linux-kernel - Re: [RFC][PATCH] iwlwifi: Add pci .shutdown() hook for iwlwifi driver

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: <CAO7dBbV=D3N31L-VPS=2Vtreqc-4drKYHT1xWrKphT3J_G5Ndw@mail.gmail.com>
Date:   Thu, 3 Aug 2023 16:45:04 +0800
From:   Tao Liu <ltao@...hat.com>
To:     gregory.greenman@...el.com, kvalo@...nel.org,
        johannes.berg@...el.com, mukesh.sisodiya@...el.com
Cc:     linux-wireless@...r.kernel.org, linux-kernel@...r.kernel.org,
        kexec@...ts.infradead.org
Subject: Re: [RFC][PATCH] iwlwifi: Add pci .shutdown() hook for iwlwifi driver

Add kexec to the CC list so kexec people can know this.

On Thu, Aug 3, 2023 at 10:55 AM Tao Liu <ltao@...hat.com> wrote:
>
> Previously no .shutdown() hook is implemented for iwlwifi driver, a
> ETIMEDOUT error will occur during the kexec kernel bootup. As a
> consequence, wifi is unusable after kexec into the new kernel.
>
> This issue is observed and patch tested on the following wireless cards:
>
> 1) Network controller: Intel Corporation Comet Lake PCH-LP CNVi WiFi,
>    Subsystem: Intel Corporation Wi-Fi 6 AX201 160MHz
> 2) Network controller: Intel Corporation Wireless-AC 9260,
>    Subsystem: Intel Corporation Device e014
>
> Signed-off-by: Tao Liu <ltao@...hat.com>
> ---
>
> Hi folks,
>
> This is a RFC patch and I'm not sure about the correctness of the code,
> especially about the pci_clear_master() part. What I want is to stop any
> ongoing DMA access, in case if the memory overwritting during kexec
> kernel bootup. But there is already pci_clear_master(pci_dev) in
> drivers/pci/pci-driver.c:pci_device_shutdown(), so I'm not sure if it is
> still needed in the driver side. And I only tested the patch against the
> above 2 wireless cards and worked OK, not sure if it can work for others.
> Please review the patch, thanks in advance!
>
> Thanks,
> Tao Liu
>
> ---
>  drivers/net/wireless/intel/iwlwifi/pcie/drv.c | 13 +++++++++++++
>  1 file changed, 13 insertions(+)
>
> diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/drv.c b/drivers/net/wireless/intel/iwlwifi/pcie/drv.c
> index 73c1fb3c0c5e..24c4c2dd7cb0 100644
> --- a/drivers/net/wireless/intel/iwlwifi/pcie/drv.c
> +++ b/drivers/net/wireless/intel/iwlwifi/pcie/drv.c
> @@ -1513,6 +1513,18 @@ static void iwl_pci_remove(struct pci_dev *pdev)
>         iwl_trans_pcie_free(trans);
>  }
>
> +static void iwl_pci_shutdown(struct pci_dev *pdev)
> +{
> +       struct iwl_trans *trans = pci_get_drvdata(pdev);
> +
> +       if (!trans)
> +               return;
> +
> +       iwl_drv_stop(trans->drv);
> +
> +       pci_clear_master(pdev);
> +}
> +
>  #ifdef CONFIG_PM_SLEEP
>
>  static int iwl_pci_suspend(struct device *device)
> @@ -1583,6 +1595,7 @@ static struct pci_driver iwl_pci_driver = {
>         .id_table = iwl_hw_card_ids,
>         .probe = iwl_pci_probe,
>         .remove = iwl_pci_remove,
> +       .shutdown = iwl_pci_shutdown,
>         .driver.pm = IWL_PM_OPS,
>  };
>
> --
> 2.40.1
>