The dwc3->gadget_driver is not initialized during the dwc3 probe
process. This leads to a warning when the runtime power management (PM)
attempts to suspend the gadget using dwc3_gadget_suspend().
This patch adds a check to prevent the warning.
Cc: [email protected]
Fixes: 61a348857e86 ("usb: dwc3: gadget: Fix NULL pointer dereference in dwc3_gadget_suspend")
Signed-off-by: Ray Chi <[email protected]>
---
drivers/usb/dwc3/gadget.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index 28f49400f3e8..de987cffe1ec 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -4708,6 +4708,9 @@ int dwc3_gadget_suspend(struct dwc3 *dwc)
unsigned long flags;
int ret;
+ if (!dwc->gadget_driver)
+ return 0;
+
ret = dwc3_gadget_soft_disconnect(dwc);
if (ret)
goto err;
--
2.44.0.rc0.258.g7320e95886-goog
On Tue, Feb 20, 2024 at 4:12 PM Ray Chi <[email protected]> wrote:
>
> The dwc3->gadget_driver is not initialized during the dwc3 probe
> process. This leads to a warning when the runtime power management (PM)
> attempts to suspend the gadget using dwc3_gadget_suspend().
>
> This patch adds a check to prevent the warning.
>
> Cc: [email protected]
> Fixes: 61a348857e86 ("usb: dwc3: gadget: Fix NULL pointer dereference in dwc3_gadget_suspend")
> Signed-off-by: Ray Chi <[email protected]>
> ---
> drivers/usb/dwc3/gadget.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> index 28f49400f3e8..de987cffe1ec 100644
> --- a/drivers/usb/dwc3/gadget.c
> +++ b/drivers/usb/dwc3/gadget.c
> @@ -4708,6 +4708,9 @@ int dwc3_gadget_suspend(struct dwc3 *dwc)
> unsigned long flags;
> int ret;
>
> + if (!dwc->gadget_driver)
> + return 0;
> +
> ret = dwc3_gadget_soft_disconnect(dwc);
> if (ret)
> goto err;
> --
> 2.44.0.rc0.258.g7320e95886-goog
>
Attached the warning message.
02-20 06:01:24.227725 root 86 86 W Call trace:
02-20 06:01:24.227730 root 86 86 W :
dwc3_ep0_out_start+0xec/0x22c
02-20 06:01:24.227738 root 86 86 W :
dwc3_ep0_stall_and_restart+0xb8/0xd4
02-20 06:01:24.227747 root 86 86 W :
dwc3_gadget_soft_disconnect+0x1a8/0x284
02-20 06:01:24.227756 root 86 86 W :
dwc3_gadget_suspend+0x18/0xe4
02-20 06:01:24.227764 root 86 86 W :
dwc3_suspend_common+0x5c/0x320
02-20 06:01:24.227774 root 86 86 W :
dwc3_runtime_suspend+0x30/0x4c
02-20 06:01:24.227783 root 86 86 W :
pm_generic_runtime_suspend+0x3c/0x54
02-20 06:01:24.227793 root 86 86 W : __rpm_callback+0xfc/0x76c
02-20 06:01:24.227802 root 86 86 W : rpm_suspend+0x534/0xd78
02-20 06:01:24.227811 root 86 86 W : dwc3_runtime_idle+0xcc/0x104
02-20 06:01:24.227819 root 86 86 W : rpm_idle+0x228/0x4e0
02-20 06:01:24.227827 root 86 86 W : update_autosuspend+0x50/0xb8
02-20 06:01:24.227836 root 86 86 W :
__pm_runtime_use_autosuspend+0x50/0x6c
On Tue, Feb 20, 2024 at 04:12:04PM +0800, Ray Chi wrote:
> The dwc3->gadget_driver is not initialized during the dwc3 probe
> process. This leads to a warning when the runtime power management (PM)
> attempts to suspend the gadget using dwc3_gadget_suspend().
What type of warning happens?
> This patch adds a check to prevent the warning.
>
> Cc: [email protected]
> Fixes: 61a348857e86 ("usb: dwc3: gadget: Fix NULL pointer dereference in dwc3_gadget_suspend")
> Signed-off-by: Ray Chi <[email protected]>
> ---
> drivers/usb/dwc3/gadget.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> index 28f49400f3e8..de987cffe1ec 100644
> --- a/drivers/usb/dwc3/gadget.c
> +++ b/drivers/usb/dwc3/gadget.c
> @@ -4708,6 +4708,9 @@ int dwc3_gadget_suspend(struct dwc3 *dwc)
> unsigned long flags;
> int ret;
>
> + if (!dwc->gadget_driver)
> + return 0;
> +
This directly reverts part of the commit you say this fixes, are you
SURE about this? Why?
thanks,
greg k-h
On Tue, Feb 20, 2024 at 04:20:50PM +0800, Ray Chi wrote:
> On Tue, Feb 20, 2024 at 4:12 PM Ray Chi <[email protected]> wrote:
> >
> > The dwc3->gadget_driver is not initialized during the dwc3 probe
> > process. This leads to a warning when the runtime power management (PM)
> > attempts to suspend the gadget using dwc3_gadget_suspend().
> >
> > This patch adds a check to prevent the warning.
> >
> > Cc: [email protected]
> > Fixes: 61a348857e86 ("usb: dwc3: gadget: Fix NULL pointer dereference in dwc3_gadget_suspend")
> > Signed-off-by: Ray Chi <[email protected]>
> > ---
> > drivers/usb/dwc3/gadget.c | 3 +++
> > 1 file changed, 3 insertions(+)
> >
> > diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> > index 28f49400f3e8..de987cffe1ec 100644
> > --- a/drivers/usb/dwc3/gadget.c
> > +++ b/drivers/usb/dwc3/gadget.c
> > @@ -4708,6 +4708,9 @@ int dwc3_gadget_suspend(struct dwc3 *dwc)
> > unsigned long flags;
> > int ret;
> >
> > + if (!dwc->gadget_driver)
> > + return 0;
> > +
> > ret = dwc3_gadget_soft_disconnect(dwc);
> > if (ret)
> > goto err;
> > --
> > 2.44.0.rc0.258.g7320e95886-goog
> >
>
> Attached the warning message.
>
> 02-20 06:01:24.227725 root 86 86 W Call trace:
> 02-20 06:01:24.227730 root 86 86 W :
> dwc3_ep0_out_start+0xec/0x22c
> 02-20 06:01:24.227738 root 86 86 W :
> dwc3_ep0_stall_and_restart+0xb8/0xd4
> 02-20 06:01:24.227747 root 86 86 W :
> dwc3_gadget_soft_disconnect+0x1a8/0x284
> 02-20 06:01:24.227756 root 86 86 W :
> dwc3_gadget_suspend+0x18/0xe4
> 02-20 06:01:24.227764 root 86 86 W :
> dwc3_suspend_common+0x5c/0x320
> 02-20 06:01:24.227774 root 86 86 W :
> dwc3_runtime_suspend+0x30/0x4c
> 02-20 06:01:24.227783 root 86 86 W :
> pm_generic_runtime_suspend+0x3c/0x54
> 02-20 06:01:24.227793 root 86 86 W : __rpm_callback+0xfc/0x76c
> 02-20 06:01:24.227802 root 86 86 W : rpm_suspend+0x534/0xd78
> 02-20 06:01:24.227811 root 86 86 W : dwc3_runtime_idle+0xcc/0x104
> 02-20 06:01:24.227819 root 86 86 W : rpm_idle+0x228/0x4e0
> 02-20 06:01:24.227827 root 86 86 W : update_autosuspend+0x50/0xb8
> 02-20 06:01:24.227836 root 86 86 W :
> __pm_runtime_use_autosuspend+0x50/0x6c
That's a kernel warning? Odd wrapping :(
Can you clean it up a bit and put it in the changelog? There should be
more info before the "Call trace" portion, right? What is causing the
warning exactly?
thanks,
greg k-h
On 2/20/2024 2:04 PM, Greg KH wrote:
> On Tue, Feb 20, 2024 at 04:12:04PM +0800, Ray Chi wrote:
>> The dwc3->gadget_driver is not initialized during the dwc3 probe
>> process. This leads to a warning when the runtime power management (PM)
>> attempts to suspend the gadget using dwc3_gadget_suspend().
>
> What type of warning happens?
>
>> This patch adds a check to prevent the warning.
>>
>> Cc: [email protected]
>> Fixes: 61a348857e86 ("usb: dwc3: gadget: Fix NULL pointer dereference in dwc3_gadget_suspend")
>> Signed-off-by: Ray Chi <[email protected]>
>> ---
>> drivers/usb/dwc3/gadget.c | 3 +++
>> 1 file changed, 3 insertions(+)
>>
>> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
>> index 28f49400f3e8..de987cffe1ec 100644
>> --- a/drivers/usb/dwc3/gadget.c
>> +++ b/drivers/usb/dwc3/gadget.c
>> @@ -4708,6 +4708,9 @@ int dwc3_gadget_suspend(struct dwc3 *dwc)
>> unsigned long flags;
>> int ret;
>>
>> + if (!dwc->gadget_driver)
>> + return 0;
>> +
>
> This directly reverts part of the commit you say this fixes, are you
> SURE about this? Why?
>
Hi Ray,
Thinh sent a patch recently addressing the issue in soft disconnect.
Can you check if it helps:
https://lore.kernel.org/all/e3be9b929934e0680a6f4b8f6eb11b18ae9c7e07.1708043922.git.Thinh.Nguyen@synopsys.com/
Regards,
Krishna,
Hi Krishna,
I verified the Thinh's patch and the warning could be
fixed. Thanks for the information.
Regards,
Ray
On Tue, Feb 20, 2024 at 4:40 PM Krishna Kurapati PSSNV
<[email protected]> wrote:
>
>
>
> On 2/20/2024 2:04 PM, Greg KH wrote:
> > On Tue, Feb 20, 2024 at 04:12:04PM +0800, Ray Chi wrote:
> >> The dwc3->gadget_driver is not initialized during the dwc3 probe
> >> process. This leads to a warning when the runtime power management (PM)
> >> attempts to suspend the gadget using dwc3_gadget_suspend().
> >
> > What type of warning happens?
> >
> >> This patch adds a check to prevent the warning.
> >>
> >> Cc: [email protected]
> >> Fixes: 61a348857e86 ("usb: dwc3: gadget: Fix NULL pointer dereference in dwc3_gadget_suspend")
> >> Signed-off-by: Ray Chi <[email protected]>
> >> ---
> >> drivers/usb/dwc3/gadget.c | 3 +++
> >> 1 file changed, 3 insertions(+)
> >>
> >> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> >> index 28f49400f3e8..de987cffe1ec 100644
> >> --- a/drivers/usb/dwc3/gadget.c
> >> +++ b/drivers/usb/dwc3/gadget.c
> >> @@ -4708,6 +4708,9 @@ int dwc3_gadget_suspend(struct dwc3 *dwc)
> >> unsigned long flags;
> >> int ret;
> >>
> >> + if (!dwc->gadget_driver)
> >> + return 0;
> >> +
> >
> > This directly reverts part of the commit you say this fixes, are you
> > SURE about this? Why?
> >
>
> Hi Ray,
>
> Thinh sent a patch recently addressing the issue in soft disconnect.
> Can you check if it helps:
>
> https://lore.kernel.org/all/e3be9b929934e0680a6f4b8f6eb11b18ae9c7e07.1708043922.git.Thinh.Nguyen@synopsys.com/
>
> Regards,
> Krishna,
On Tue, Feb 20, 2024 at 05:42:56PM +0800, Ray Chi wrote:
> Hi Krishna,
>
> I verified the Thinh's patch and the warning could be
> fixed. Thanks for the information.
Can you provide a tested-by for that one?
And please do not top post :(
thanks,
greg k-h
On Tue, Feb 20, 2024 at 9:56 PM Greg KH <[email protected]> wrote:
>
> On Tue, Feb 20, 2024 at 05:42:56PM +0800, Ray Chi wrote:
> > Hi Krishna,
> >
> > I verified the Thinh's patch and the warning could be
> > fixed. Thanks for the information.
>
> Can you provide a tested-by for that one?
Since the solution has been merged, do I still need to provide tested-by?
If tested-by is required, should I reply to the email thread for the
merged patch or the reported patch?
> And please do not top post :(
Thanks for the tips and your patience.
Regards,
Ray