linux-kernel - Re: [PATCH v3 4/5] drm/panel: samsung-s6e88a0-ams427ap24: Add brightness control

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CACRpkdYOgymfjOD3cAMXt7u8SH0vvVzwt75gamJvXuyyjdsMPw@mail.gmail.com>
Date: Fri, 25 Oct 2024 21:27:11 +0200
From: Linus Walleij <linus.walleij@...aro.org>
To: Jakob Hauser <jahau@...ketmail.com>
Cc: Neil Armstrong <neil.armstrong@...aro.org>, Jessica Zhang <quic_jesszhan@...cinc.com>, 
	Rob Herring <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley <conor+dt@...nel.org>, 
	Thierry Reding <thierry.reding@...il.com>, 
	Laurent Pinchart <laurent.pinchart+renesas@...asonboard.com>, 
	Andrzej Hajda <andrzej.hajda@...el.com>, dri-devel@...ts.freedesktop.org, 
	devicetree@...r.kernel.org, linux-kernel@...r.kernel.org, 
	~postmarketos/upstreaming@...ts.sr.ht
Subject: Re: [PATCH v3 4/5] drm/panel: samsung-s6e88a0-ams427ap24: Add
 brightness control

Hi Jakob,

thanks for your patch!

On Thu, Oct 24, 2024 at 5:18 AM Jakob Hauser <jahau@...ketmail.com> wrote:
> +static const int s6e88a0_ams427ap24_br_to_cd[NUM_STEPS_CANDELA] = {

(...)
> +       /* brightness till, candela */

Brightness to candela conversion table? Edit comment?

> +static const u8 s6e88a0_ams427ap24_aid[NUM_STEPS_AID][SEQ_LENGTH_AID] = {

If you know that the sequence 0xb2, 0x40, 0x08, 0x20 means "set AID"
(or is it AOR??) you can #define

#define S6E88A0_SET_AID 0xb2

Then make a small buffer:

u8 set_aid[5] = { S6E88A0_SET_AID, 0x40, 0x08, 0x20, 0x00, 0x00 };

then you can strip the first three bytes from the entire table,
just copy in the two relevant bytes into set_aor[]
and send that.

> +static const u8 s6e88a0_ams427ap24_elvss[NUM_STEPS_ELVSS][SEQ_LENGTH_ELVSS] = {
> +       { 0x28, 0x14 }, /* 10CD ~ 111CD */
> +       { 0x28, 0x13 }, /* 119CD */

Command 0xb6 is
#define S6E88A0_SET_LVSS 0xb6

Same comment: just define
u8 set_lvss[2] = {S6E88A0_SET_LVSS, 0x28, 0x00};
copy the second byte into the command, this becomes
an array of single bytes.

> +static const u8 s6e88a0_ams427ap24_gamma[NUM_STEPS_CANDELA][SEQ_LENGTH_GAMMA] = {
> +       /* 10CD */
> +       { 0x00, 0xc8, 0x00, 0xc4, 0x00, 0xc5, 0x8c, 0x8a, 0x8a, 0x8c, 0x8b,
> +         0x8c, 0x87, 0x89, 0x89, 0x88, 0x87, 0x8c, 0x80, 0x82, 0x88, 0x7b,
> +         0x72, 0x8c, 0x60, 0x68, 0x8c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 },

This array looks fine though, it seems to be all unique gamma calibration.

> +static int s6e88a0_ams427ap24_set_brightness(struct backlight_device *bd)
> +{
> +       struct s6e88a0_ams427ap24 *ctx = bl_get_data(bd);
> +       struct mipi_dsi_device *dsi = ctx->dsi;
> +       struct mipi_dsi_multi_context dsi_ctx = { .dsi = dsi };
> +       struct device *dev = &dsi->dev;
> +       int brightness = bd->props.brightness;
> +       int candela_enum;
> +       u8 b2[SEQ_LENGTH_AID + 1];
> +       u8 b6[SEQ_LENGTH_ELVSS + 1];
> +       u8 ca[SEQ_LENGTH_GAMMA + 1];

Rename them to something like my suggestions so we understand what it is
all about. It seems the infrastructure for what I suggested is mostly already
there.

See comment above how to modify arrays to contain stuff that is always
the same.

> +       /* get aid */
> +       b2[0] = 0xb2;

Use a define per above.

> +       switch (candela_enum) {
> +       case CANDELA_10CD ... CANDELA_105CD:
> +               memcpy(&b2[1], s6e88a0_ams427ap24_aid[candela_enum],
> +                      SEQ_LENGTH_AID);
> +               break;
> +       case CANDELA_111CD ... CANDELA_172CD:
> +               memcpy(&b2[1], s6e88a0_ams427ap24_aid[CANDELA_111CD],
> +                      SEQ_LENGTH_AID);
> +               break;
> +       case CANDELA_183CD ... CANDELA_300CD:
> +               memcpy(&b2[1], s6e88a0_ams427ap24_aid[CANDELA_111CD + 1],
> +                      SEQ_LENGTH_AID);
> +               break;
> +       default:
> +               dev_err(dev, "Failed to get aid data\n");
> +               return -EINVAL;
> +       }
> +
> +       /* get elvss */
> +       b6[0] = 0xb6;

Use a define per above.

> +       if (candela_enum <= CANDELA_111CD) {
> +               memcpy(&b6[1], s6e88a0_ams427ap24_elvss[0], SEQ_LENGTH_ELVSS);
> +       } else {
> +               memcpy(&b6[1], s6e88a0_ams427ap24_elvss[candela_enum - CANDELA_111CD],
> +                      SEQ_LENGTH_ELVSS);
> +       }
> +
> +       /* get gamma */
> +       ca[0] = 0xca;

#define S6E88A0_SET_GAMMA 0xca

> +       memcpy(&ca[1], s6e88a0_ams427ap24_gamma[candela_enum], SEQ_LENGTH_GAMMA);
> +
> +       /* write: key on, aid, acl off, elvss, gamma, gamma update, key off */
> +       mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0xf0, 0x5a, 0x5a);

0xf0 is clearly an unlocking key as per comment in the previous patch.

> +       mipi_dsi_dcs_write_buffer_multi(&dsi_ctx, b2, ARRAY_SIZE(b2));
> +       mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0x55, 0x00);

0x55 is MIPI_DCS_WRITE_POWER_SAVE in <video/mipi_display.h>

> +       mipi_dsi_dcs_write_buffer_multi(&dsi_ctx, b6, ARRAY_SIZE(b6));
> +       mipi_dsi_dcs_write_buffer_multi(&dsi_ctx, ca, ARRAY_SIZE(ca));

> +       mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0xf7, 0x03);
> +       mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0xf0, 0xa5, 0xa5);

Clearly this locks the L2 access again.

Yours,
Linus Walleij