[Deepin-Kernel-SIG] [linux 6.6.y] [Upstream] Sync bugfix for intel mei from upstream

[opsiff]

Arnd Bergmann (1):
mei: fix vsc dependency

Hans de Goede (8):
mei: vsc: Use "wakeuphostint" when getting the host wakeup GPIO
mei: vsc: Do not re-enable interrupt from vsc_tp_reset()
mei: vsc: Don't re-init VSC from mei_vsc_hw_reset() on stop
mei: vsc: Destroy mutex after freeing the IRQ
mei: vsc: Event notifier fixes
mei: vsc: Unset the event callback on remove and probe errors
mei: vsc: Run event callback from a workqueue
mei: vsc: Fix "BUG: Invalid wait context" lockdep error

drivers/misc/mei/Kconfig | 1 +
drivers/misc/mei/platform-vsc.c | 8 ++++++-
drivers/misc/mei/vsc-tp.c | 41 ++++++++++++++++++++-------------
3 files changed, 33 insertions(+), 17 deletions(-)

Summary by Sourcery

Synchronize upstream Intel MEI VSC bug fixes, improving interrupt handling, event notification, and reset behavior for the VSC transport, and aligning GPIO usage with the upstream driver.

Bug Fixes:

• Fix VSC interrupt handling by moving event notification to a workqueue and avoiding invalid wait-context usage in the IRQ thread.
• Ensure VSC reset and init logic do not re-enable or reinitialize interrupts incorrectly based on the intr_enable flag.
• Correct the VSC wakeup GPIO name to use the upstream "wakeuphostint" line for host wake events.
• Prevent use-after-free and ordering issues by canceling work and unregistering event callbacks on probe failure and driver remove, and destroying mutexes after IRQ teardown.

[sourcery-ai]

Reviewer's Guide

Syncs multiple upstream fixes for the Intel MEI VSC transport: adjusts IRQ handling and event notifications to use a workqueue with proper locking and teardown, fixes GPIO naming and reset/interrupt behavior, and ensures event callbacks are cleared on probe errors and device removal to avoid races and lockdep issues.

Sequence diagram for IRQ-driven VSC event notification via workqueue

sequenceDiagram
    participant SPI as SpiController
    participant IRQ as IRQSubsystem
    participant TP as VscTpDriver
    participant WQ as SystemWorkqueue
    participant MEI as MeiVscDriver

    SPI->>IRQ: Assert wakeuphostint (falling edge)
    IRQ->>TP: vsc_tp_isr(irq, tp)
    activate TP
    TP->>TP: wake_up(xfer_wait)
    TP->>WQ: schedule_work(event_work)
    deactivate TP

    WQ->>TP: vsc_tp_event_work(event_work)
    activate TP
    TP->>TP: lock(event_notify_mutex)
    alt event_notify is set
        TP->>MEI: event_notify(event_notify_context)
    else event_notify is NULL
        TP->>TP: skip callback
    end
    TP->>TP: unlock(event_notify_mutex)
    deactivate TP

Loading

Updated class diagram for VSC transport and MEI VSC driver structures

classDiagram
    class vsc_tp {
        atomic_t assert_cnt
        wait_queue_head_t xfer_wait
        work_struct event_work
        vsc_tp_event_cb_t event_notify
        void* event_notify_context
        mutex event_notify_mutex
        mutex mutex
        void vsc_tp_reset()
        int vsc_tp_register_event_cb(vsc_tp* tp, vsc_tp_event_cb_t event_cb, void* context)
    }

    class VscTpDriver {
        irqreturn_t vsc_tp_isr(int irq, void* data)
        void vsc_tp_event_work(work_struct* work)
        int vsc_tp_probe(spi_device* spi)
        void vsc_tp_remove(spi_device* spi)
    }

    class mei_vsc_hw {
        vsc_tp* tp
    }

    class mei_device {
        void* dev
    }

    class MeiVscDriver {
        int mei_vsc_hw_reset(mei_device* mei_dev, bool intr_enable)
        int mei_vsc_probe(platform_device* pdev)
        void mei_vsc_remove(platform_device* pdev)
    }

    vsc_tp <|-- VscTpDriver
    mei_vsc_hw o-- vsc_tp
    MeiVscDriver o-- mei_vsc_hw
    MeiVscDriver o-- mei_device

Loading

Flow diagram for MEI VSC hardware reset with conditional interrupt handling

flowchart TD
    A["mei_vsc_hw_reset(mei_dev, intr_enable)"] --> B["Call vsc_tp_reset(hw->tp)"]
    B --> C{"intr_enable?"}
    C -->|No| D["Return 0"]
    C -->|Yes| E["Call vsc_tp_init(hw->tp, mei_dev->dev)"]
    E --> F["Return result"]

Loading

File-Level Changes

Change	Details	Files
Move VSC TP event notification out of IRQ context into a workqueue with dedicated locking and proper init/teardown.	Include workqueue header and add a work_struct plus an event_notify_mutex to the vsc_tp struct to protect callback and context fields separately from the command mutex. Change the IRQ handler to wake the transfer wait queue and schedule work instead of returning IRQ_WAKE_THREAD, and implement a workqueue callback that safely calls the event_notify callback under the new mutex. Update vsc_tp_register_event_cb to guard updates to the event callback and context with event_notify_mutex. Initialize event_notify_mutex and event_work in probe, and on both probe‑error and remove paths cancel the work and destroy both mutexes after freeing the IRQ.	drivers/misc/mei/vsc-tp.c
Adjust VSC TP GPIO, IRQ, and reset behavior to match upstream semantics and avoid spurious re-enables.	Change the wakeup GPIO requested in probe from "wakeuphost" to "wakeuphostint". Remove re-enabling of the SPI IRQ from vsc_tp_reset to avoid enabling interrupts at the wrong time. Request the IRQ as a purely threaded handler (no hard-IRQ handler function) using vsc_tp_isr as the threaded function.	drivers/misc/mei/vsc-tp.c
Align MEI VSC platform driver’s reset and event-callback lifecycle with the transport layer to avoid re-init and stale callbacks.	Modify mei_vsc_hw_reset so that when intr_enable is false it returns early after vsc_tp_reset instead of calling vsc_tp_init again, preventing unwanted VSC re-initialization. On mei_vsc_probe error and mei_vsc_remove, explicitly clear the VSC TP event callback by registering a NULL callback and context before disabling interrupts.	drivers/misc/mei/platform-vsc.c
Fix Kconfig dependency so MEI VSC is correctly tied to the VSC transport option.	Update MEI Kconfig to add the missing dependency ensuring the VSC-based MEI driver only builds when the corresponding VSC transport is enabled.	drivers/misc/mei/Kconfig

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[deepin-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please ask for approval from opsiff. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• deepin/OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[sourcery-ai]

Hey - I've reviewed your changes and they look great!

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[Copilot]

Pull request overview

Syncs upstream fixes for the Intel MEI VSC transport/path in the 6.6.y kernel, focusing on safer event notification, IRQ/reset handling, and Kconfig dependency correctness.

Changes:

• Move VSC event notification off the IRQ thread into a workqueue; add locking around the event callback registration/invocation.
• Adjust reset/interrupt behavior and GPIO line name usage for host-wakeup signaling.
• Add Kconfig dependency to ensure the VSC HW transport is built with the base MEI stack.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
drivers/misc/mei/vsc-tp.c	Reworks IRQ/event notification flow (workqueue), updates GPIO lookup, and adjusts teardown ordering.
drivers/misc/mei/platform-vsc.c	Tweaks hw_reset behavior and unsets event callback on probe-error/remove.
drivers/misc/mei/Kconfig	Adds missing dependency between VSC HW transport and INTEL_MEI.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

request_threaded_irq() is switched to run vsc_tp_isr() as the threaded handler, but vsc_tp_request_irq() (used by suspend/resume) still references the removed vsc_tp_thread_isr, which will cause a build failure. Please update vsc_tp_request_irq() to match the new IRQ registration pattern (primary handler NULL, threaded handler vsc_tp_isr()) or reintroduce the missing handler as appropriate.

[Copilot]

In vsc_tp_probe(), the IRQ is requested before event_notify_mutex is initialized and before INIT_WORK(&tp->event_work, ...). Since the IRQ handler schedules tp->event_work, an interrupt firing immediately after request_threaded_irq() can queue/execute an uninitialized work item and use an uninitialized mutex. Initialize the mutex and work_struct before requesting the IRQ, or request the IRQ with IRQF_NO_AUTOEN and explicitly enable it after initialization.

[Copilot]

vsc_tp_reset() no longer re-enables the IRQ after the reset sequence (the enable_irq() call was dropped). Since the function still disables the IRQ earlier, this can leak IRQ disable-depth across reset/stop/resume flows and potentially leave the IRQ permanently disabled. Please ensure the IRQ enable/disable balance is preserved (e.g., manage IRQ state only via the MEI intr_enable/intr_disable ops, or otherwise restore the prior IRQ state after reset).

Suggested change

	atomic_set(&tp->assert_cnt, 0);
	atomic_set(&tp->assert_cnt, 0);
	enable_irq(tp->spi->irq);

[Copilot]

Since VSC events are now delivered via vsc_tp's workqueue, mei_stop()/mei_reset() no longer implicitly synchronize with the event callback (it used to run in the IRQ thread). In mei_vsc_remove(), the callback is unset only after mei_stop(), leaving a window where a previously queued VSC event-work item can still run mei_vsc_event_cb() while teardown is in progress. Consider unsetting the callback before calling mei_stop() (and similarly in other teardown paths) so that any queued work will observe a NULL callback and exit quickly.