Repository navigation
arm64: dts: apple: t6050: Add shared USB-C display routes and four boards - #2283
RyanTheTide wants to merge 1 commit into
Conversation
…ards External displays over USB-C (DisplayPort alt mode) are described only on the J714s and the J716c. Every other T6050 board's ADT has the same four external display engines on die 0 (dcpext0..3, with their dart-dcpextN and dart-dispextN), the same display crossbar and the same ATC1/ATC2 Type-C PHYs, but their trees leave the engines bare and the connectors without a DisplayPort endpoint. Add t6050-display.dtsi with the board-agnostic part: - dcpext1..3: firmware index, USB-C output on die 0's DPTX, no clocks; their DCP and display DARTs' DMA range (vm-base 1 TiB, vm-size 3 TiB) and the display DARTs' single register window, as dcpext0's; - DisplayPort alt mode routes for dcpext0 and dcpext2 to ATC2 (route@0) and ATC1 (route@1), on crossbar source 0 and 4, with the DP IN controls, in the J714s's phy and crossbar order; - the ATC1 and ATC2 PHYs' DP clock source and DP power states, and both ports' display crossbars. The engines, their mailboxes and DARTs stay disabled: the boot loader hands each one over when its checks pass, as on the J714s. Each board adds only its connectors' DisplayPort endpoints. J714c, J716s, K114c and K116s include the file and link their left-front (ATC1) and right (ATC2) connectors. On the J714c the right connector's endpoint from t6050-max-atc2.dtsi is replaced. The K116s drops its own copy of the dcpext1..3 clock and DART range properties, now in the shared file. All values match each board's ADT. The J714s and J716c keep their own display files and their DTBs are byte-identical. J714c, J716s, K114c and K116s change; J714s, J716c, J775c, J775d, J873s, K114s and K116c are byte-identical. The routes, crossbar indices and PHY values were written by DjDeveloperr, Isaac Duarte and Erik Bourget for the J714s and J716c. Signed-off-by: Ryan Murray <ryan@aurorasilicon.org>
|
Looks good; approve once the K116s and K114c regression rows below are on the head. Checked on current integration/aurora: only J714c, J716s, K114c and K116s change, and the dispext DART single window, the crossbar/mux lists, route@0/route@1 and the ATC DP power states match J714s/J716c exactly. dt-checks and the ADT coverage check pass. The engines stay disabled, and
HW row: K116s and K114c on the head: USB 3 device and charging on the left-front and right ports (both orientations), 3x s2idle with lid wake, idle power within noise of the base (SMC PSTR, 5 min lid-closed-display-off average), internal panel unchanged; 0 AVC, 0 failed units, no new kernel errors. External display rows wait for the loader rows named in the body. Assisted-by: LLM (Claude:Opus 5.5) |
Problem
USB-C external displays (DisplayPort alt mode through the dcpext engines) are described only on the J714s and the J716c. Every other T6050 board's ADT has the same four external display engines on die 0 (dcpext0..3 with dart-dcpextN and dart-dispextN), the same display crossbar and the same ATC1/ATC2 Type-C PHYs. Their trees leave those engines bare, and their connectors have no DisplayPort endpoint.
Change
t6050-display.dtsiholds the part that is the same on every board:t6050-max-atc2.dtsiis replaced.t6050-j714s-display.dtsi(arm64: dts: apple: t6050-j714s: the left-back USB-C port (USB, USB4, DisplayPort) #2017, J714s: keep the USB4 dock port's USB host in place through a deep sleep #2072). They can include the shared file once those land.Testing
make dtbs CHECK_DTBS=y(dtschema) for every T6050 DTB on this branch: builds clean, no dtc warnings, and no schema message that integration/aurora does not already have.port0message). No new compatibles.Still to run
Each board needs this owner row: a USB-C DisplayPort monitor on the left-front port, then on the right port. On each port,
modetest -M apple -cshould list a connected DP connector with the monitor's modes, and a mode set (modetest -s) should light the monitor. Unplugging and replugging should report disconnected/connected (hotplug), and the monitor should come back.The kernel side is ready, but nothing hands these engines over yet. The loader hands external engines over only after an internal panel handoff with an image row that has external engines on:
Per-board values and DTB diffs
Values compared in each board's ADT (J714c, J716s, K114c, K116s, against J714s and J716c):
Changes per board, all of them in the merged tree:
Board drivers that limit, but do not refuse, the new boards in drivers/gpu/drm/apple:
dcp_j714s_ext: crash reconnect, deep-sleep refusal, link report)v27_j714s: A407/A500 method table)Assisted-by: LLM (Claude:Opus 5.5)