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: <1521706339.3717.141.camel@mhfsdcap03>
Date:   Thu, 22 Mar 2018 16:12:19 +0800
From:   Chunfeng Yun <chunfeng.yun@...iatek.com>
To:     Roger Quadros <rogerq@...com>
CC:     Mathias Nyman <mathias.nyman@...el.com>,
        Martin Blumenstingl <martin.blumenstingl@...glemail.com>,
        Matthias Brugger <matthias.bgg@...il.com>,
        <linux-kernel@...r.kernel.org>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-usb@...r.kernel.org>, <linux-mediatek@...ts.infradead.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Subject: Re: [usb-next PATCH] usb: xhci: skip phys initialization of shared
 hcd

hi,

On Wed, 2018-03-21 at 14:27 +0200, Roger Quadros wrote:
> On 21/03/18 13:50, Chunfeng Yun wrote:
> > Hi Roger,
> > 
> > On Wed, 2018-03-21 at 11:05 +0200, Roger Quadros wrote:
> >> Hi Chunfeng,
> >>
> >> On 21/03/18 08:12, Chunfeng Yun wrote:
> >>> The phys has already been initialized when add primary hcd,
> >>> including usb2 phys and usb3 phys also if exist, so needn't
> >>> re-parse "phys" property again.
> >>>
> >>> Signed-off-by: Chunfeng Yun <chunfeng.yun@...iatek.com>
> >>> ---
> >>>  drivers/usb/host/xhci-mtk.c  | 1 +
> >>>  drivers/usb/host/xhci-plat.c | 1 +
> >>>  2 files changed, 2 insertions(+)
> >>>
> >>> diff --git a/drivers/usb/host/xhci-mtk.c b/drivers/usb/host/xhci-mtk.c
> >>> index 7334da9..6bb23fb 100644
> >>> --- a/drivers/usb/host/xhci-mtk.c
> >>> +++ b/drivers/usb/host/xhci-mtk.c
> >>> @@ -554,6 +554,7 @@ static int xhci_mtk_probe(struct platform_device *pdev)
> >>>  	if (HCC_MAX_PSA(xhci->hcc_params) >= 4)
> >>>  		xhci->shared_hcd->can_do_streams = 1;
> >>>  
> >>> +	xhci->shared_hcd->skip_phy_initialization = 1;
> >>>  	ret = usb_add_hcd(xhci->shared_hcd, irq, IRQF_SHARED);
> >>>  	if (ret)
> >>>  		goto dealloc_usb2_hcd;
> >>> diff --git a/drivers/usb/host/xhci-plat.c b/drivers/usb/host/xhci-plat.c
> >>> index 6700e5e..65a4294 100644
> >>> --- a/drivers/usb/host/xhci-plat.c
> >>> +++ b/drivers/usb/host/xhci-plat.c
> >>> @@ -294,6 +294,7 @@ static int xhci_plat_probe(struct platform_device *pdev)
> >>>  	if (HCC_MAX_PSA(xhci->hcc_params) >= 4)
> >>>  		xhci->shared_hcd->can_do_streams = 1;
> >>>  
> >>> +	xhci->shared_hcd->skip_phy_initialization = 1;
> >>
> >> I think this is unnecessary.
> >> There aren't separate PHYs for the primary and shared HCDs right?
> > Yes, the primary and shared HCDs will get the same PHYs provided by
> > "phys" property.
> > 
> >>
> >> Also how can you be sure that phy_init() has been done for all platform HCD users?
> > Here I assume that "phys" and "usb-phy" properties are not used at the
> > same time. There are two cases:
> > 1. Using "usb-phy": no "phys" to be got by shared HCD.
> > 2. Using "phys": the primary HCD will get those phys, so shared HCD can
> > skip them.
> 
> Now I see the problem.
> But the right fix should be in core/hcd.c
> 
> diff --git a/drivers/usb/core/hcd.c b/drivers/usb/core/hcd.c
> index 2884607..1d1da12 100644
> --- a/drivers/usb/core/hcd.c
> +++ b/drivers/usb/core/hcd.c
> @@ -2757,7 +2757,7 @@ int usb_add_hcd(struct usb_hcd *hcd,
>  		}
>  	}
>  
> -	if (!hcd->skip_phy_initialization) {
> +	if (!hcd->skip_phy_initialization && usb_hcd_is_primary_hcd(hcd)) {
It's a better way.
Thanks
>  		hcd->phy_roothub = usb_phy_roothub_init(hcd->self.sysdev);
>  		if (IS_ERR(hcd->phy_roothub)) {
>  			retval = PTR_ERR(hcd->phy_roothub);
> 
> > 
> >>
> >>>  	ret = usb_add_hcd(xhci->shared_hcd, irq, IRQF_SHARED);
> >>>  	if (ret)
> >>>  		goto dealloc_usb2_hcd;
> >>>
> >>
> > 
> > 
> 


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ