Hi Russell,

Thanks for your comments :-D

On 01/19/2016 01:15 AM, Russell King - ARM Linux wrote:

Hi,

Some comments below...

On Mon, Jan 18, 2016 at 10:42:20PM +0800, Yakir Yang wrote:

quoted

+static int inno_hdmi_config_video_avi(struct inno_hdmi *hdmi)
+{
+	char info[HDMI_SIZE_AVI_INFOFRAME] = {0};
+	int avi_color_mode;
+	int i;
+
+	hdmi_writeb(hdmi, HDMI_CONTROL_PACKET_BUF_INDEX, INFOFRAME_AVI);
+
+	info[0] = 0x82;
+	info[1] = 0x02;
+	info[2] = 0x0D;
+	info[3] = info[0] + info[1] + info[2];
+
+	if (hdmi->hdmi_data.enc_out_format == HDMI_COLORSPACE_RGB)
+		avi_color_mode = AVI_COLOR_MODE_RGB;
+	else if (hdmi->hdmi_data.enc_out_format == HDMI_COLORSPACE_YUV444)
+		avi_color_mode = AVI_COLOR_MODE_YCBCR444;
+	else if (hdmi->hdmi_data.enc_out_format == HDMI_COLORSPACE_YUV422)
+		avi_color_mode = AVI_COLOR_MODE_YCBCR422;
+	else
+		avi_color_mode = AVI_COLOR_MODE_RGB;
+
+	info[4] = (avi_color_mode << 5);
+	info[5] = (AVI_COLORIMETRY_NO_DATA << 6) |
+		  (AVI_CODED_FRAME_ASPECT_NO_DATA << 4) |
+		  ACTIVE_ASPECT_RATE_SAME_AS_CODED_FRAME;
+
+	info[6] = 0;
+	info[7] = hdmi->hdmi_data.vic;
+
+	if (hdmi->hdmi_data.vic == 6 || hdmi->hdmi_data.vic == 7 ||
+	    hdmi->hdmi_data.vic == 21 || hdmi->hdmi_data.vic == 22)
+		info[8] = 1;
+	else
+		info[8] = 0;
+
+	/* Calculate avi info frame checKsum */
+	for (i = 4; i < HDMI_SIZE_AVI_INFOFRAME; i++)
+		info[3] += info[i];
+	info[3] = 0x100 - info[3];
+
+	for (i = 0; i < HDMI_SIZE_AVI_INFOFRAME; i++)
+		hdmi_writeb(hdmi, HDMI_CONTROL_PACKET_ADDR + i, info[i]);
+
+	return 0;

Is there a reason why the helpers in drivers/video/hdmi.c can't be used
to generate this?

hdmi_avi_infoframe_init / hdmi_avi_infoframe_pack and
drm_hdmi_avi_infoframe_from_display_mode ?

Great, thanks for point out, it would make code much clean with those 
helper functions.

quoted

+}
+
+static int inno_hdmi_config_video_vsi(struct inno_hdmi *hdmi)
+{
+	char info[HDMI_SIZE_VSI_INFOFRAME] = {0};
+	int i;
+
+	hdmi_modb(hdmi, HDMI_PACKET_SEND_AUTO, m_PACKET_VSI_EN,
+		  v_PACKET_VSI_EN(0));
+
+	hdmi_writeb(hdmi, HDMI_CONTROL_PACKET_BUF_INDEX, INFOFRAME_VSI);
+
+	/* Header Bytes */
+	info[0] = 0x81;
+	info[1] = 0x01;
+
+	/* PB1 - PB3 contain the 24bit IEEE Registration Identifier */
+	info[4] = 0x03;
+	info[5] = 0x0c;
+	info[6] = 0x00;
+
+	/* PB4 - HDMI_Video_Format into bits 7:5 */
+	info[7] = 0;
+
+	/*
+	 * PB5 - Depending on the video format, this byte will contain
+	 * either the HDMI_VIC code in buts 7:0, OR the 3D_Structure in
+	 * bits 7:4.
+	 */
+	info[2] = 0x06 - 2;
+	info[8] = 0;
+	info[9] = 0;
+
+	info[3] = info[0] + info[1] + info[2];
+
+	/* Calculate info frame checKsum */
+	for (i = 4; i < HDMI_SIZE_VSI_INFOFRAME; i++)
+		info[3] += info[i];
+	info[3] = 0x100 - info[3];
+
+	for (i = 0; i < HDMI_SIZE_VSI_INFOFRAME; i++)
+		hdmi_writeb(hdmi, HDMI_CONTROL_PACKET_ADDR + i, info[i]);
+
+	hdmi_modb(hdmi, HDMI_PACKET_SEND_AUTO, m_PACKET_VSI_EN,
+		  v_PACKET_VSI_EN(1));
+
+	return 0;

hdmi_vendor_infoframe_init / hdmi_vendor_infoframe_pack?

You can probably use the same function to upload the control packet
too - something like:

static int inno_hdmi_upload_frame(struct inno_hdmi *hdmi, int setup_rc,
	union hdmi_infoframe *frame, u32 mask, u32 disable, u32 enable)
{
	if (mask)
		hdmi_modb(hdmi, HDMI_PACKET_SEND_AUTO, mask, disable);

	if (setup_rc >= 0) {
		u8 packed_frame[YOUR_MAXIMUM_INFO_FRAME_SIZE];
		ssize_t rc, i;

		rc = hdmi_infoframe_pack(frame, packed_frame,
					 sizeof(packed_frame));
		if (rc < 0)
			return rc;

		for (i = 0; i < rc; i++)
			hdmi_writeb(hdmi, HDMI_CONTROL_PACKET_ADDR + i, buf[i]);

		if (mask)
			hdmi_modb(hdmi, HDMI_PACKET_SEND_AUTO, mask, enable);
	}

	return setup_rc;
}

static int inno_hdmi_config_video_vsi(struct inno_hdmi *hdmi)
{
	union hdmi_infoframe frame;

	rc = drm_hdmi_vendor_infoframe_from_display_mode(&frame.vendor.hdmi,
							 the_drm_display_mode);

	return inno_hdmi_upload_frame(hdmi, rc, &frame, m_PACKET_VSI_EN,
				      v_PACKET_VSI_EN(0), v_PACKET_VSI_EN(1));
}

Ah.... appreciate for the code,  it's great suggest, thanks.

Also I found a mistaken with previous "inno_hdmi_config_video_vsi"
function that we don't need to send the HDMI vendor infoframes if
we are not using a 4K or stereoscopic 3D mode.  :)

quoted

+static int inno_hdmi_i2c_wait(struct inno_hdmi *hdmi)
+{
+	struct inno_hdmi_i2c *i2c = hdmi->i2c;
+	int stat;
+
+	stat = wait_for_completion_timeout(&i2c->cmp, HZ / 10);
+	if (!stat) {
+		stat = wait_for_completion_timeout(&i2c->cmp, HZ / 10);
+		if (!stat)
+			return -EAGAIN;

What's the reason for this double wait_for_completion?  This probably
needs a comment, otherwise it looks like some weird cut-n-paste error.

Yep, you 're right, don't need double wait here.

quoted

+static int inno_hdmi_i2c_write(struct inno_hdmi *hdmi, struct i2c_msg *msgs)
+{
+	struct inno_hdmi_i2c *i2c = hdmi->i2c;

You seem to have a local i2c pointer...

Ah, got it  ;)

quoted

+
+	/*
+	 * The DDC module only support read EDID message, so
+	 * we assume that each word write to this i2c adapter
+	 * should be the offset of EDID word address.
+	 */
+	if ((msgs->len != 1) ||
+	    ((msgs->addr != DDC_ADDR) && (msgs->addr != DDC_SEGMENT_ADDR)))
+		return -EINVAL;
+
+	reinit_completion(&i2c->cmp);
+
+	if (msgs->addr == DDC_SEGMENT_ADDR)
+		hdmi->i2c->segment_addr = msgs->buf[0];
+	if (msgs->addr == DDC_ADDR)
+		hdmi->i2c->ddc_addr = msgs->buf[0];
+
+	/* Set edid fifo first addr */
+	hdmi_writeb(hdmi, HDMI_EDID_FIFO_OFFSET, 0x00);
+
+	/* Set edid word address 0x00/0x80 */
+	hdmi_writeb(hdmi, HDMI_EDID_WORD_ADDR, hdmi->i2c->ddc_addr);
+
+	/* Set edid segment pointer */
+	hdmi_writeb(hdmi, HDMI_EDID_SEGMENT_POINTER, hdmi->i2c->segment_addr);

but then you don't use it in the four locations above.

Thanks,
- Yakir


_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel