[ICL] TC port in legacy/static mode can't be detected due TCCOLD

When ICL SoC is in PC10, hotplugging HDMI adapter to Type-C DP, port can't be detected: "Port C/TC#1: PHY in TCCOLD, nothing connected"

The issue doesn't happen when SoC is not in PC10. dmesg

changed title from When SoC is in PC10, hotplugging Type-C to HDMI adapter to Type-C DP can't be detected: "Port C/TC#1: PHY in TCCOLD, nothing connected" to When ICL SoC is in PC10, hotplugging Type-C to HDMI adapter to Type-C DP can't be detected: "Port C/TC#1: PHY in TCCOLD, nothing connected"

changed the description

changed title from When ICL SoC is in PC10, hotplugging Type-C to HDMI adapter to Type-C DP can't be detected: "Port C/TC#1: PHY in TCCOLD, nothing connected" to When ICL SoC is in PC10, hotplugging Type-C to HDMI adapter to Type-C DP port can't be detected: "Port C/TC#1: PHY in TCCOLD, nothing connected"

added platform: ICL label

Add ICL TC Cold exit patch for OSV testing. https://patchwork.kernel.org/patch/11167917/

Yes this patch solves the issue.

changed title from When ICL SoC is in PC10, hotplugging Type-C to HDMI adapter to Type-C DP port can't be detected: "Port C/TC#1: PHY in TCCOLD, nothing connected" to [ICL] TC port in legacy/static mode can't be detected due TCCOLD

changed the description

I was wandering if a system with this TC legacy port would ever hit commercial products.

Apparently yes :)

Let me know if you want to be Cc in our private discussions.

added Community feature: display/USB-C labels

assigned to @zehortigoza

Just a headsup I'm working on this right now. Might have it tomorrow in the mailist, just need a litle bit more of testing.

Patches sent https://patchwork.freedesktop.org/series/74851/ can you give a try and send at least a Tested-by?

Thanks

Hi Jose, Per OSV tested, this patches got failure in validation. log is as attached. dmesg.5.6.0-rc6-tccold.log

Applied https://patchwork.freedesktop.org/series/74851/ to drm-tip commit 5e893da0b8c2, still cannot light up external screen during PC10 residency.

The diff bellow should fix the error/typo but it is still exiting the TCCOLD according to the logs. diff.txt

Can you share a log when running with https://patchwork.kernel.org/patch/11167917/ ?

dmesg.5.6.0-5-generic.patchwork-11167917 This succeeded to light up external for the first hotplug but failed all followings.

Thanks again.

Could you give another try with this branch and then share logs? https://github.com/zehortigoza/linux/tree/tccold

Sorry for keep asking tests but we don't any machine with TCCOLD working.

dmesg.5.6.0-rc7-zehortigoza-18292ed6e34a This works for me 10/10. Thank you.

Awesome thanks, just sent it to the mail list with the fixes https://patchwork.freedesktop.org/series/75034/ Can you give a Tested-by?

Done. Thank you. :)

Patches merged.

closed

Hi @vicamo, @cooperch,

could you test if everything works as expected with the latest drm-tip and on top of that the following patch: remove-legacy-timeout-exception.patch It would be great if you could also provide the drm.debug=0x1e dmesg logs.

Thanks!

Sure, I can have a test tomorrow.

• dmesg.5.7.0-rc2-drmtip-1f3ffd7683d5: drm-tip HEAD 1f3ffd7683d5
• dmesg.5.7.0-rc2-imredeak-1f3ffd7683d5: drm-tip HEAD 1f3ffd7683d5 with remove-legacy-timeout-exception.patch

Reboot => sudo powertop --auto-tune => watch cat /sys/kernel/debug/pmc_core/package_cstate_show => Lock screen => Plug HDMI cable with Type C dongle => Wake system

@vicamo , thanks looks ok, no warn on the legacy AUX power enabling.

mentioned in commit tzimmermann/linux@feb7e0ef