Hello Marek,
Thank you for your feedback,
On Tue, May 12, 2026 at 08:29:55PM +0200, Marek Vasut wrote:
> On 5/12/26 6:46 PM, Boerge Struempfel wrote:
>
> > +/* Maps register value (index) to drive-strength-microamp DT property
> > value */
> > +static const u32 lt9211_hsdrv_microamp[] = {
>
> This can be u8 .
>
you are right, I'll change that in the v2.
> > + 12, 14, 16, 17, 19, 20, 22, 23, 25, 27, 28, 30, 31, 33, 34, 36
> > +};
> > +
> > struct lt9211 {
> > struct drm_bridge bridge;
> > struct device *dev;
> > @@ -50,6 +55,7 @@ struct lt9211 {
> > struct regulator *vccio;
> > bool lvds_dual_link;
> > bool lvds_dual_link_even_odd_swap;
> > + u8 lvds_hsdrv_isel;
> > };
> >
> > static const struct regmap_range lt9211_rw_ranges[] = {
> > @@ -374,7 +380,8 @@ static int lt9211_configure_tx(struct lt9211 *ctx, bool
> > jeida,
> > /* BIT(7) is LVDS dual-port */
> > { 0x823b, 0x38 | (ctx->lvds_dual_link ? BIT(7) : 0) },
> > { 0x823e, 0x92 },
> > - { 0x823f, 0x48 },
> > + /* bits 3:0: RG_MLTX_HSDRV_ISEL, LVDS TX driver current */
> > + { 0x823f, 0x40 | ctx->lvds_hsdrv_isel },
> > { 0x8240, 0x31 },
> > { 0x8243, 0x80 },
> > { 0x8244, 0x00 },
> > @@ -629,7 +636,9 @@ static int lt9211_parse_dt(struct lt9211 *ctx)
> > struct device *dev = ctx->dev;
> > struct drm_panel *panel;
> > int dual_link;
> > + u32 microamp;
> > int ret;
> > + int i;
> >
> > ctx->vccio = devm_regulator_get(dev, "vccio");
> > if (IS_ERR(ctx->vccio))
> > @@ -666,6 +675,23 @@ static int lt9211_parse_dt(struct lt9211 *ctx)
> >
> > ctx->panel_bridge = panel_bridge;
> >
> > + ctx->lvds_hsdrv_isel = 8; /* default: 25 uA */
> > + ret = of_property_read_u32(dev->of_node, "drive-strength-microamp",
> > + µamp);
>
> if ret != 0 , then what happens here ?
>
if ret != 0, we will keep the default value of 8 (corresponding to 25 uA
as written above), which is the behavior that the driver had previously.
According to the documentation, of_property_read_u32() will return 0 on
success, -EINVAL if the property does not exist, -ENODATA if property
does not have a value, and -EOVERFLOW if the property data isn't large
enough. I think we only would need to give a warning or similar if the
-ENODATA or -EOVERFLOW cases. However, I personally do not think that is
necessary, as the usage is specified unambiguously in the devicetree
bindings.
If you feel like we should add a warning in those cases, I can add them
however.
> > + if (!ret) {
> > + for (i = 0; i < ARRAY_SIZE(lt9211_hsdrv_microamp); i++) {
> > + if (lt9211_hsdrv_microamp[i] == microamp) {
> > + ctx->lvds_hsdrv_isel = i;
> > + break;
> > + }
> > + }
> > + if (i == ARRAY_SIZE(lt9211_hsdrv_microamp)) {
> > + dev_err(dev, "Invalid drive-strength-microamp value
> > %u\n",
> > + microamp);
> > + return -EINVAL;
> > + }
> > + }
> > +
> > return 0;
> [...]
>
--
Best regards,
Börge Strümpfel
DATA MODUL AG
Landsberger Str. 322
80687 München
