Skip to content

arm64: dts: apple: t6050-j714s: Add the left-front router's DP IN interrupts - #2002

Closed
Isaac-Duarte wants to merge 1 commit into
aurora-silicon:integration/aurorafrom
Isaac-Duarte:j714s/left-dock-dpin
Closed

Isaac-Duarte wants to merge 1 commit into
aurora-silicon:integration/aurorafrom
Isaac-Duarte:j714s/left-dock-dpin

Conversation

@Isaac-Duarte

@Isaac-Duarte Isaac-Duarte commented Nov 15, 2026 •

Copy link
Copy Markdown
Contributor

What

One commit, one file (arch/arm64/boot/dts/apple/t6050-j714s-display.dtsi), J714s only.

  • &acio1 (the left-front port's USB4 router, added in arm64: dts: apple: t6050-j714s: Add the left-front port's USB4 router #1791) gets its four DP IN interrupts, dpin0..dpin3: AIC 1635, 1636, 1641, 1642. They sit next to &acio2's, in the same form.
  • The file's header now says the left-front port carries USB4 DP tunnels too (it said the right port only), reflowed to 100 columns.

Why

#1791 described acio1 without the DP IN interrupts that acio2 has, so the left-front router doesn't match the right one or the T6050 inventory.

This is not a functional fix. apple_cio_probe() looks up dpinN with platform_get_irq_byname_optional(). With no line, a mapped DP IN is polled every 20 ms (APPLE_DPIN_POLL_MS). The IRQ thread and the poll both call apple_dpin_irq_dispatch(), and are started and stopped together (apple_dpin_irq_start()/_stop()). c->irq isn't used anywhere else. With the interrupts:

  • hot plug and IRQ_HPD events arrive without the up to 20 ms poll delay;
  • there is no idle poll per mapped DP IN.

Where the values come from:

  • Interrupt numbers:
    • atc1_dpin0..3 in t6050-peripherals.dtsi (the T6050 inventory).
    • J716c's acio1 has the same four (t6050-j716c-left-usb4.dtsi).
    • They follow acio2's, 32 lower, like the rest of the router's interrupts. acio2 has NHI rings 1643-1666, DART 1670, parity 1672, DP IN 1667/1668/1673/1674. acio1 has rings 1611-1634, DART 1638, parity 1640, DP IN 1635/1636/1641/1642.
    • I had no J714sAP ADT dump to check them against.
  • Addresses: those inventory DP IN blocks (0x2a9e50000, +0x8000, +0x18000, +0x20000) are where the driver maps DP IN 0..3 from acio1's rc region (APPLE_DPIN_OFFSET 0x390000 + apple_dpin_offsets[]).
  • The display side is already in place: dcpext0 and dcpext2 have typec-routes to typec1 with the typec1-dpin0..3 controls on atc1_dpxbar. The router finds that connector through its port@1 endpoint (typec1 port@3).

There is no DT binding for apple,t6050-usb4-acio in the tree, so no schema covers these properties. acio2 uses the same names.

Separately, there's an earlier report that a dock's displays stay dark on the left-front port. It's unverified, with no logs, and given the polling above this change isn't expected to fix it.

Tested

Not run on hardware. Only build checks:

  • DTB: t6050-j714s.dtb built at the base (fef1b4c8d6) and at this head with cpp + dtc 1.8.1, using kbuild's include paths. Both builds have no dtc warnings. dtc -I dtb -O dts shows that the only difference is in acio@2a9ac0000: interrupts = <0 0 0x663 4>, <0 0 0x664 4>, <0 0 0x669 4>, <0 0 0x66a 4> and the four interrupt-names.
    make wasn't usable here: my worktree is a sparse checkout, without scripts/ or the Kconfig tree.
  • dt-schema: not installed, so CHECK_DTBS was not run. No binding covers the acio node anyway.
  • checkpatch --strict: 0 errors, 0 warnings, 0 checks.

To run before merge

  • J714s, a Thunderbolt dock on the left-front port: the 2a9ac0000.acio-dpinN lines in /proc/interrupts count up on hot plug, and displays behave as before (or better). If displays stay dark, these driver messages say why: no Type-C connector, no display pipeline free, waiting for display driver/firmware readiness, DP tunnel setup failed.
  • J714s: the same dock on the right port, to check nothing changed there.
  • J714s: a DP alt mode display on the left-front port still lights.

…errupts

The left-front port's USB4 router (acio1) came with the port's USB4
support, but without the DP IN interrupts that the right port's router
(acio2) has. With no "dpinN" interrupt, the router driver polls a mapped
DP IN's status every 20 ms instead; the interrupt and the poll run the
same handler. This is not a functional fix: with the interrupts, hot plug
and IRQ_HPD events arrive without the up to 20 ms poll delay, and there
is no idle poll per mapped DP IN.

Give acio1 the four DP IN interrupts, as acio2 has. The numbers are those
of atc1-dpin0..3 in the T6050 inventory and of J716c's acio1, and follow
acio2's, 32 lower like the rest of the router's interrupts. The DP IN
blocks there sit at the offsets the driver maps from the router's "rc"
region. The display side is already in place: dcpext0 and dcpext2 have
routes to this port with its DP IN crossbar controls (typec1-dpin0..3 on
atc1_dpxbar). Say in the file's header that the left-front port carries
DP tunnels too.

Signed-off-by: Isaac Duarte <31712685+Isaac-Duarte@users.noreply.github.com>
@Isaac-Duarte

Copy link
Copy Markdown
Contributor Author

Folded into #2006 (commit 0e0169b) to avoid the header-comment conflict with #1921/#2006; the change is unchanged.

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.

1 participant