > +
> > > +#include <video/mipi_display.h>
> > > +#include <video/of_videomode.h>
> > > +#include <video/videomode.h>
> > > +
> > > +#include <drm/drmP.h>
> > > +#include <drm/drm_panel.h>
> > > +#include <drm/drm_mipi_dsi.h>
> > > +
> > > +static const char * const regulator_names[] =3D {
> > > +	"vdda",
> > > +	"vdispp",
> > > +	"vdispn"
> > > +};
> [Abhinav] Will add a ',' here as per jordan's comment
> > > +
> > > +static unsigned long const regulator_enable_loads[] =3D {
> > > +	62000,
> > > +	100000,
> > > +	100000,
> > > +};
> > > +
> > > +static unsigned long const regulator_disable_loads[] =3D {
> > > +	80,
> > > +	100,
> > > +	100,
> > > +};
> > > +
> > > +struct truly_wqxga {
> > > +	struct device *dev;
> > > +	struct drm_panel panel;
> > > +
> > > +	struct regulator_bulk_data supplies[ARRAY_SIZE(regulator_names)];
> > > +
> > > +	struct gpio_desc *reset_gpio;
> > > +	struct gpio_desc *mode_gpio;
> > > +
> > > +	struct backlight_device *backlight;
> > > +	struct videomode vm;
> > > +
> > > +	struct mipi_dsi_device *dsi[2];
> > > +
> > > +	bool prepared;
> > > +	bool enabled;
> > > +};
> > > +
> > > +static inline struct truly_wqxga *panel_to_truly_wqxga(struct
> > > drm_panel *panel)
> > > +{
> > > +	return container_of(panel, struct truly_wqxga, panel);
> > > +}
> > > +
> > > +static int truly_wqxga_power_on(struct truly_wqxga *ctx)
> > > +{
> > > +	int ret, i;
> > > +
> > > +	for (i =3D 0; i < ARRAY_SIZE(ctx->supplies); i++) {
> > > +		ret =3D regulator_set_load(ctx->supplies[i].consumer,
> > > +					regulator_enable_loads[i]);
> > > +		if (ret)
> > > +			return ret;
> > > +	}
> > > +
> > > +	ret =3D regulator_bulk_enable(ARRAY_SIZE(ctx->supplies),
> > > ctx->supplies);
> > > +	if (ret < 0)
> > > +		return ret;
> > > +
> > > +	msleep(20);
> > > +	gpiod_set_value(ctx->reset_gpio, 1);
> > > +	msleep(20);
> > > +	gpiod_set_value(ctx->reset_gpio, 0);
> > > +	msleep(20);
> > > +	gpiod_set_value(ctx->reset_gpio, 1);
> > > +	msleep(50);
> >=20
> > Why is this needed? Could you please comment?
> >=20
> > Also, it seems like this is active low? You should specify that in the
> > dt, and
> > let gpiod translate the value.
>=20
> [Abhinav] This is the panel's reset sequence which demands this delay. I =
can
> comment the same.
> Its active HIGH. I can leave a comment in the bindings mentioning the sam=
e.
> Let me know if this is sufficient.

According to the code that reset is active low, otherwise that last
gpiod_set_value() would leave the reset asserted. In either case I think
what Sean meant to say was that the code here should be assuming that
the reset is active high (i.e. the reset deassert should be
gpiod_set_value(..., 0) and the device tree should have GPIO_ACTIVE_LOW
in the specifier for the reset GPIO, which will then cause the gpiod
library to properly do the inversion for you.

>=20
> >=20
> > > +
> > > +	return 0;
> > > +}
> > > +
> > > +static int truly_wqxga_power_off(struct truly_wqxga *ctx)
> > > +{
> > > +	int ret, i;
> > > +
> > > +	gpiod_set_value(ctx->reset_gpio, 0);
> > > +
> > > +	for (i =3D 0; i < ARRAY_SIZE(ctx->supplies); i++) {
> > > +		ret =3D regulator_set_load(ctx->supplies[i].consumer,
> > > +				regulator_disable_loads[i]);
> > > +		if (ret)
> > > +			return ret;
> > > +	}
> > > +
> > > +	return regulator_bulk_disable(ARRAY_SIZE(ctx->supplies),
> > > ctx->supplies);
> > > +}
> > > +
> > > +static int truly_wqxga_disable(struct drm_panel *panel)
> > > +{
> > > +	struct truly_wqxga *ctx =3D panel_to_truly_wqxga(panel);
> > > +
> > > +	if (!ctx->enabled)
> > > +		return 0;
> > > +
> > > +	if (ctx->backlight) {
> > > +		ctx->backlight->props.power =3D FB_BLANK_POWERDOWN;
> > > +		backlight_update_status(ctx->backlight);
> >=20
> > You can use the backlight_enable/disable helpers, and please check the
> > return
> > value.
> >=20
> [Abhinav] Can you please provide me the reference to the helper you are
> referring to?
> Other panels were using backlight_update_status(), Hence i am not able to
> locate it.
> Yes, I will check the return value in v5.

I think Sean was referring to backlight_enable() and backlight_disable()
in include/linux/backlight.h. You should use those instead of the low-
level backlight_update_status(). If you look at the implementation of
the new helpers, you'll understand why. =3D)

>=20
> > > +	}
> > > +
> > > +	ctx->enabled =3D false;
> > > +	return 0;
> > > +}
> > > +
> > > +static int truly_wqxga_unprepare(struct drm_panel *panel)
> > > +{
> > > +	struct truly_wqxga *ctx =3D panel_to_truly_wqxga(panel);
> > > +	struct mipi_dsi_device **dsis =3D ctx->dsi;
> >=20
> > IMO, this obfuscation isn't too useful (and prevents us from using
> > ARRAY_SIZE
> > below).
> [Abhinav] Agreed, will fix this in v5.
> >=20
> > > +	int ret =3D 0, i;
> > > +
> > > +	if (!ctx->prepared)
> > > +		return 0;
> > > +
> > > +	dsis[0]->mode_flags =3D 0;
> > > +	dsis[1]->mode_flags =3D 0;
> > > +
> > > +	for (i =3D 0; i < 2; i++)
> >=20
> > s/2/ARRAY_SIZE(ctx->dsi)/
> >=20
> > > +		if (mipi_dsi_dcs_set_display_off(dsis[i]) < 0)
> > > +			ret =3D -ECOMM;
> >=20
> > I guess it's alright that we're continuing through errors, however we
> > should
> > probably log the actual error value as well as the channel number in a
> > DEBUG
> > message.
> >=20
> [Abhinav] Agreed, will print an error message.
> > > +
> > > +	msleep(120);
> >=20
> > Please comment.
> [Abhinav] Agreed, will explain this.
> >=20
> > > +
> > > +	for (i =3D 0; i < 2; i++)
> >=20
> > ARRAY_SIZE nit here.
> >=20
> [Abhinav] Agreed, will use ARRAY_SIZE here
> > > +		if (mipi_dsi_dcs_enter_sleep_mode(dsis[i]) < 0)
> > > +			ret =3D -ECOMM;
> > > +
> > > +	truly_wqxga_power_off(ctx);
> >=20
> > You don't check the return value.
> [Abhinav] Agreed, will check return value.
> >=20
> > > +
> > > +	ctx->prepared =3D false;
> > > +	return ret;
> > > +}
> > > +
> > > +#define MAX_LEN	5

That's odd, you set MAX_LEN to 5, but the longest command in the
sequence below is 4 bytes. That means you're effectively wasting at
least 1 byte per command.

> > > +struct {
> > > +	u8 commands[MAX_LEN];
> > > +	int size;

This can be unsigned.

> > > +} panel_cmds[] =3D { /* CMD2_P0 */

You probably want to make this static and const?

> > > +		   { { 0xff, 0x20 }, 2 },
> > > +		   { { 0xfb, 0x01 }, 2 },
> > > +		   { { 0x00, 0x01 }, 2 },
> > > +		   { { 0x01, 0x55 }, 2 },
> > > +		   { { 0x02, 0x45 }, 2 },
> > > +		   { { 0x05, 0x40 }, 2 },
> > > +		   { { 0x06, 0x19 }, 2 },
> > > +		   { { 0x07, 0x1e }, 2 },
> > > +		   { { 0x0b, 0x73 }, 2 },
> > > +		   { { 0x0c, 0x73 }, 2 },
> > > +		   { { 0x0e, 0xb0 }, 2 },
> > > +		   { { 0x0f, 0xae }, 2 },
> > > +		   { { 0x11, 0xb8 }, 2 },
> > > +		   { { 0x13, 0x00 }, 2 },
> > > +		   { { 0x58, 0x80 }, 2 },
> > > +		   { { 0x59, 0x01 }, 2 },
> > > +		   { { 0x5a, 0x00 }, 2 },
> > > +		   { { 0x5b, 0x01 }, 2 },
> > > +		   { { 0x5c, 0x80 }, 2 },
> > > +		   { { 0x5d, 0x81 }, 2 },
> > > +		   { { 0x5e, 0x00 }, 2 },
> > > +		   { { 0x5f, 0x01 }, 2 },
> > > +		   { { 0x72, 0x11 }, 2 },
> > > +		   { { 0x68, 0x03 }, 2 },
> > > +		   /* CMD2_P4 */
> > > +		   { { 0xFF, 0x24 }, 2 },
> > > +		   { { 0xFB, 0x01 }, 2 },
> > > +		   { { 0x00, 0x1C }, 2 },
> > > +		   { { 0x01, 0x0B }, 2 },
> > > +		   { { 0x02, 0x0C }, 2 },
> > > +		   { { 0x03, 0x01 }, 2 },
> > > +		   { { 0x04, 0x0F }, 2 },
> > > +		   { { 0x05, 0x10 }, 2 },
> > > +		   { { 0x06, 0x10 }, 2 },
> > > +		   { { 0x07, 0x10 }, 2 },
> > > +		   { { 0x08, 0x89 }, 2 },
> > > +		   { { 0x09, 0x8A }, 2 },
> > > +		   { { 0x0A, 0x13 }, 2 },
> > > +		   { { 0x0B, 0x13 }, 2 },
> > > +		   { { 0x0C, 0x15 }, 2 },
> > > +		   { { 0x0D, 0x15 }, 2 },
> > > +		   { { 0x0E, 0x17 }, 2 },
> > > +		   { { 0x0F, 0x17 }, 2 },
> > > +		   { { 0x10, 0x1C }, 2 },
> > > +		   { { 0x11, 0x0B }, 2 },
> > > +		   { { 0x12, 0x0C }, 2 },
> > > +		   { { 0x13, 0x01 }, 2 },
> > > +		   { { 0x14, 0x0F }, 2 },
> > > +		   { { 0x15, 0x10 }, 2 },
> > > +		   { { 0x16, 0x10 }, 2 },
> > > +		   { { 0x17, 0x10 }, 2 },
> > > +		   { { 0x18, 0x89 }, 2 },
> > > +		   { { 0x19, 0x8A }, 2 },
> > > +		   { { 0x1A, 0x13 }, 2 },
> > > +		   { { 0x1B, 0x13 }, 2 },
> > > +		   { { 0x1C, 0x15 }, 2 },
> > > +		   { { 0x1D, 0x15 }, 2 },
> > > +		   { { 0x1E, 0x17 }, 2 },
> > > +		   { { 0x1F, 0x17 }, 2 },
> > > +		   /* STV */
> > > +		   { { 0x20, 0x40 }, 2 },
> > > +		   { { 0x21, 0x01 }, 2 },
> > > +		   { { 0x22, 0x00 }, 2 },
> > > +		   { { 0x23, 0x40 }, 2 },
> > > +		   { { 0x24, 0x40 }, 2 },
> > > +		   { { 0x25, 0x6D }, 2 },
> > > +		   { { 0x26, 0x40 }, 2 },
> > > +		   { { 0x27, 0x40 }, 2 },
> > > +		   /* Vend */
> > > +		   { { 0xE0, 0x00 }, 2 },
> > > +		   { { 0xDC, 0x21 }, 2 },
> > > +		   { { 0xDD, 0x22 }, 2 },
> > > +		   { { 0xDE, 0x07 }, 2 },
> > > +		   { { 0xDF, 0x07 }, 2 },
> > > +		   { { 0xE3, 0x6D }, 2 },
> > > +		   { { 0xE1, 0x07 }, 2 },
> > > +		   { { 0xE2, 0x07 }, 2 },
> > > +		   /* UD */
> > > +		   { { 0x29, 0xD8 }, 2 },
> > > +		   { { 0x2A, 0x2A }, 2 },
> > > +		   /* CLK */
> > > +		   { { 0x4B, 0x03 }, 2 },
> > > +		   { { 0x4C, 0x11 }, 2 },
> > > +		   { { 0x4D, 0x10 }, 2 },
> > > +		   { { 0x4E, 0x01 }, 2 },
> > > +		   { { 0x4F, 0x01 }, 2 },
> > > +		   { { 0x50, 0x10 }, 2 },
> > > +		   { { 0x51, 0x00 }, 2 },
> > > +		   { { 0x52, 0x80 }, 2 },
> > > +		   { { 0x53, 0x00 }, 2 },
> > > +		   { { 0x56, 0x00 }, 2 },
> > > +		   { { 0x54, 0x07 }, 2 },
> > > +		   { { 0x58, 0x07 }, 2 },
> > > +		   { { 0x55, 0x25 }, 2 },
> > > +		   /* Reset XDONB */
> > > +		   { { 0x5B, 0x43 }, 2 },
> > > +		   { { 0x5C, 0x00 }, 2 },
> > > +		   { { 0x5F, 0x73 }, 2 },
> > > +		   { { 0x60, 0x73 }, 2 },
> > > +		   { { 0x63, 0x22 }, 2 },
> > > +		   { { 0x64, 0x00 }, 2 },
> > > +		   { { 0x67, 0x08 }, 2 },
> > > +		   { { 0x68, 0x04 }, 2 },
> > > +		   /* Resolution:1440x2560 */
> > > +		   { { 0x72, 0x02 }, 2 },
> > > +		   /* mux */
> > > +		   { { 0x7A, 0x80 }, 2 },
> > > +		   { { 0x7B, 0x91 }, 2 },
> > > +		   { { 0x7C, 0xD8 }, 2 },
> > > +		   { { 0x7D, 0x60 }, 2 },
> > > +		   { { 0x7F, 0x15 }, 2 },
> > > +		   { { 0x75, 0x15 }, 2 },
> > > +		   /* ABOFF */
> > > +		   { { 0xB3, 0xC0 }, 2 },
> > > +		   { { 0xB4, 0x00 }, 2 },
> > > +		   { { 0xB5, 0x00 }, 2 },
> > > +		   /* Source EQ */
> > > +		   { { 0x78, 0x00 }, 2 },
> > > +		   { { 0x79, 0x00 }, 2 },
> > > +		   { { 0x80, 0x00 }, 2 },
> > > +		   { { 0x83, 0x00 }, 2 },
> > > +		   /* FP BP */
> > > +		   { { 0x93, 0x0A }, 2 },
> > > +		   { { 0x94, 0x0A }, 2 },
> > > +		   /* Inversion Type */
> > > +		   { { 0x8A, 0x00 }, 2 },
> > > +		   { { 0x9B, 0xFF }, 2 },
> > > +		   /* IMGSWAP =3D1 @PortSwap=3D1 */
> > > +		   { { 0x9D, 0xB0 }, 2 },
> > > +		   { { 0x9F, 0x63 }, 2 },
> > > +		   { { 0x98, 0x10 }, 2 },
> > > +		   /* FRM */
> > > +		   { { 0xEC, 0x00 }, 2 },
> > > +		   /* CMD1 */
> > > +		   { { 0xFF, 0x10 }, 2 },
> > > +		    /* VBP+VSA=3D,VFP =3D 10H */
> > > +		   { { 0x3B, 0x03, 0x0A, 0x0A, }, 4 },
> > > +		   /* FTE on */
> > > +		   { { 0x35, 0x00 }, 2 },
> > > +		   /* EN_BK =3D1(auto black) */
> > > +		   { { 0xE5, 0x01 }, 2 },
> > > +		   /* CMD mode(10) VDO mode(03) */
> > > +		   { { 0xBB, 0x03 }, 2 },
> > > +		   /* Non Reload MTP */
> > > +		   { { 0xFB, 0x01 }, 2 },
> > > +};
> >=20
> > How many of these commands are specific to MTP? Decoding the cryptic
> > comments,
> > it seems like this probably wouldn't work on any other boards?
> >=20
> > I see a couple of duplicate writes as well.
> >=20
> > Finally, all but one command is 2 bytes, with the outlier being 4 bytes.
> > Yet,
> > MAX_LEN is 5. If we _must_ keep the huge list of obfuscated register
> > writes,
> > could we eliminate the size member?
>=20
> [Abhinav] The name is probably misleading you here. MTP mentioned here is
> something
> specific to the panel. It has no relation to our MTP board.
> Hence none of these commands are specific to our MTP board. They are all
> related to
> the panel.
>=20
> There are no duplicates. The sequences of the panel often demand the same
> command to
> be sent again in different places.
>=20
> You are right, all the commands except one have 2 bytes of payload.
> However, the 4 byte long command has to be sent in between.
> If we want to get rid of the extra member element, I can break the comman=
ds
> into 3
> sets.
> First set of x commands with 2 byte payload, then the 4 byte payload
> command, then the
> remaining 2 byte payload commands.
> Let me know if that works for you.
>=20
> I have tried to leave as much comments as possible about the command sets.
> It will be difficult to explain more.

Given your explanation, I think the above is fine. Modulo the MAX_LEN,
signedness and static const comments.

> > > +static int truly_wqxga_probe(struct mipi_dsi_device *dsi)
> > > +{
> > > +	struct device *dev =3D &dsi->dev;
> > > +	struct truly_wqxga *ctx;
> > > +	struct mipi_dsi_device *secondary;
> > > +	struct device_node *dsi1;
> > > +	struct mipi_dsi_host *dsi1_host;
> > > +	int ret =3D 0;
> > > +	const struct mipi_dsi_device_info info =3D {.type =3D "trulynt35597=
",
> > > +		.channel =3D 0,
> > > +		.node =3D NULL,
> > > +	};
> > > +
> > > +	ctx =3D devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
> > > +
> > > +	if (!ctx)
> > > +		return -ENOMEM;
> > > +
> > > +	/* This device represents itself as one with
> > > +	 * two input ports which are fed by the output
> > > +	 * ports of the two DSI controllers . The DSI0
> > > +	 * is the master controller and has most of the
> > > +	 * panel related info in its child node.
> > > +	 */
> > > +
> > > +	dsi1 =3D of_graph_get_remote_node(dsi->dev.of_node, 1, -1);
> > > +	if (!dsi1) {
> > > +		dev_err(dev, "failed to get remote node for secondary DSI\n");
> > > +		ret =3D -ENODEV;
> > > +		goto err_get_remote;
> > > +	}
> > > +
> > > +	dsi1_host =3D of_find_mipi_dsi_host_by_node(dsi1);
> > > +	if (!dsi1_host) {
> > > +		dev_err(dev, "failed to find dsi host\n");
> > > +		ret =3D -EPROBE_DEFER;
> > > +		goto err_host;
> > > +	}
> > > +
> > > +	of_node_put(dsi1);
> > > +
> > > +	/* register the second DSI device */
> > > +	secondary =3D mipi_dsi_device_register_full(dsi1_host, &info);

Since this does dual DSI, are you aware of the dual-channel DSI bindings
that Archit (Cc'ed) has recently proposed?

	https://patchwork.freedesktop.org/patch/237057/

> > > +
> >=20
> > nit: extra line
> >=20
> > While I'm nitpicking, this name leaves something to be desired.
> > Everything else
> > is dsi1, so perhaps this could be dsi1_device to be consistent.
> >=20
> [Abhinav] Agreed, will do.
>=20
> > > +	if (IS_ERR(secondary)) {
> > > +		dev_err(dev, "failed to create dsi device\n");
> > > +		ret =3D PTR_ERR(dsi);
> > > +		goto err_dsi_device;
> > > +	}
> > > +
> > > +	mipi_dsi_set_drvdata(dsi, ctx);
> > > +
> > > +	ctx->dev =3D dev;
> > > +	ctx->dsi[0] =3D dsi;
> > > +	ctx->dsi[1] =3D secondary;
> > > +
> > > +	ret =3D truly_wqxga_panel_add(ctx);
> > > +	if (ret) {
> > > +		dev_err(dev, "failed to add panel\n");
> > > +		goto err_panel_add;
> > > +	}
> > > +
> > > +	/* configure master DSI device */
> > > +	dsi->lanes =3D 4;
> > > +	dsi->format =3D MIPI_DSI_FMT_RGB888;
> > > +	dsi->mode_flags =3D MIPI_DSI_MODE_VIDEO |
> > > MIPI_DSI_CLOCK_NON_CONTINUOUS |
> > > +						MIPI_DSI_MODE_LPM;
> > > +
> > > +	ret =3D mipi_dsi_attach(dsi);
> > > +	if (ret < 0) {
> > > +		dev_err(dev, "master dsi attach failed\n");
> > > +		goto err_dsi_attach;
> > > +	}
> > > +
> > > +	/* configure secondary DSI device */
> > > +	secondary->lanes =3D 4;
> > > +	secondary->format =3D MIPI_DSI_FMT_RGB888;
> > > +	secondary->mode_flags =3D MIPI_DSI_MODE_VIDEO |
> > > +		MIPI_DSI_CLOCK_NON_CONTINUOUS |
> > > +		MIPI_DSI_MODE_LPM;
> > > +
> > > +	ret =3D mipi_dsi_attach(secondary);
> > > +	if (ret < 0) {
> > > +		dev_err(dev, "mipi_dsi_attach on secondary failed\n");
> > > +		goto err_dsi_attach_sec;
> > > +	}
> >=20
> > Replace the duplicated attach code with a loop:
> >=20
> >         for (i =3D 0; i < ARRAY_SIZE(ctx->dsi); i++) {
> >                 struct mipi_dsi_device *dsi_dev =3D ctx->dsi[i];
> >=20
> >                 dsi_dev->lanes =3D 4;
> >                 dsi_dev->format =3D MIPI_DSI_FMT_RGB888;
> >                 dsi_dev->mode_flags =3D MIPI_DSI_MODE_VIDEO |
> > MIPI_DSI_MODE_LPM |
> >                                       MIPI_DSI_CLOCK_NON_CONTINUOUS;
> >=20
> >                 ret =3D mipi_dsi_attach(dsi_dev);
> >                 if (ret < 0) {
> >                         DRM_DEV_ERROR(dev, "master dsi attach
> > failed\n");
> >                         goto err_dsi_attach;
> >                 }
> >         }
> >=20
> [Abhinav] Agreed, will do.
> > > +
> > > +	return 0;
> > > +
> > > +err_dsi_attach_sec:
> > > +	mipi_dsi_detach(ctx->dsi[0]);
> > > +err_dsi_attach:
> > > +	truly_wqxga_panel_del(ctx);
> > > +err_panel_add:
> > > +	mipi_dsi_device_unregister(secondary);
> > > +err_dsi_device:
> > > +err_host:
> > > +	of_node_put(dsi1);
> > > +err_get_remote:
> > > +	return ret;
> > > +}
> > > +
> > > +static int truly_wqxga_remove(struct mipi_dsi_device *dsi)
> > > +{
> > > +	struct truly_wqxga *ctx =3D mipi_dsi_get_drvdata(dsi);
> > > +
> > > +	if (ctx->dsi[0])
> > > +		mipi_dsi_detach(ctx->dsi[0]);
> > > +	if (ctx->dsi[1]) {
> > > +		mipi_dsi_detach(ctx->dsi[1]);
> > > +		mipi_dsi_device_unregister(ctx->dsi[1]);
> > > +	}
> > > +	truly_wqxga_panel_del(ctx);
> > > +
> > > +	return 0;
> > > +}
> > > +
> > > +static const struct of_device_id truly_wqxga_of_match[] =3D {
> > > +	{ .compatible =3D "truly,nt35597", },
> > > +	{ }
> > > +};
> > > +MODULE_DEVICE_TABLE(of, truly_wqxga_of_match);
> > > +
> > > +static struct mipi_dsi_driver truly_wqxga_driver =3D {
> > > +	.driver =3D {
> > > +		.name =3D "panel_truly_nt35597",

I generally prefer dashes to underscores in driver names, in this case:
"panel-truly-nt35597".

Thierry

--2B/JsCI69OhZNC5r
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
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=kOWp
-----END PGP SIGNATURE-----

--2B/JsCI69OhZNC5r--

--===============1143224104==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: base64
Content-Disposition: inline

X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KZHJpLWRldmVs
IG1haWxpbmcgbGlzdApkcmktZGV2ZWxAbGlzdHMuZnJlZWRlc2t0b3Aub3JnCmh0dHBzOi8vbGlz
dHMuZnJlZWRlc2t0b3Aub3JnL21haWxtYW4vbGlzdGluZm8vZHJpLWRldmVsCg==

--===============1143224104==--