Hi,
On 22/04/2020 23:12, Martin Blumenstingl wrote:
Hi Neil,
On Tue, Apr 21, 2020 at 3:44 PM Neil Armstrong narmstrong@baylibre.com wrote: [...]
diff --git a/drivers/gpu/drm/meson/meson_dw_hdmi.c b/drivers/gpu/drm/meson/meson_dw_hdmi.c index e8c94915a4fc..dc3d5122475a 100644 --- a/drivers/gpu/drm/meson/meson_dw_hdmi.c +++ b/drivers/gpu/drm/meson/meson_dw_hdmi.c @@ -695,6 +695,13 @@ dw_hdmi_mode_valid(struct drm_connector *connector, dev_dbg(connector->dev->dev, "%s: vclk:%d phy=%d venc=%d hdmi=%d\n", __func__, phy_freq, vclk_freq, venc_freq, hdmi_freq);
/* Check against soc revision/package limits */if (priv->limits) {if (priv->limits->max_hdmi_phy_freq &&phy_freq > priv->limits->max_hdmi_phy_freq)return MODE_CLOCK_HIGH;}I think that this will also be useful for the 32-bit SoCs as well. is there a chance you can move it to meson_vclk_vic_supported_freq (called right below), where all the existing frequency limit checks are already?
It would need to add priv to meson_vclk_vic_supported_freq(), but indeed, would be cleaner.
And the meson_vclk_dmt_supported_freq() would also need this test aswell.
I'll resend with these fixed.
Neil
Thank you! Martin