On Wed, Jul 30, 2025 at 07:02:47PM +0200, Louis Chauvet wrote:
> As stated in the AM62x Technical Reference Manual (SPRUIV7B), the data
> sampling edge needs to be configured in two distinct registers: one in the
> TIDSS IP and another in the memory-mapped control register modules. Since
> the latter is not within the same address range, a phandle to a syscon
> device is used to access the regmap.
>
> Fixes: 32a1795f57ee ("drm/tidss: New driver for TI Keystone platform Display
> SubSystem")
> Signed-off-by: Louis Chauvet <louis.chau...@bootlin.com>
>
> ---
>
> Cc: sta...@vger.kernel.org
Please read docs how to add stable tags.
> ---
> drivers/gpu/drm/tidss/tidss_dispc.c | 14 ++++++++++++++
> 1 file changed, 14 insertions(+)
>
> diff --git a/drivers/gpu/drm/tidss/tidss_dispc.c
> b/drivers/gpu/drm/tidss/tidss_dispc.c
> index
> c0277fa36425ee1f966dccecf2b69a2d01794899..65ca7629a2e75437023bf58f8a1bddc24db5e3da
> 100644
> --- a/drivers/gpu/drm/tidss/tidss_dispc.c
> +++ b/drivers/gpu/drm/tidss/tidss_dispc.c
> @@ -498,6 +498,7 @@ struct dispc_device {
> const struct dispc_features *feat;
>
> struct clk *fclk;
> + struct regmap *clk_ctrl;
>
> bool is_enabled;
>
> @@ -1267,6 +1268,11 @@ void dispc_vp_enable(struct dispc_device *dispc, u32
> hw_videoport,
> FLD_VAL(mode->vdisplay - 1, 27, 16));
>
> VP_REG_FLD_MOD(dispc, hw_videoport, DISPC_VP_CONTROL, 1, 0, 0);
> +
> + if (dispc->clk_ctrl) {
> + regmap_update_bits(dispc->clk_ctrl, 0, 0x100, ipc ? 0x100 :
> 0x000);
> + regmap_update_bits(dispc->clk_ctrl, 0, 0x200, rf ? 0x200 :
> 0x000);
> + }
> }
>
> void dispc_vp_disable(struct dispc_device *dispc, u32 hw_videoport)
> @@ -3012,6 +3018,14 @@ int dispc_init(struct tidss_device *tidss)
>
> dispc_init_errata(dispc);
>
> + dispc->clk_ctrl =
> syscon_regmap_lookup_by_phandle_optional(tidss->dev->of_node,
> +
> "ti,clk-ctrl");
> + if (IS_ERR(dispc->clk_ctrl)) {
> + r = dev_err_probe(dispc->dev, PTR_ERR(dispc->clk_ctrl),
> + "DISPC: syscon_regmap_lookup_by_phandle
> failed.\n");
> + return r;
This breaks ABI. Commit msg mentions the reason but without
justification - was everything broken? Nothing was working? Was it ever
tested?
And anyway ABI impact must be clearly documented.
Best regards,
Krzysztof
