Linux USB
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Atharva Tiwari" <atharvatiwarilinuxdev@gmail.com>
Cc: "open list:USB XHCI DRIVER" <linux-usb@vger.kernel.org>
Subject: Re: [PATCH v2] usb: xhci-pci: Add Power handling for Card readers on Apple systems
Date: Tue, 17 Nov 2026 00:31:44 +0000	[thread overview]
Message-ID: <sashiko-outbox-218004@kernel.org> (raw)
In-Reply-To: <20261117000034.7641-1-atharvatiwarilinuxdev@gmail.com>

Thank you for your contribution! Sashiko AI review found 3 potential
issues to consider:
- [Medium] usb: xhci-pci: xhci_pci_suspend() powers off Apple card
  reader in runtime PM
- [Medium] usb: xhci-pci: missing xhci_apple_card_reader_power() off on
  teardown
- [Low] usb: xhci-pci: duplicate SPWR call and early power-on in
  xhci_pci_quirks()
--

commit 9d162606edd929676ba53f7638bfed5c72201d2d
Author: Atharva Tiwari <atharvatiwarilinuxdev@gmail.com>

usb: xhci-pci: Add Power handling for Card readers on Apple systems

This commit adds an XHCI_APPLE_CARD_READER_POWER quirk on x86 Apple
machines to toggle card reader power via the ACPI SPWR method during
probe, suspend, and resume.

> diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c
> index e278db940a583..31961de9761f8 100644
> --- a/drivers/usb/host/xhci-pci.c
> +++ b/drivers/usb/host/xhci-pci.c
[ ... ]
> @@ -517,6 +518,18 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci)
>  	if (xhci->hci_version >= 0x120)
>  		xhci->quirks |= XHCI_DEFAULT_PM_RUNTIME_ALLOW;
>  
> +#ifdef CONFIG_ACPI
> +	/* On Apple Machine with Card readers, we have to use SPWR(0/1) to turn
> +	 * off/on the card reader, And on boot the card reader is off, so we turn
> +	 * it on.
> +	 */
> +	if (x86_apple_machine && ACPI_HANDLE(&pdev->dev) &&
> +			acpi_has_method(ACPI_HANDLE(&pdev->dev), "SPWR")) {
> +		acpi_execute_simple_method(ACPI_HANDLE(&pdev->dev), "SPWR", 1);
> +		xhci->quirks |= XHCI_APPLE_CARD_READER_POWER;
> +	}
> +#endif /* CONFIG_ACPI */
> +
>  	if (xhci->quirks & XHCI_RESET_ON_RESUME)
>  		xhci_dbg_trace(xhci, trace_xhci_dbg_quirks,
>  				"QUIRK: Resetting on resume");

[Severity: Medium]
Does this leave the card reader powered on if probe fails later in
xhci_gen_setup() or xhci_pci_common_probe(), or when the driver is unbound
or the system shuts down?

Neither the probe error paths, xhci_pci_remove(), nor xhci_pci_shutdown()
call xhci_apple_card_reader_power(pdev, 0), and pci_device_remove() wakes
the controller via pm_runtime_get_sync() -> xhci_pci_resume() before
calling xhci_pci_remove(), turning SPWR(1) back on before the HCD is
stopped.

> @@ -562,10 +575,14 @@ static void xhci_find_lpm_incapable_ports(struct usb_hcd *hcd, struct usb_device
>  		}
>  	}
>  }
> -
> +static void xhci_apple_card_reader_power(struct pci_dev *pdev, u64 on)
> +{
> +	acpi_execute_simple_method(ACPI_HANDLE(&pdev->dev), "SPWR", on);
> +}
>  #else
>  static void xhci_pme_acpi_rtd3_enable(struct pci_dev *dev) { }
>  static void xhci_find_lpm_incapable_ports(struct usb_hcd *hcd, struct usb_device *hdev) { }
> +static void xhci_apple_card_reader_power(struct pci_dev *pdev, u64 on) { }
>  #endif /* CONFIG_ACPI */

[Severity: Low]
This isn't a bug, but xhci_pci_quirks() adds a separate #ifdef CONFIG_ACPI
block and open-codes acpi_execute_simple_method() instead of using
xhci_apple_card_reader_power(), and it powers on the device before
xhci_gen_setup() halts and resets the controller.

Since ACPI_HANDLE() and acpi_has_method() already have !CONFIG_ACPI stubs,
xhci_pci_quirks() could just set XHCI_APPLE_CARD_READER_POWER and let
xhci_pci_setup() call xhci_apple_card_reader_power(pdev, 1) after
xhci_gen_setup() succeeds, matching XHCI_PME_STUCK_QUIRK and
xhci_pme_acpi_rtd3_enable().

> @@ -853,6 +870,9 @@ static int xhci_pci_suspend(struct usb_hcd *hcd, bool do_wakeup)
>  	/* synchronize irq when using MSI-X */
>  	xhci_msix_sync_irqs(xhci);
>  
> +	if (!ret && (xhci->quirks & XHCI_APPLE_CARD_READER_POWER))
> +		xhci_apple_card_reader_power(pdev, 0);
> +
>  	if (ret && (xhci->quirks & XHCI_SSIC_PORT_UNUSED))
>  		xhci_ssic_port_unused_quirk(hcd, false);

[Severity: Medium]
Does calling xhci_apple_card_reader_power(pdev, 0) unconditionally in
xhci_pci_suspend() break runtime PM and wake-on-disconnect on the card
reader port?

hcd_pci_runtime_suspend() also calls suspend_common() ->
xhci_pci_suspend() (and controllers with hci_version >= 0x120 enable
runtime PM by default via XHCI_DEFAULT_PM_RUNTIME_ALLOW), so cutting power
with SPWR(0) during runtime suspend prevents the card reader from
detecting card insertion or signaling remote wakeup while the controller
is idle.

Furthermore, when do_wakeup is true, xhci_bus_suspend() arms PORT_WKDISC_E
on connected ports and xhci_suspend() -> xhci_disable_hub_port_wake()
preserves PORT_WAKE_BITS; cutting power to an attached USB card reader
after xhci_suspend() causes a disconnect on the port that can trigger a
spurious wake event and drops the device's power session without notifying
USB core via usb_root_hub_lost_power() on resume.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261117000034.7641-1-atharvatiwarilinuxdev@gmail.com?part=1

      reply	other threads:[~2026-11-17  0:31 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-17  0:00 [PATCH v2] usb: xhci-pci: Add Power handling for Card readers on Apple systems Atharva Tiwari
2026-11-17  0:31 ` sashiko-bot [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=sashiko-outbox-218004@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=atharvatiwarilinuxdev@gmail.com \
    --cc=linux-usb@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox