mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Kent Gibson <warthog618@gmail.com>
To: Adrian Ambrozewicz <aambrozewicz@nvidia.com>
Cc: Bartosz Golaszewski <brgl@kernel.org>,
	Linus Walleij <linusw@kernel.org>,
	Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>,
	"Paul E. McKenney" <paulmck@kernel.org>,
	Herve Codina <herve.codina@bootlin.com>,
	Jia Jia <physicalmtea@gmail.com>,
	Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
	linux-gpio@vger.kernel.org, linux-kernel@vger.kernel.org,
	stable@vger.kernel.org, Sashiko <sashiko-bot@kernel.org>
Subject: Re: [PATCH v7 8/8] gpiolib: remove the shared GPIO proxies before freeing the remaining IRQs
Date: Sat, 21 Nov 2026 13:08:18 +0800	[thread overview]
Message-ID: <20261121050818.GC17013@rigel> (raw)
In-Reply-To: <20261120-gpiolib-chip-removal-v7-8-aambrozewicz@nvidia.com>

On Fri, Nov 20, 2026 at 10:14:00AM +0200, Adrian Ambrozewicz wrote:
> gpio_shared_proxy_to_irq() returns the parent chip's IRQ, so an edge
> request on a proxy line holds an IRQ of the parent.
> 
> When the parent is removed, gpiochip_remove() frees the IRQs still
> requested on its lines before it removes the proxies in
> gpio_device_teardown_shared(). The release of a proxy line request is
> serialized only with the proxy's own removal notifier, so if
> gpiod_free_irqs() frees the IRQ between the release's irq_has_action()
> check and its free_irq(), free_irq() warns:
> 
>   Trying to free already-free IRQ 22
> 
> Remove the proxies first, so the character device on each proxy frees
> its IRQs before the rest are forcibly freed.
> 
> Reproduced on arm64 QEMU with a shared gpio-sim line and a test delay
> before the release's free_irq(): 3 of 3 runs warned; with this change,
> none.
> 
> Fixes: 1e4f6db614a3 ("gpiolib: support shared GPIOs in core subsystem code")
> Cc: stable@vger.kernel.org
> Reported-by: Sashiko <sashiko-bot@kernel.org>
> Closes: https://sashiko.dev/#/patchset/20261103-gpiolib-chip-removal-v5-0-aambrozewicz@nvidia.com?part=7
> Signed-off-by: Adrian Ambrozewicz <aambrozewicz@nvidia.com>
> ---
>  drivers/gpio/gpiolib.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> index e0ba17a73c4b..57024a1b042b 100644
> --- a/drivers/gpio/gpiolib.c
> +++ b/drivers/gpio/gpiolib.c
> @@ -1461,12 +1461,12 @@ void gpiochip_remove(struct gpio_chip *gc)
>  
>  	/*
>  	 * No new interrupts can be requested through the character device now.
> -	 * Have it free the ones it holds before forcibly freeing the rest, and
> -	 * both before the irqchip and its mappings go away.
> +	 * Have it free the ones it holds, also on the shared proxies, before
> +	 * forcibly freeing the rest and removing the irqchip.
>  	 */
>  	blocking_notifier_call_chain(&gdev->device_notifier, 0, NULL);
> -	gpiochip_free_remaining_irqs(gc);
>  	gpio_device_teardown_shared(gdev);
> +	gpiochip_free_remaining_irqs(gc);
>  	gpiochip_irqchip_remove(gc);
>  	acpi_gpiochip_remove(gc);
>  	of_gpiochip_remove(gc);
> -- 
> 2.43.0
> 

Everything it touches was added in patch 7, so it needs a dependency note as
per my other patch comments.

Cheers,
Kent.

      reply	other threads:[~2026-11-21  5:08 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-20  8:13 [PATCH v7 0/8] gpiolib: fix chip removal with lines still requested from user space Adrian Ambrozewicz
2026-11-20  8:13 ` [PATCH v7 1/8] gpiolib: wait for pending label frees before cleaning up desc_srcu Adrian Ambrozewicz
2026-11-20  8:13 ` [PATCH v7 2/8] gpiolib: free the label of lines released after their chip was removed Adrian Ambrozewicz
2026-11-20  8:13 ` [PATCH v7 3/8] gpiolib: cdev: do not free line IRQs again after the chip is removed Adrian Ambrozewicz
2026-11-20  8:13 ` [PATCH v7 4/8] gpiolib: cdev: don't sleep in the SRCU read section in read() Adrian Ambrozewicz
2026-11-20  8:13 ` [PATCH v7 5/8] gpiolib: cdev: don't leak the IRQ name when the chip is removed Adrian Ambrozewicz
2026-11-21  4:52   ` Kent Gibson
2026-11-20  8:13 ` [PATCH v7 6/8] gpiolib: cdev: don't report a failed line read as an edge Adrian Ambrozewicz
2026-11-20  8:13 ` [PATCH v7 7/8] gpiolib: let the character device free its IRQs first on chip removal Adrian Ambrozewicz
2026-11-21  4:46   ` Kent Gibson
2026-11-20  8:14 ` [PATCH v7 8/8] gpiolib: remove the shared GPIO proxies before freeing the remaining IRQs Adrian Ambrozewicz
2026-11-21  5:08   ` Kent Gibson [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=20261121050818.GC17013@rigel \
    --to=warthog618@gmail.com \
    --cc=aambrozewicz@nvidia.com \
    --cc=andriy.shevchenko@linux.intel.com \
    --cc=bartosz.golaszewski@oss.qualcomm.com \
    --cc=brgl@kernel.org \
    --cc=herve.codina@bootlin.com \
    --cc=linusw@kernel.org \
    --cc=linux-gpio@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=paulmck@kernel.org \
    --cc=physicalmtea@gmail.com \
    --cc=sashiko-bot@kernel.org \
    --cc=stable@vger.kernel.org \
    /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

all inboxes | Powered by JetHome®