linux-kernel - Re: [PATCH v5] gpio: bt8xx: use generic PCI PM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <aPEfR_lwO1i2wGYb@gmail.com>
Date: Thu, 16 Oct 2025 16:37:27 +0000
From: Vaibhav Gupta <vaibhavgupta40@...il.com>
To: Bjorn Helgaas <helgaas@...nel.org>
Cc: Michael Buesch <m@...s.ch>, Bartosz Golaszewski <brgl@...ev.pl>,
	Linus Walleij <linus.walleij@...aro.org>,
	linux-gpio@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5] gpio: bt8xx: use generic PCI PM

On Mon, Oct 13, 2025 at 12:43:19PM -0500, Bjorn Helgaas wrote:
> On Mon, Oct 13, 2025 at 03:52:59PM +0000, Vaibhav Gupta wrote:
> > Switch to the new generic PCI power management framework and remove legacy
> > callbacks like .suspend() and .resume(). With the generic framework, the
> > standard PCI related work like:
> > 	- pci_save/restore_state()
> > 	- pci_enable/disable_device()
> > 	- pci_set_power_state()
> > is handled by the PCI core and this driver should implement only gpio-bt8xx
> > specific operations in its respective callback functions.
> 
> Tiny nits:
> 
> s/use generic PCI PM/use generic power management/  # subject; not PCI
> s/new generic PCI power/generic power management/   # no longer "new" :)
> 
> > Signed-off-by: Vaibhav Gupta <vaibhavgupta40@...il.com>
> > ---
> >  drivers/gpio/gpio-bt8xx.c | 29 +++++++----------------------
> >  1 file changed, 7 insertions(+), 22 deletions(-)
> > 
> > diff --git a/drivers/gpio/gpio-bt8xx.c b/drivers/gpio/gpio-bt8xx.c
> > index 05401da03ca3..70b49c385b43 100644
> > --- a/drivers/gpio/gpio-bt8xx.c
> > +++ b/drivers/gpio/gpio-bt8xx.c
> > @@ -52,10 +52,8 @@ struct bt8xxgpio {
> >  	struct pci_dev *pdev;
> >  	struct gpio_chip gpio;
> >  
> > -#ifdef CONFIG_PM
> >  	u32 saved_outen;
> >  	u32 saved_data;
> > -#endif
> >  };
> >  
> >  #define bgwrite(dat, adr)	writel((dat), bg->mmio+(adr))
> > @@ -224,9 +222,9 @@ static void bt8xxgpio_remove(struct pci_dev *pdev)
> >  	pci_disable_device(pdev);
> >  }
> >  
> > -#ifdef CONFIG_PM
> > -static int bt8xxgpio_suspend(struct pci_dev *pdev, pm_message_t state)
> > +static int __maybe_unused bt8xxgpio_suspend(struct device *dev)
> 
> I think __maybe_unused is unnecessary because of this path:
> 
>   DEFINE_SIMPLE_DEV_PM_OPS
>     _DEFINE_DEV_PM_OPS
>       SYSTEM_SLEEP_PM_OPS
>         pm_sleep_ptr
> 	  PTR_IF(IS_ENABLED(CONFIG_PM_SLEEP), (_ptr))
> 
> Detailed explanation here:
>   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/include/linux/util_macros.h?id=v6.17#n86
> 
Fixed them in v6.
> >  {
> > +	struct pci_dev *pdev = to_pci_dev(dev);
> >  	struct bt8xxgpio *bg = pci_get_drvdata(pdev);
> >  
> >  	scoped_guard(spinlock_irqsave, &bg->lock) {
> 
> Unrelated to this patch, but it's not clear to me why bg->lock is
> needed here (or maybe anywhere).
> 
>   $ git grep -l "static int.*_suspend" drivers/gpio | wc -l
>   26
>   $ git grep -W "static int.*_suspend" drivers/gpio | grep lock | grep -v unlock | wc -l
> 
> It looks like only about 8 of the 26 gpio drivers that implement
> .suspend() protect it with a lock.  I would expect a lock to be needed
> in all of them or none of them.
> 

I can have a look at them.

regards,
Vaibhav
> > @@ -238,23 +236,13 @@ static int bt8xxgpio_suspend(struct pci_dev *pdev, pm_message_t state)
> >  		bgwrite(0x0, BT848_GPIO_OUT_EN);
> >  	}
> >  
> > -	pci_save_state(pdev);
> > -	pci_disable_device(pdev);
> > -	pci_set_power_state(pdev, pci_choose_state(pdev, state));
> > -
> >  	return 0;
> >  }
> >  
> > -static int bt8xxgpio_resume(struct pci_dev *pdev)
> > +static int __maybe_unused bt8xxgpio_resume(struct device *dev)
> >  {
> > +	struct pci_dev *pdev = to_pci_dev(dev);
> >  	struct bt8xxgpio *bg = pci_get_drvdata(pdev);
> > -	int err;
> > -
> > -	pci_set_power_state(pdev, PCI_D0);
> > -	err = pci_enable_device(pdev);
> > -	if (err)
> > -		return err;
> > -	pci_restore_state(pdev);
> >  
> >  	guard(spinlock_irqsave)(&bg->lock);
> >  
> > @@ -267,10 +255,8 @@ static int bt8xxgpio_resume(struct pci_dev *pdev)
> >  
> >  	return 0;
> >  }
> > -#else
> > -#define bt8xxgpio_suspend NULL
> > -#define bt8xxgpio_resume NULL
> > -#endif /* CONFIG_PM */
> > +
> > +static DEFINE_SIMPLE_DEV_PM_OPS(bt8xxgpio_pm_ops, bt8xxgpio_suspend, bt8xxgpio_resume);
> >  
> >  static const struct pci_device_id bt8xxgpio_pci_tbl[] = {
> >  	{ PCI_DEVICE(PCI_VENDOR_ID_BROOKTREE, PCI_DEVICE_ID_BT848) },
> > @@ -286,8 +272,7 @@ static struct pci_driver bt8xxgpio_pci_driver = {
> >  	.id_table	= bt8xxgpio_pci_tbl,
> >  	.probe		= bt8xxgpio_probe,
> >  	.remove		= bt8xxgpio_remove,
> > -	.suspend	= bt8xxgpio_suspend,
> > -	.resume		= bt8xxgpio_resume,
> > +	.driver.pm	= &bt8xxgpio_pm_ops,
> >  };
> >  
> >  module_pci_driver(bt8xxgpio_pci_driver);
> > -- 
> > 2.51.0
> >