linux-kernel - Re: Is there something wrong with v5.10.234 patch 664760c49d98 ("xhci: use pm_ptr() instead of #ifdef for CONFIG

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: <dac30597-6d62-46f4-9493-059344e253c4@huawei.com>
Date: Thu, 6 Mar 2025 11:35:26 +0800
From: "zhangzekun (A)" <zhangzekun11@...wei.com>
To: <gregkh@...uxfoundation.org>
CC: <akpm@...ux-foundation.org>, <jslaby@...e.cz>,
	<linux-kernel@...r.kernel.org>, <lwn@....net>, <stable@...r.kernel.org>,
	<torvalds@...ux-foundation.org>, <chenweilong@...wei.com>,
	<liuyongqiang13@...wei.com>, <arnd@...db.de>
Subject: Re: Is there something wrong with v5.10.234 patch 664760c49d98
 ("xhci: use pm_ptr() instead of #ifdef for CONFIG_PM conditionals")?

Sorry for make noise. Please ignore this messgae.
The latter patch c762b76981fc ("Partial revert of xhci: use pm_ptr() 
instead #ifdef for CONFIG_PM conditionals") patial revert the patch, So, 
there is no such problem with the patch.

Thanks,
Zekunk

在 2025/3/6 10:33, Zhang Zekun 写道:
> Hi, Greg,
> The stable patch 664760c49d98 ("xhci: use pm_ptr() instead of #ifdef for CONFIG_PM conditionals") in linux-5.10.y, is different with the mainline patch. In the following code.
> 
> mainline:
> -#ifdef CONFIG_PM
> -       xhci_pci_hc_driver.pci_suspend = xhci_pci_suspend;
> -       xhci_pci_hc_driver.pci_resume = xhci_pci_resume;
> -       xhci_pci_hc_driver.pci_poweroff_late = xhci_pci_poweroff_late;
> -       xhci_pci_hc_driver.shutdown = xhci_pci_shutdown;
> -#endif
> +       xhci_pci_hc_driver.pci_suspend = pm_ptr(xhci_pci_suspend);		
> +       xhci_pci_hc_driver.pci_resume = pm_ptr(xhci_pci_resume);		
> +       xhci_pci_hc_driver.pci_poweroff_late = pm_ptr(xhci_pci_poweroff_late);	
> +       xhci_pci_hc_driver.shutdown = pm_ptr(xhci_pci_shutdown);		
> 
> linux-5.10.y:
> -#ifdef CONFIG_PM
>   	xhci_pci_hc_driver.pci_suspend = xhci_pci_suspend;
>   	xhci_pci_hc_driver.pci_resume = xhci_pci_resume;
>   	xhci_pci_hc_driver.shutdown = xhci_pci_shutdown;
> -#endif
> 
> xhci_pci_shutdown() in mainline is wrapped with pm_ptr(), which is NULL if CONFIG_PM is not enabled, but for linux-5.10.y, xhci_pci_shutdown() it will not be converted to a NULL pointer. The .shutdown() function seems has been used in usb_hcd_platform_shutdown() in usb platfrom device shutdown routine:
> 
> drivers/usb/host/ehci-atmel.c:  .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-brcm.c:   .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-exynos.c: .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-fsl.c:    .shutdown = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-grlib.c:  .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-mxc.c:    .shutdown = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-npcm7xx.c:        .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-omap.c:   .shutdown               = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-orion.c:  .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-platform.c:       .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-ppc-of.c: .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-spear.c:  .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-st.c:     .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ehci-xilinx-of.c:      .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-at91.c:   usb_hcd_platform_shutdown(pdev);
> drivers/usb/host/ohci-at91.c:   .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-da8xx.c:  .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-omap.c:   .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-platform.c:       .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-ppc-of.c: .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-pxa27x.c: .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-s3c2410.c:        .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-sm501.c:  .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-st.c:     .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/ohci-tmio.c:   .shutdown       = usb_hcd_platform_shutdown,
> drivers/usb/host/xhci-plat.c:   .shutdown = usb_hcd_platform_shutdown,
>   
> For the mainline patch seems does not want to call xhci_pci_shutdown() when CONFIG_PM is diabled, is there something wrong with the stable patch?