On Wed, Aug 28, 2013 at 03:34:53PM +0300, Mikko Perttunen wrote:
On 08/28/2013 03:07 PM, Thierry Reding wrote:
- PGP Signed by an unknown key
On Wed, Aug 28, 2013 at 01:40:56PM +0300, Mikko Perttunen wrote:
Use EDID data to determine whether the display supports HDMI or just DVI. This used to be hardcoded to be HDMI, which broke support for DVI displays that couldn't understand the interspersed audio/other data.
If the EDID data isn't available, default to DVI, which should be a safer choice.
Signed-off-by: Mikko Perttunen mperttunen@nvidia.com
drivers/gpu/host1x/drm/hdmi.c | 8 ++++++++ 1 file changed, 8 insertions(+)
diff --git a/drivers/gpu/host1x/drm/hdmi.c b/drivers/gpu/host1x/drm/hdmi.c index d81fac8..140339b 100644 --- a/drivers/gpu/host1x/drm/hdmi.c +++ b/drivers/gpu/host1x/drm/hdmi.c @@ -702,6 +702,14 @@ static int tegra_output_hdmi_enable(struct tegra_output *output) unsigned long value; int retries = 1000; int err;
struct drm_property_blob *edid_blob = output->connector.edid_blob_ptr;
if (edid_blob && edid_blob->data &&
drm_detect_hdmi_monitor((struct edid *)edid_blob->data)) {hdmi->dvi = false;} else {
hdmi->dvi = true;}
pclk = mode->clock * 1000; h_sync_width = mode->hsync_end - mode->hsync_start;
Odd, now that I see that code I remember that there was a similar patch a few months back, but it was never applied for some reason:
http://lists.freedesktop.org/archives/dri-devel/2013-January/033509.html
That was already reviewed by me and Jon Mayo, so I'll go ahead and apply that one instead.
Thierry
- Unknown Key
- 0x7F3EB3A1
That patch seems to cause a warning for me: drivers/gpu/host1x/drm/hdmi.c: In function ‘tegra_output_hdmi_enable’: drivers/gpu/host1x/drm/hdmi.c:706:2: warning: passing argument 1 of ‘drm_detect_hdmi_monitor’ discards ‘const’ qualifier from pointer target type [enabled by default] include/drm/drm_crtc.h:1037:13: note: expected ‘struct edid *’ but argument is of type ‘const struct edid *’
Given the discussion about this back in January, I think the best way, at least for now, is to keep a copy of the EDID data so that it can actually be modified.
But looking at the problem more closely, I don't think the patch from January is quite correct. The problem is that it uses the EDID stored within the struct tegra_output. That's problematic because that field is not updated when EDID is probed via DDC. That would make the patch that you proposed more correct.
Thierry