[26.04_linux-nvidia-bos] i2c: mediatek: add ACPI/MT8901 support and gpiolib zero-debounce fix - #518
Conversation
PR Validation ReportPatchscan ✅ No Missing FixesAll cherry-picked commits checked — no missing upstream fixes found. PR Lint ✅ All checks passedDetailsChecking 2 commits... Cherry-pick digest: ┌──────────────┬──────────────────────────────────────────────────────────────────┬────────────┬─────────┬───────────────────────────┐ │ Local │ Referenced upstream / Patch subject │ Patch-ID │ Subject │ SoB chain │ ├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤ │ be4ca719d290 │ [SAUCE] gpiolib: acpi: route acpi_dev_gpio_irq_wake_get_by() deb │ N/A │ N/A │ kmaddara │ ├──────────────┼──────────────────────────────────────────────────────────────────┼────────────┼─────────┼───────────────────────────┤ │ dfc63029af72 │ [SAUCE] i2c: mediatek: add acpi/mt8901 support and firmware-mana │ N/A │ N/A │ zhang, kmaddara │ └──────────────┴──────────────────────────────────────────────────────────────────┴────────────┴─────────┴───────────────────────────┘ Lint: all checks passed. |
BaseOS Kernel ReviewSummaryNo issues found across the reviewed commits. Findings: no problems found Latest watcher review: open review Generated test plan: open test plan Kernel deb build: successful (download debs, 4 files) Head: This comment is maintained by nv-pr-bot. It is updated when the GitHub watcher publishes a newer review. |
Additionally, my review with Codex had one finding: Medium: drivers/i2c/busses/i2c-mt65xx.c:1390 now defaults missing clock-div to 1 for both ACPI and DT. The comment says this is for ACPI, but DT still uses this same parser, and the binding requires clock-div at Documentation/devicetree/bindings/i2c/i2c-mt65xx.yaml:110. Previously a missing or bad DT clock-div failed probe; now it silently probes with different timing. I’d gate the default on has_acpi_companion(i2c->dev) and keep returning the property-read error for DT. |
|
@nvmochs
What is the target platform for this? Spark?
Related, is there a NVbug associated with these patches?
This same PR needs to also be submitted against the 7.0-lts kernel (26.04_linux-nvidia). |
@kmaddaraki Thanks for clarifying. Please address the codex/claude findings listed here and in PR 507 and then we should be able to get this merged. |
5877724 to
cd36a16
Compare
|
The Codex finding is fixed in v2 — the clock-div default now applies only when an ACPI companion is present; the DT path returns the property-read error as before, so DT behavior is unchanged. |
Thanks! I confirmed this has been fixed along with the other findings from PR 507.
|
|
Same question I left in #507: Does the |
…naged clocks On MT8901-based platforms the I2C controllers are described through ACPI (HID NVDA0200) rather than device tree. The upstream i2c-mt65xx driver is DT-only: no acpi_match_table, no propagation of the ACPI fwnode to the i2c_adapter device, and clock/timing properties read through of_property_read_*() helpers that don't operate on ACPI nodes. Compounding that, on these platforms firmware keeps the controller's "main" and "dma" clocks running and does not expose them via the Linux clk framework, so devm_clk_get() returns -ENOENT for both. Consequently the driver does not bind on ACPI systems: no i2c_adapter is created, HID-over-I2C never enumerates child devices, and the internal I2C keyboard stays silent. Wire up ACPI and accept firmware-managed clocks: * Add mt8901_compat (v3 register layout, default_parent_rate=124.8 MHz as a stand-in for clk_get_rate() when no clock provider is exposed) and an acpi_match_table entry NVDA0200 -> mt8901_compat. * Replace of_device_get_match_data() with device_get_match_data() in probe (with a NULL-match check), and switch mtk_i2c_parse_dt() to the fwnode-aware device_property_read_*() helpers so DT and ACPI share the same probe path. * Call device_set_node(&adap->dev, dev_fwnode(&pdev->dev)) so the ACPI fwnode reaches the i2c_adapter device; without it, has_acpi_companion() on the adapter returns 0 and i2c_acpi_register_devices() exits early, leaving HID children unenumerated. * Gate devm_clk_get() for "main"/"dma" on has_acpi_companion() - use devm_clk_get_optional() in the ACPI branch so absent clocks are accepted, and substitute i2c->dev_comp->default_parent_rate when mtk_i2c_set_speed() has no clock handle to query. Signed-off-by: Housong Zhang <housong.zhang@mediatek.com> Signed-off-by: Kiran Maddaraki <kmaddaraki@nvidia.com>
…ebounce through the warn-only wrapper Mainline commit e4a77f9 ("gpiolib: acpi: Make set debounce errors non fatal") introduced acpi_gpio_set_debounce_timeout() and converted two of the three ACPI call sites that program a debounce timeout to use it. The wrapper downgrades a set_config(PIN_CONFIG_INPUT_DEBOUNCE) failure to a dev_warn() so that GPIO controllers which reject that config (e.g. the MediaTek EINT block used behind pinctrl-paris, which returns -EINVAL for any debounce request on an edge-triggered pin) do not fail the whole GpioInt lookup. The third call site in acpi_dev_gpio_irq_wake_get_by() was not converted by that commit and still returns the raw -EINVAL from gpio_set_debounce_timeout(). On ACPI systems whose _CRS declares a GpioInt with any DebounceTimeout on an edge-triggered pinctrl backend, that error propagates through acpi_dev_gpio_irq_get() and its callers, and the affected consumer (e.g. i2c_hid_acpi) fails to bind - no HID children enumerate. Route this call site through the same warn-only wrapper so all three ACPI debounce paths behave consistently. Fixes: 8dcb7a1 ("gpiolib: acpi: Take into account debounce settings") Signed-off-by: Kiran Maddaraki <kmaddaraki@nvidia.com>
cd36a16 to
be4ca71
Compare
|
Yes, agreed — added Fixes: 8dcb7a1 ("gpiolib: acpi: Take into account debounce settings") |
|
|
nirmoy
added
has_2_acks
and removed
help wanted
|
|
6 participants
Two small SAUCE patches to bring up the MediaTek MT8901 I²C host controller under ACPI on 26.04 linux-nvidia-bos.
gpiolib: acpi: route acpi_dev_gpio_irq_wake_get_by() debounce through the warn-only wrapper
Mainline commit e4a77f9 ("gpiolib: acpi: Make set debounce errors non fatal") added the warn-only acpi_gpio_set_debounce_timeout() wrapper and converted two of the three ACPI call sites that program a debounce timeout. The third, in acpi_dev_gpio_irq_wake_get_by(), still returns the raw error: on firmware whose _CRS declares a GpioInt with a DebounceTimeout on an edge-triggered pin, the MediaTek EINT driver rejects set_config(PIN_CONFIG_INPUT_DEBOUNCE) with -EINVAL, the error propagates through acpi_dev_gpio_irq_get(), and i2c_hid_acpi fails probe with HID over i2c has not been provided an Int IRQ — the internal keyboard never enumerates. Convert the remaining call site to the same wrapper so all three ACPI debounce paths behave consistently: the rejection is logged and the IRQ is still delivered. Core gpiolib.c semantics are untouched.
i2c: mediatek: add ACPI/MT8901 support and firmware-managed clocks
Adds ACPI binding to the DT-only i2c-mt65xx driver so the MT8901 I²C controllers enumerate under NVDA0200:
New mt8901_compat (v3 register layout, default_parent_rate=124.8 MHz) and an acpi_match_table entry NVDA0200 → mt8901_compat.
Switch probe to device_get_match_data() and fwnode-aware device_property_read_*() so DT and ACPI share the same code path.
device_set_node(&adap->dev, dev_fwnode(&pdev->dev)) so i2c_acpi_register_devices() walks the child HID nodes and instantiates them.
Gate devm_clk_get() for main/dma on has_acpi_companion() — use devm_clk_get_optional() for the ACPI path (firmware keeps those clocks running and does not expose them via the Linux clk framework), and fall back to dev_comp->default_parent_rate in mtk_i2c_set_speed() when no clock handle is available.
DT behaviour is unchanged: both device_get_match_data() and device_property_read_*() fall through to the existing OF helpers for DT-described nodes, and the DT clock path is preserved.
LP: https://bugs.launchpad.net/ubuntu/+source/linux-nvidia-7.0/+bug/2163407