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: <20230102063536.GD74470@ubuntu>
Date:   Mon, 2 Jan 2023 15:35:36 +0900
From:   Jung Daehwan <dh10.jung@...sung.com>
To:     Arnd Bergmann <arnd@...db.de>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Thinh Nguyen <Thinh.Nguyen@...opsys.com>,
        Mathias Nyman <mathias.nyman@...el.com>,
        Felipe Balbi <balbi@...nel.org>,
        "open list:USB SUBSYSTEM" <linux-usb@...r.kernel.org>,
        "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" 
        <devicetree@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>, sc.suh@...sung.com,
        taehyun.cho@...sung.com, jh0801.jung@...sung.com,
        eomji.oh@...sung.com
Subject: Re: [RFC PATCH v2 1/3] usb: support Samsung Exynos xHCI Controller

On Thu, Dec 29, 2022 at 03:44:02PM +0100, Arnd Bergmann wrote:
> On Thu, Dec 29, 2022, at 10:57, Daehwan Jung wrote:
> > Currently, dwc3 invokes just xhci platform driver without any data.
> > We add xhci node as child of dwc3 node in order to get data from
> > device tree. It populates "xhci" child by name during initialization
> > of host. This patch only effects if dwc3 node has a child named "xhci"
> > not to disturb original path.
> 
> Using child nodes is not the normal way of abstracting a soc specific
> variant of a device, though there are some USB host drivers that
> do this. Just use the node itself and add whatever samsung specific
> properties are needed based on the compatible string.
> 

I've tried to use existing platform drivers(dwc3, xhci-plat).
Dwc3 has host mode and I think it's necessary to have child node.
Currently we can't use compatible string but can just use xhci platform driver
itself. That's why I modified to have a xhci child. It makes us to use
specific properties.

I don't find a better way even if it's not the normal way.
I'm going to talk with other maintainers(dwc3, xhci-plat) to solve the
problem.

> > @@ -86,6 +90,33 @@ static void xhci_plat_quirks(struct device *dev, 
> > struct xhci_hcd *xhci)
> >  	xhci->quirks |= XHCI_PLAT | priv->quirks;
> >  }
> > 
> > +static int xhci_plat_bus_suspend(struct usb_hcd *hcd)
> > +{
> > +	struct xhci_hcd	*xhci = hcd_to_xhci(hcd);
> > +
> > +	if (xhci->quirks & XHCI_ROOTHUB_WAKEUP) {
> > +		if (hcd == xhci->main_hcd)
> > +			__pm_relax(xhci->main_wakelock);
> > +		else
> > +			__pm_relax(xhci->shared_wakelock);
> > +	}
> > +
> > +	return xhci_bus_suspend(hcd);
> > +}
> > +
> > +static int xhci_plat_bus_resume(struct usb_hcd *hcd)
> > +{
> > +	struct xhci_hcd	*xhci = hcd_to_xhci(hcd);
> > +
> > +	if (xhci->quirks & XHCI_ROOTHUB_WAKEUP) {
> > +		if (hcd == xhci->main_hcd)
> > +			__pm_stay_awake(xhci->main_wakelock);
> > +		else
> > +			__pm_stay_awake(xhci->shared_wakelock);
> > +	}
> > +	return xhci_bus_resume(hcd);
> > +}
> 
> It looks like these are no longer tied to the Samsung
> device type, which would be a step in the right direction,
> but I think adding this should be a separate patch since
> it is not a hardware specific change but a new feature.
> 
>     Arnd
> 

Thanks for the comment. I will separate and fix commit msg on next
submission.

Best Regards,
Jung Deahwan


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ