linux-kernel - Re: [PATCH] pci: add pci_dev_is

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

Message-ID: <20210525140854.GA1186411@bjorn-Precision-5520>
Date:   Tue, 25 May 2021 09:08:54 -0500
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Lambert Wang <lambert.q.wang@...il.com>
Cc:     Bjorn Helgaas <bhelgaas@...gle.com>, linux-pci@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Krzysztof Wilczyński <kw@...ux.com>
Subject: Re: [PATCH] pci: add pci_dev_is_alive API

[+cc Krzysztof]

On Tue, May 25, 2021 at 08:59:25PM +0800, Lambert Wang wrote:
> Device drivers use this API to proactively check if the device
> is alive or not. It is helpful for some PCI devices to detect
> surprise removal and do recovery when Hotplug function is disabled.
> 
> Note: Device in power states larger than D0 is also treated not alive
> by this function.
> 
> Signed-off-by: Lambert Wang <lambert.q.wang@...il.com>
> ---
>  drivers/pci/pci.c   | 23 +++++++++++++++++++++++
>  include/linux/pci.h |  1 +
>  2 files changed, 24 insertions(+)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index b717680377a9..8a7c039b1cd5 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -4659,6 +4659,29 @@ int pcie_flr(struct pci_dev *dev)
>  }
>  EXPORT_SYMBOL_GPL(pcie_flr);
>  
> +/**
> + * pci_dev_is_alive - check the pci device is alive or not
> + * @pdev: the PCI device
> + *
> + * Device drivers use this API to proactively check if the device
> + * is alive or not. It is helpful for some PCI devices to detect
> + * surprise removal and do recovery when Hotplug function is disabled.
> + *
> + * Note: Device in power state larger than D0 is also treated not alive
> + * by this function.
> + *
> + * Returns true if the device is alive.
> + */
> +bool pci_dev_is_alive(struct pci_dev *pdev)
> +{
> +	u16 vendor;
> +
> +	pci_read_config_word(pdev, PCI_VENDOR_ID, &vendor);
> +
> +	return vendor == pdev->vendor;
> +}
> +EXPORT_SYMBOL(pci_dev_is_alive);

This is quite similar to pci_device_is_present().  Does that not do
what you need?

I'm not a big fan of either pci_device_is_present() or
pci_dev_is_alive() because they are racy.  You must assume that even
if they return "true," the device may disappear because of a surprise
removal before you can act on that information.

>  static int pci_af_flr(struct pci_dev *dev, int probe)
>  {
>  	int pos;
> diff --git a/include/linux/pci.h b/include/linux/pci.h
> index c20211e59a57..2a3ba06a7347 100644
> --- a/include/linux/pci.h
> +++ b/include/linux/pci.h
> @@ -1227,6 +1227,7 @@ u32 pcie_bandwidth_available(struct pci_dev *dev, struct pci_dev **limiting_dev,
>  void pcie_print_link_status(struct pci_dev *dev);
>  bool pcie_has_flr(struct pci_dev *dev);
>  int pcie_flr(struct pci_dev *dev);
> +bool pci_dev_is_alive(struct pci_dev *pdev);
>  int __pci_reset_function_locked(struct pci_dev *dev);
>  int pci_reset_function(struct pci_dev *dev);
>  int pci_reset_function_locked(struct pci_dev *dev);
> -- 
> 2.30.2
>