[PATCH v3 12/15] drm/bridge: Add ITE IT6263 LVDS to HDMI converter
Liu Ying
victor.liu at nxp.com
Tue Oct 22 08:10:51 UTC 2024
Hi Maxime,
On 10/22/2024, Maxime Ripard wrote:
> On Tue, Oct 22, 2024 at 03:36:47PM +0800, Liu Ying wrote:
>> Hi Maxime,
>>
>> On 10/21/2024, Maxime Ripard wrote:
>>> On Mon, Oct 21, 2024 at 02:44:43PM +0800, Liu Ying wrote:
>>>> +static int it6263_bridge_atomic_check(struct drm_bridge *bridge,
>>>> + struct drm_bridge_state *bridge_state,
>>>> + struct drm_crtc_state *crtc_state,
>>>> + struct drm_connector_state *conn_state)
>>>> +{
>>>> + struct drm_display_mode *mode = &crtc_state->adjusted_mode;
>>>> + int ret;
>>>> +
>>>> + ret = drm_atomic_helper_connector_hdmi_check(conn_state->connector,
>>>> + conn_state->state);
>>>> + if (ret)
>>>> + return ret;
>>>> +
>>>> + return mode->clock > MAX_PIXEL_CLOCK_KHZ ? -EINVAL : 0;
>>>
>>> drm_atomic_helper_connector_hdmi_check will already make that check, so
>>> it's redundant.
>>
>> MAX_PIXEL_CLOCK_KHZ is 150MHz. With 150MHz pixel clock rate, we'll get
>> 150MHz HDMI character rate for 8bpc and 187.5MHz HDMI character rate
>> for 10bpc, both are lower than MAX_HDMI_TMDS_CHAR_RATE_HZ = 225MHz.
>
> I guess? I have no idea how that's relevant though. Where are those
> constraints coming from, and why aren't you checking for them in
> tmds_char_rate_valid?
All constraints come from IT6263 data sheet. They are also mentioned
in IT6263 product link(commit message contains the link).
https://www.ite.com.tw/en/product/cate1/IT6263
"
LVDS RX
Support input clock rate up to 150 MHz
HDMI TX
Support link speeds of up to 2.25 Gbps (link clock rate of 225 MHz)
"
If no objection, I'll check mode clock rate against
MAX_PIXEL_CLOCK_KHZ in tmds_char_rate_valid.
>
>> So, it looks like pixel clock rate is the bottleneck.
>
> The bottleneck to what?
To the IT6263 video processing throughput capability.
>
>> Remove drm_atomic_helper_connector_hdmi_check() or keep this as-is?
>
> No, like I said, remove the final check for mode->clock.
>
> Maxime
--
Regards,
Liu Ying
More information about the dri-devel
mailing list