Skip to content

Conversation

@miczyg1
Copy link
Contributor

@miczyg1 miczyg1 commented May 7, 2025

No description provided.

@miczyg1 miczyg1 marked this pull request as draft May 7, 2025 15:54
@miczyg1 miczyg1 requested a review from mkopec May 8, 2025 08:02
@miczyg1 miczyg1 marked this pull request as ready for review May 8, 2025 08:02
@miczyg1
Copy link
Contributor Author

miczyg1 commented May 8, 2025

Fixes small misconfiguration, but does not fix the problem with disappearing devices in the dock in Windows

@miczyg1 miczyg1 changed the title WIP: Mtl dock fixes MTL TCSS fixes May 8, 2025
@miczyg1 miczyg1 force-pushed the mtl_dock_fixes branch 4 times, most recently from 2f3e898 to 58209fb Compare May 15, 2025 13:59
Copy link
Member

@mkopec mkopec left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Tested on NovaCustom V540TU with Wavlink UMD05 Pro Rev. E, seems to fix issue Dasharo/dasharo-issues#1316

miczyg1 added 15 commits May 15, 2025 17:22
The EnableTcssCovTypeA UPD definitions changed compared to the
ADL/RPL FSP and the PCH XHCI port being mapped is not in a
separate UPD anymore. Instead, it is passed in the same field
as (previously boolean UINT8) EnableTcssCovTypeA.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
All TCSS XHCI ports are connected to a retimer (either TBT or USB3).
These retimers already handle SBU pins and the orientation. Clear
retimer bits:

- for port 3 and 4 as they are not used
- for port 2 as it has USB3 retimer with SBU orientation support

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
The default state of TBT force power GPIO was not consistent across
variants. Assert TBT force power by default before FSP so that the
initialization will be done when the retimer is active.

De-assert TBT force power before handing off to the payload, so that
the RTD3 logic will work in OS.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Match upstream version. The non-TBT USB2.0 port is routed the same
way as for TBT port.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Mimic Intel's reference code.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
TCSS XHCI ports that have a retimer connected should disable U1 and U2
transitions, as per Intel's reference code.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Mimic Intel's reference code.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Document #741494 says it is required for TCSS.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Report IRQ rotues as in reference code.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
There was a bug which assignd IRQ for device 2.2 instead of 2.0
by using wrong macro.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Document #741494 says to set these options.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
If DP tulleing and DP_ALT monitoring is enabeld in FSP for IOM,
ACPI can notify IGD about HPD assertion and wake IGD from D3 hot.

Add ACPI code to detect the HPD assertions and notify IGD if needed.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Enable TCSS to notify the IGD about hot plug events and eventually
wake it up from D3 hot state.

Disabel PS_ON feature, as it is only available on desktops. In
Intel reference code it was involved in inhibiting the power gating
of certain IP blocks.

Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
@mkopec mkopec merged commit 3c6d02e into dasharo May 15, 2025
2 of 3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants