| 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
| ||
|
Message-ID: <SN1PR0301MB1583296CF246F14DCA2F5EEDE0700@SN1PR0301MB1583.namprd03.prod.outlook.com>
Date: Mon, 10 Aug 2015 09:23:09 +0000
From: Ramneek Mehresh <ramneek.mehresh@...escale.com>
To: Alan Stern <stern@...land.harvard.edu>
CC: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"balbi@...com" <balbi@...com>,
"gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
"linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
Li Leo <LeoLi@...escale.com>
Subject: RE: [PATCH 3/8][v2]usb:fsl:otg: Add support to add/remove usb host
driver
> -----Original Message-----
> From: Alan Stern [mailto:stern@...land.harvard.edu]
> Sent: Saturday, August 08, 2015 2:04 AM
> To: Mehresh Ramneek-B31383
> Cc: linux-kernel@...r.kernel.org; balbi@...com;
> gregkh@...uxfoundation.org; linux-usb@...r.kernel.org; Li Yang-Leo-R58472
> Subject: Re: [PATCH 3/8][v2]usb:fsl:otg: Add support to add/remove usb
> host driver
>
> On Wed, 15 Jul 2015, Ramneek Mehresh wrote:
>
> > Add workqueue to add/remove host driver (outside interrupt context)
> > upon each id change.
> >
> > Signed-off-by: Li Yang <leoli@...escale.com>
> > Signed-off-by: Ramneek Mehresh <ramneek.mehresh@...escale.com>
> > ---
> > drivers/usb/host/ehci-fsl.c | 83
> > ++++++++++++++++++++++++++++++++++-----------
> > drivers/usb/host/ehci-fsl.h | 20 +++++++++++
> > 2 files changed, 84 insertions(+), 19 deletions(-)
> >
> > diff --git a/drivers/usb/host/ehci-fsl.c b/drivers/usb/host/ehci-fsl.c
> > index 5352e74..81e4bf5 100644
> > --- a/drivers/usb/host/ehci-fsl.c
> > +++ b/drivers/usb/host/ehci-fsl.c
> > @@ -44,6 +44,34 @@
> >
> > static struct hc_driver __read_mostly fsl_ehci_hc_driver;
> >
> > +#if defined(CONFIG_FSL_USB2_OTG) ||
> > +defined(CONFIG_FSL_USB2_OTG_MODULE)
>
> You've got these #if lines all over the place. They look ugly and make the
> code hard to read. Consider removing them. Or even if you can't remove
> them entirely, removing most of them would help.
>
> Also, instead of testing both CONFIG_FSL_USB2_OTG and
> CONFIG_FSL_USB2_OTG_MODULE, how about testing a single symbol? For
> example:
>
> #if defined(CONFIG_FSL_USB2_OTG) ||
> defined(CONFIG_FSL_USB2_OTG_MODULE)
> #define CHANGE_HCD 1
> #else
> #define CHANGE_HCD 0
> #endif
>
> Then all you need later on is "#if CHANGE_HCD". Or if it's inside a code block,
> just "if (CHANGE_HCD)".
>
Ok, let me see what can be done to minimize these macros. Thanks
> > +static struct ehci_fsl *hcd_to_ehci_fsl(struct usb_hcd *hcd) {
> > + return (struct ehci_fsl *)hcd_to_ehci(hcd)->priv; }
> > +
> > +static void do_change_hcd(struct work_struct *work) {
> > + struct ehci_fsl *ehci_fsl = container_of(work, struct ehci_fsl,
> > + change_hcd_work);
> > + struct usb_hcd *hcd = ehci_fsl->hcd;
> > + void __iomem *non_ehci = hcd->regs;
> > + int retval;
> > +
> > + if (ehci_fsl->hcd_add && !ehci_fsl->have_hcd) {
> > + writel(USBMODE_CM_HOST, non_ehci +
> FSL_SOC_USB_USBMODE);
> > + /* host, gadget and otg share same int line */
> > + retval = usb_add_hcd(hcd, hcd->irq, IRQF_SHARED);
> > + if (retval == 0)
> > + ehci_fsl->have_hcd = 1;
> > + } else if (!ehci_fsl->hcd_add && ehci_fsl->have_hcd) {
> > + usb_remove_hcd(hcd);
> > + ehci_fsl->have_hcd = 0;
>
> Don't you have to turn off the USBMODE_CM_HOST bit here? It looks
> strange to turn it on above but not turn it off again.
>
USBMODE bit defines mode for the controller. It gets auto cleared when controller is
reset in host/gadget driver. After resetting, it's mandatory to write this bit once to define
controller's mode.
> > + }
> > +}
> > +#endif
>
>
> > static int ehci_fsl_drv_suspend(struct device *dev) {
> > struct usb_hcd *hcd = dev_get_drvdata(dev);
> > - struct ehci_fsl *ehci_fsl = hcd_to_ehci_fsl(hcd);
> > void __iomem *non_ehci = hcd->regs;
> > +#if defined(CONFIG_FSL_USB2_OTG) ||
> defined(CONFIG_FSL_USB2_OTG_MODULE)
> > + struct ehci_fsl *ehci_fsl = hcd_to_ehci_fsl(hcd);
> > + struct usb_bus host = hcd->self;
> > +#endif
> > +
> >
> > if (of_device_is_compatible(dev->parent->of_node,
> > "fsl,mpc5121-usb2-dr")) {
> > return ehci_fsl_mpc512x_drv_suspend(dev);
> > }
> >
> > +#if defined(CONFIG_FSL_USB2_OTG) ||
> defined(CONFIG_FSL_USB2_OTG_MODULE)
> > + if (host.is_otg) {
> > + /* remove hcd */
> > + ehci_fsl->hcd_add = 0;
> > + schedule_work(&ehci_fsl->change_hcd_work);
> > + host.is_otg = 0;
>
> Why do you set host.is_otg to 0 here? Why not do it in the work routine?
Well, it can be done there also. I just wanted to keep host-bus house-keeping out
of otg role-changing specific workqueue (where host driver is brought-up/shut-down).
>
> > + return 0;
> > + }
> > +#endif
> > +
> > ehci_prepare_ports_for_controller_suspend(hcd_to_ehci(hcd),
> > device_may_wakeup(dev));
> > if (!fsl_deep_sleep())
> > @@ -540,15 +571,29 @@ static int ehci_fsl_drv_suspend(struct device
> > *dev) static int ehci_fsl_drv_resume(struct device *dev) {
> > struct usb_hcd *hcd = dev_get_drvdata(dev);
> > - struct ehci_fsl *ehci_fsl = hcd_to_ehci_fsl(hcd);
> > struct ehci_hcd *ehci = hcd_to_ehci(hcd);
> > void __iomem *non_ehci = hcd->regs;
> > +#if defined(CONFIG_FSL_USB2_OTG) ||
> defined(CONFIG_FSL_USB2_OTG_MODULE)
> > + struct ehci_fsl *ehci_fsl = hcd_to_ehci_fsl(hcd);
> > + struct usb_bus host = hcd->self;
> > +#endif
> >
> > if (of_device_is_compatible(dev->parent->of_node,
> > "fsl,mpc5121-usb2-dr")) {
> > return ehci_fsl_mpc512x_drv_resume(dev);
> > }
> >
> > +#if defined(CONFIG_FSL_USB2_OTG) ||
> defined(CONFIG_FSL_USB2_OTG_MODULE)
> > + if (host.is_otg) {
> > + /* add hcd */
> > + ehci_fsl->hcd_add = 1;
> > + schedule_work(&ehci_fsl->change_hcd_work);
> > + usb_hcd_resume_root_hub(hcd);
> > + host.is_otg = 0;
>
> Again, why change host.is_otg here? And for that matter, where does
> host.is_otg ever get set to 1?
host.is_otg sets to 1 in phy driver before calling ehci drv suspend/resume.
This is donr to make it clear in host drv suspend/resume that these functions
Are called in context of otg drv
>
> Also, what is the reason for calling usb_hcd_resume_root_hub()? It won't
> do anything, because it will run before the scheduled work, so there won't
> be a root hub for it to resume.
>
Well, you're right...it's not a surety that usb_hcd_resume_root_hub() will run after
scheduled work...i'll move it inside do_change_hcd() after usb_add_hcd(). Thanks.
> > + return 0;
> > + }
> > +#endif
> > +
> > ehci_prepare_ports_for_controller_resume(ehci);
> > if (!fsl_deep_sleep())
> > return 0;
> > diff --git a/drivers/usb/host/ehci-fsl.h b/drivers/usb/host/ehci-fsl.h
> > index dbd292e..3fd1fd0 100644
> > --- a/drivers/usb/host/ehci-fsl.h
> > +++ b/drivers/usb/host/ehci-fsl.h
> > @@ -62,4 +62,24 @@
> > #define UTMI_PHY_EN (1<<9)
> > #define ULPI_PHY_CLK_SEL (1<<10)
> > #define PHY_CLK_VALID (1<<17)
> > +
> > +struct ehci_fsl {
> > +#ifdef CONFIG_PM
> > + /* Saved USB PHY settings, need to restore after deep sleep. */
> > + u32 usb_ctrl;
> > +#endif
> > + struct usb_hcd *hcd;
> > +#if defined(CONFIG_FSL_USB2_OTG) ||
> defined(CONFIG_FSL_USB2_OTG_MODULE)
> > + struct work_struct change_hcd_work;
> > +#endif
>
> Again, try to eliminate these #if's. There really isn't anything wrong with
> allocating the space for these things even in a non-OTG build.
>
> > + /*
> > + * store current hcd state for otg;
> > + * have_hcd is true when host drv al already part of otg framework,
> > + * otherwise false;
> > + * hcd_add is true when otg framework wants to add host
> > + * drv as part of otg;flase when it wants to remove it
> > + */
> > + unsigned have_hcd:1;
> > + unsigned hcd_add:1;
> > +};
> > #endif /* _EHCI_FSL_H */
>
> Alan Stern
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists