[PATCH] usb: xhci-mtk: fix AHB bus hang up caused by roothubs polling

[PATCH] usb: xhci-mtk: fix AHB bus hang up caused by roothubs polling

[Chunfeng Yun]

when ip fail to enter sleep mode, register access protection will
be disabed, at the same time if all clocks are disabled, access
register will hang up AHB bus.
the common case causes ip sleep fail is that after all ports enter
U3 but before ip enters sleep mode, a port receives a resume
signal('K'). this will happens when such as clicks mouse to try to
do remote wakeup to stop system enter suspend.
so stop polling roothubs to avoid access xHCI register on bus
suspend, and restart it when bus resume.

Signed-off-by: Chunfeng Yun <chunfeng.yun-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
---
 drivers/usb/host/xhci-mtk.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/drivers/usb/host/xhci-mtk.c b/drivers/usb/host/xhci-mtk.c
index c9ab6a4..38635fb 100644
--- a/drivers/usb/host/xhci-mtk.c
+++ b/drivers/usb/host/xhci-mtk.c
@@ -696,9 +696,24 @@ static int xhci_mtk_remove(struct platform_device *dev)
 }
 
 #ifdef CONFIG_PM_SLEEP
+/*
+ * if ip sleep fail, and all clocks are disabled, access register will hang
+ * AHB bus, so stop poll roothubs to avoid regs access on bus suspend.
+ * and no need to check whether ip sleep fail or not; this will cause SPM to
+ * wakeup system immediately after system suspend complete if ip sleep
+ * fail, it is what we wanted.
+ */
 static int xhci_mtk_suspend(struct device *dev)
 {
 	struct xhci_hcd_mtk *mtk = dev_get_drvdata(dev);
+	struct usb_hcd *hcd = mtk->hcd;
+	struct xhci_hcd *xhci = hcd_to_xhci(hcd);
+
+	xhci_dbg(xhci, "%s: stop port polling\n", __func__);
+	clear_bit(HCD_FLAG_POLL_RH, &hcd->flags);
+	del_timer_sync(&hcd->rh_timer);
+	clear_bit(HCD_FLAG_POLL_RH, &xhci->shared_hcd->flags);
+	del_timer_sync(&xhci->shared_hcd->rh_timer);
 
 	xhci_mtk_host_disable(mtk);
 	xhci_mtk_phy_power_off(mtk);
@@ -710,11 +725,19 @@ static int xhci_mtk_suspend(struct device *dev)
 static int xhci_mtk_resume(struct device *dev)
 {
 	struct xhci_hcd_mtk *mtk = dev_get_drvdata(dev);
+	struct usb_hcd *hcd = mtk->hcd;
+	struct xhci_hcd *xhci = hcd_to_xhci(hcd);
 
 	usb_wakeup_disable(mtk);
 	xhci_mtk_clks_enable(mtk);
 	xhci_mtk_phy_power_on(mtk);
 	xhci_mtk_host_enable(mtk);
+
+	xhci_dbg(xhci, "%s: restart port polling\n", __func__);
+	set_bit(HCD_FLAG_POLL_RH, &hcd->flags);
+	usb_hcd_poll_rh_status(hcd);
+	set_bit(HCD_FLAG_POLL_RH, &xhci->shared_hcd->flags);
+	usb_hcd_poll_rh_status(xhci->shared_hcd);
 	return 0;
 }
 
-- 
1.8.1.1.dirty

Re: [PATCH] usb: xhci-mtk: fix AHB bus hang up caused by roothubs polling

[Sergei Shtylyov]

Hello.

    Sorry for the grammar nitpicking but since it's in the comments, I felt 
the necessity to comment.

On 12/4/2015 5:40 AM, Chunfeng Yun wrote:

> when ip fail to enter sleep mode, register access protection will

    Fails.

> be disabed, at the same time if all clocks are disabled, access
      ^^^^^^^ disabled

> register will hang up AHB bus.
> the common case causes ip sleep fail is that after all ports enter

    Failure.

> U3 but before ip enters sleep mode, a port receives a resume
> signal('K'). this will happens when such as clicks mouse to try to
> do remote wakeup to stop system enter suspend.

    Wake up.

> so stop polling roothubs to avoid access xHCI register on bus

    Root hubs. Accessing.

> suspend, and restart it when bus resume.

    Resumes. Or "is resumed", maybe?

> Signed-off-by: Chunfeng Yun <chunfeng.yun-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
> ---
>   drivers/usb/host/xhci-mtk.c | 23 +++++++++++++++++++++++
>   1 file changed, 23 insertions(+)
>
> diff --git a/drivers/usb/host/xhci-mtk.c b/drivers/usb/host/xhci-mtk.c
> index c9ab6a4..38635fb 100644
> --- a/drivers/usb/host/xhci-mtk.c
> +++ b/drivers/usb/host/xhci-mtk.c
> @@ -696,9 +696,24 @@ static int xhci_mtk_remove(struct platform_device *dev)
>   }
>
>   #ifdef CONFIG_PM_SLEEP
> +/*
> + * if ip sleep fail, and all clocks are disabled, access register will hang

    Fails.

> + * AHB bus, so stop poll roothubs to avoid regs access on bus suspend.

    Polling.

> + * and no need to check whether ip sleep fail or not; this will cause SPM to

    Failed.

> + * wakeup system immediately after system suspend complete if ip sleep

    Wake up.

> + * fail, it is what we wanted.

    Fails.

> + */
[...]

MBR, Sergei

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] usb: xhci-mtk: fix AHB bus hang up caused by roothubs polling

[chunfeng yun]

Hi,
On Fri, 2015-12-04 at 16:24 +0300, Sergei Shtylyov wrote:
> Hello.
> 
>     Sorry for the grammar nitpicking but since it's in the comments, I felt 
> the necessity to comment.
> 
> On 12/4/2015 5:40 AM, Chunfeng Yun wrote:
> 
> > when ip fail to enter sleep mode, register access protection will
> 
>     Fails.
> 
> > be disabed, at the same time if all clocks are disabled, access
>       ^^^^^^^ disabled
> 
> > register will hang up AHB bus.
> > the common case causes ip sleep fail is that after all ports enter
> 
>     Failure.
> 
> > U3 but before ip enters sleep mode, a port receives a resume
> > signal('K'). this will happens when such as clicks mouse to try to
> > do remote wakeup to stop system enter suspend.
> 
>     Wake up.
> 
> > so stop polling roothubs to avoid access xHCI register on bus
> 
>     Root hubs. Accessing.
> 
> > suspend, and restart it when bus resume.
> 
>     Resumes. Or "is resumed", maybe?
> 
> > Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
> > ---
> >   drivers/usb/host/xhci-mtk.c | 23 +++++++++++++++++++++++
> >   1 file changed, 23 insertions(+)
> >
> > diff --git a/drivers/usb/host/xhci-mtk.c b/drivers/usb/host/xhci-mtk.c
> > index c9ab6a4..38635fb 100644
> > --- a/drivers/usb/host/xhci-mtk.c
> > +++ b/drivers/usb/host/xhci-mtk.c
> > @@ -696,9 +696,24 @@ static int xhci_mtk_remove(struct platform_device *dev)
> >   }
> >
> >   #ifdef CONFIG_PM_SLEEP
> > +/*
> > + * if ip sleep fail, and all clocks are disabled, access register will hang
> 
>     Fails.
> 
> > + * AHB bus, so stop poll roothubs to avoid regs access on bus suspend.
> 
>     Polling.
> 
> > + * and no need to check whether ip sleep fail or not; this will cause SPM to
> 
>     Failed.
> 
> > + * wakeup system immediately after system suspend complete if ip sleep
> 
>     Wake up.
> 
> > + * fail, it is what we wanted.
> 
>     Fails.
> 

I do need to pay more attention to the grammar from now on.

Thanks a lot

> > + */
> [...]
> 
> MBR, Sergei
>