[PATCH 1/2] drm: bridge/dw_hdmi: Cache edid data

[Russell King - ARM Linux]

On Fri, Aug 07, 2015 at 11:40:28PM +0100, Russell King - ARM Linux wrote:
> However, if we do this, then we might as well just modify
> dw_hdmi_connector_get_modes(), and arrange for a HPD de-assertion to
> NULL and free hdmi->edid so the next get_modes() call triggers a re-read.

Testing with the AV amp seems to be working okay.

For reference, he's the RXSENSE / HPD activity, with the AV amp
configured for HDMI passthrough in standby mode.  This output is from
the IRQ thread rather than the hardware IRQ function, so timings may not
be accurate.

unplug:
[ 9128.933266] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,---)
[ 9128.935883] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX----,---)

plugin:
[ 9167.072748] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,HPD)
[ 9167.121773] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,---)
[ 9167.122323] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX----,---)
[ 9168.320609] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,HPD)

av poweron:
[ 9185.696008] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,---)
[ 9185.696376] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX----,---)
[ 9187.286580] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,HPD)

tv poweron:
[ 9207.701588] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,---)
[ 9207.701953] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX----,---)
[ 9210.638573] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,HPD)

av standby:
[ 9230.588697] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,---)
[ 9230.589062] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX----,---)
[ 9231.788225] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX3210,HPD)

tv standby:
(nothing)

... and a bit later, when the AV amp switches into a lower standby mode:
[ 9290.700523] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX----,HPD)
[ 9290.944270] dwhdmi-imx 120000.hdmi: dw_hdmi_irq(RX----,---)

The interesting one is the plug-in event, where we see HPD dropped after
about 50ms.

I've previously had the opportunity to test the older code with another
manufacturer's AV amp, which I no longer have access to.  My notes for
that setup are as below - I didn't note down the kernel logs in this
case...

TV (no AV amp):
imxboot: initial(+rxsense,+hpd) ... -rxsense
standby->on: -hpd, +rxsense, +hpd, -hpd, -rxsense, (+rxsense +hpd)
on->unplug: -hpd, -rxsense
unplug->plug: (+rxsense +hpd)
plug->standby: no effect
standby->unplug: -hpd, -rxsense
unplug->plug: +rxsense, +hpd

AV (AV + TV in standby mode initially, AV connected to TV):
unplugged: -rxsense,-hpd
unplugged->plugged: +rxsense, +hpd
        EDID reads from AV
AV standby,TV standby->TV on: -hpd, 2s, +hpd
        EDID reads from TV
TV on,AV standby->AV on: -hpd, 1.2s, +hpd
	EDID reads from TV with AV capabilities merged
TV on,AV on->AV standby: -hpd, 1.2s, +hpd

I have a whole bunch of further patches for dw-hdmi, so if you don't
mind, I'll take your two, merge them on the front of my series, and
re-post them.  I think we're getting close to the time where we can
see initial audio support merged for dw-hdmi.

-- 
FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up
according to speedtest.net.