Re: [PATCH 01/33] drm/omap: HDMI: change enable/disable to avoid sync-losts

[Tomi Valkeinen <tomi.valkeinen@ti.com>]

[-- Attachment #1.1: Type: text/plain, Size: 3111 bytes --]



On 23/02/16 12:22, Laurent Pinchart wrote:
> Hi Tomi,
> 
> Thank you for the patch.
> 
> On Friday 19 February 2016 11:47:36 Tomi Valkeinen wrote:
>> We occasionally see DISPC sync-lost errors when enabling and disabling
>> HDMI. Sometimes we get only a few, which get handled (ignored) by the
>> driver, but sometimes there's a flood of the errors which doesn't seem
>> to stop.
>>
>> The HW team has root caused this to the order in which HDMI and DISPC
>> are enabled/disabled. Currently we enable HDMI first, and then DISPC,
>> and vice versa when disabling. HW team's suggestion is to do it the
>> other way around.
>>
>> This patch changes the order, but this has two side effects as the pixel
>> clock is produced by HDMI, and the clock is not running when we
>> enable/disable DISPC:
>>
>> * When enabling DISPC first, we don't get vertical sync events
>> * When disabling DISPC last, we don't get FRAMEDONE event
>>
>> At the moment we use both of those to verify that DISPC has been
>> enabled/disabled properly. Thus this patch also needs to change the
>> omapdrm and omapdss which handle the DISPC side.

<snip>

>> +	if (omap_crtc->mgr->output->output_type == OMAP_DISPLAY_TYPE_HDMI) {
>> +		dispc_mgr_enable(channel, enable);
>> +		return;
>> +	}
> 
> This effectively bypasses the wait until the DISPC outputs the first vsync 
> interrupt below. How does HDMI differ from other outputs in such a way to make 
> the wait unneeded ?

There's to parts here. Enabling the output and disabling the output.

Enable:

We don't strictly need the wait after enable for any output. The output
works after setting the enable bit.

There are two reasons for the waiting:

1) A sanity check that the configuration is ok. If the config is broken
(which shouldn't happen, of course, as the driver should verify the
config), we won't see vsync. At the moment we only print an error in
that case.

2) OMAP_DSS_CHANNEL_DIGIT is a bit problematic. That channel is used for
analog tv-out (VENC) in older DSS versions, and for HDMI for more recent
ones. With VENC we always get a few sync lost errors when enabling the
output, so with the wait we can ignore those errors (this sync-lost
ignoring is only done for OMAP_DSS_CHANNEL_DIGIT).

We have seen similar sync losts with HDMI too, but apparently it is
possible to support HDMI without any sync losts. That's what this patch
is doing.

With this patch we lose both 1) and 2) above, but 1) is not strictly
needed and 2) shouldn't happen for HDMI after this patch.

We could implement 1) in the HDMI driver too, using the HDMI IP's VSYNC
interrupt, but I don't think it's really necessary.

Disable:

When disabling the output, we do want to wait until the DSS has finished
the work at the end of the frame. This is done in the
omap_crtc_set_enabled() function for all outputs, using FRAMEDONE
interrupt when available, or VSYNC if not.

For HDMI we can do it also in the HDMI driver. The HDMI IP has its own
FRAMEDONE interrupt, which we wait for in hdmi_wp_video_stop().

 Tomi


[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

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