unhandled interrupt from otg port

unhandled interrupt from otg port

[Marc Dietrich]

Hi,

I sometimes see unhandled irq from the otg port on my AC100 (kernel 3.8.0-rc7).
This happens randomly when the a cable is pluged into the mini usb port. Seems
the port generates an interrupt on probe which isn't handled.

Any ideas?

Marc

[    1.983545] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
[    1.996613] unable to find transceiver of type USB2 PHY
[    2.004693] tegra-ehci tegra-ehci.0: Tegra EHCI Host Controller
[    2.013466] tegra-ehci tegra-ehci.0: new USB bus registered, assigned bus number 1
[    2.167205] irq 52: nobody cared (try booting with the "irqpoll" option)
[    2.176740] [<c001b46c>] (unwind_backtrace+0x0/0xf8) from [<c00ad640>] (__report_bad_irq+0x28/0xc4)
[    2.188686] [<c00ad640>] (__report_bad_irq+0x28/0xc4) from [<c00ad904>] (note_interrupt+0x13c/0x190)
[    2.200753] [<c00ad904>] (note_interrupt+0x13c/0x190) from [<c00ab6fc>] (handle_irq_event_percpu+0xcc/0x258)
[    2.213559] [<c00ab6fc>] (handle_irq_event_percpu+0xcc/0x258) from [<c00ab8d8>] (handle_irq_event+0x50/0x70)
[    2.226425] [<c00ab8d8>] (handle_irq_event+0x50/0x70) from [<c00ae554>] (handle_fasteoi_irq+0x7c/0x154)
[    2.238865] [<c00ae554>] (handle_fasteoi_irq+0x7c/0x154) from [<c00ab060>] (generic_handle_irq+0x20/0x30)
[    2.251521] [<c00ab060>] (generic_handle_irq+0x20/0x30) from [<c0013f24>] (handle_IRQ+0x38/0x90)
[    2.263442] [<c0013f24>] (handle_IRQ+0x38/0x90) from [<c0008508>] (gic_handle_irq+0x28/0x5c)
[    2.275071] [<c0008508>] (gic_handle_irq+0x28/0x5c) from [<c05d8c80>] (__irq_svc+0x40/0x50)
[    2.286650] Exception stack(0xdf865cb0 to 0xdf865cf8)
[    2.294904] 5ca0:                                     c09391c0 3f75ff7c 00000000 00000000
[    2.306340] 5cc0: 00000002 0000001d df864000 fe000100 00000034 df8077b0 60000113 c08ae6a0
[    2.317817] 5ce0: 002617f4 df865cf8 c00393cc c0038ef8 20000113 ffffffff
[    2.327777] [<c05d8c80>] (__irq_svc+0x40/0x50) from [<c0038ef8>] (__do_softirq+0x80/0x254)
[    2.339470] [<c0038ef8>] (__do_softirq+0x80/0x254) from [<c00393cc>] (irq_exit+0x80/0x88)
[    2.351074] [<c00393cc>] (irq_exit+0x80/0x88) from [<c0013f28>] (handle_IRQ+0x3c/0x90)
[    2.362403] [<c0013f28>] (handle_IRQ+0x3c/0x90) from [<c0008508>] (gic_handle_irq+0x28/0x5c)
[    2.374270] [<c0008508>] (gic_handle_irq+0x28/0x5c) from [<c05d8c80>] (__irq_svc+0x40/0x50)
[    2.386056] Exception stack(0xdf865d70 to 0xdf865db8)
[    2.394520] 5d60:                                     df8077d0 60000113 00000000 00070007
[    2.406178] 5d80: df807780 dd676d40 df8077d0 c040d27c 00000034 df8077b0 60000113 c08ae6a0
[    2.417840] 5da0: 00000000 df865db8 c00acc38 c05d8718 60000113 ffffffff
[    2.427932] [<c05d8c80>] (__irq_svc+0x40/0x50) from [<c05d8718>] (_raw_spin_unlock_irqrestore+0x28/0x2c)
[    2.440981] [<c05d8718>] (_raw_spin_unlock_irqrestore+0x28/0x2c) from [<c00acc38>] (__setup_irq+0x1c4/0x40c)
[    2.454462] [<c00acc38>] (__setup_irq+0x1c4/0x40c) from [<c00acf34>] (request_threaded_irq+0xb4/0x144)
[    2.467460] [<c00acf34>] (request_threaded_irq+0xb4/0x144) from [<c040d1bc>] (usb_hcd_request_irqs+0x68/0x128)
[    2.481219] [<c040d1bc>] (usb_hcd_request_irqs+0x68/0x128) from [<c040ec60>] (usb_add_hcd.part.15+0x2c0/0x310)
[    2.495034] [<c040ec60>] (usb_add_hcd.part.15+0x2c0/0x310) from [<c0420698>] (tegra_ehci_probe+0x340/0x4fc)
[    2.508638] [<c0420698>] (tegra_ehci_probe+0x340/0x4fc) from [<c0373bd4>] (platform_drv_probe+0x18/0x1c)
[    2.522034] [<c0373bd4>] (platform_drv_probe+0x18/0x1c) from [<c0372838>] (really_probe+0x60/0x1e0)
[    2.535012] [<c0372838>] (really_probe+0x60/0x1e0) from [<c0372ab4>] (driver_probe_device+0x30/0x48)
[    2.548106] [<c0372ab4>] (driver_probe_device+0x30/0x48) from [<c0372b58>] (__driver_attach+0x8c/0x90)
[    2.561401] [<c0372b58>] (__driver_attach+0x8c/0x90) from [<c037121c>] (bus_for_each_dev+0x50/0x7c)
[    2.574458] [<c037121c>] (bus_for_each_dev+0x50/0x7c) from [<c0372158>] (bus_add_driver+0x174/0x240)
[    2.587638] [<c0372158>] (bus_add_driver+0x174/0x240) from [<c0373058>] (driver_register+0x78/0x144)
[    2.600825] [<c0373058>] (driver_register+0x78/0x144) from [<c087d600>] (ehci_hcd_init+0x58/0x88)
[    2.613735] [<c087d600>] (ehci_hcd_init+0x58/0x88) from [<c00085fc>] (do_one_initcall+0x34/0x174)
[    2.626634] [<c00085fc>] (do_one_initcall+0x34/0x174) from [<c085a90c>] (do_basic_setup+0x94/0xc8)
[    2.639624] [<c085a90c>] (do_basic_setup+0x94/0xc8) from [<c085aa30>] (kernel_init_freeable+0xf0/0x18c)
[    2.653064] [<c085aa30>] (kernel_init_freeable+0xf0/0x18c) from [<c05c01f4>] (kernel_init+0x8/0xe4)
[    2.666146] [<c05c01f4>] (kernel_init+0x8/0xe4) from [<c0012fd8>] (ret_from_fork+0x14/0x3c)
[    2.678515] handlers:
[    2.684728] [<c040d27c>] usb_hcd_irq
[    2.692204] Disabling IRQ #52

Re: unhandled interrupt from otg port

[Stephen Warren]

On 02/17/2013 04:48 AM, Marc Dietrich wrote:
> Hi,
> 
> I sometimes see unhandled irq from the otg port on my AC100 (kernel 3.8.0-rc7).
> This happens randomly when the a cable is pluged into the mini usb port. Seems
> the port generates an interrupt on probe which isn't handled.

Perhaps Venu can explain this; he is our USB expert.

> Any ideas?

I'm slightly confused by your description though; I assume this problem
happens when you boot with a USB cable already plugged in, not when you
first boot the system, then plug in a cable? Is the cable connected to a
host PC, or to a device?

Re: unhandled interrupt from otg port

[Marc Dietrich]

Stephen,

On Tuesday 19 February 2013 10:03:47 Stephen Warren wrote:
> On 02/17/2013 04:48 AM, Marc Dietrich wrote:
> > Hi,
> > 
> > I sometimes see unhandled irq from the otg port on my AC100 (kernel 3.8.0-rc7).
> > This happens randomly when the a cable is pluged into the mini usb port. Seems
> > the port generates an interrupt on probe which isn't handled.
> 
> Perhaps Venu can explain this; he is our USB expert.
> 
> > Any ideas?
> 
> I'm slightly confused by your description though; I assume this problem
> happens when you boot with a USB cable already plugged in, not when you
> first boot the system, then plug in a cable? Is the cable connected to a
> host PC, or to a device?

the cable is connected to a host PC. It is not shown when connecting after the
system has booted. Unfortunately, I cannot generate the error reproducible, e.g.
I just tried 10 times and nothing was shown. Maybe just forget it for now if
no one has an obvious idea. I'll try to find some reproducable situation and will
report back.

Thanks

Marc

Re: unhandled interrupt from otg port

[Marc Dietrich]

On Tuesday 19 February 2013 23:12:53 Marc Dietrich wrote:
> Stephen,
> 
> On Tuesday 19 February 2013 10:03:47 Stephen Warren wrote:
> > On 02/17/2013 04:48 AM, Marc Dietrich wrote:
> > > Hi,
> > > 
> > > I sometimes see unhandled irq from the otg port on my AC100 (kernel 3.8.0-rc7).
> > > This happens randomly when the a cable is pluged into the mini usb port. Seems
> > > the port generates an interrupt on probe which isn't handled.
> > 
> > Perhaps Venu can explain this; he is our USB expert.
> > 
> > > Any ideas?
> > 
> > I'm slightly confused by your description though; I assume this problem
> > happens when you boot with a USB cable already plugged in, not when you
> > first boot the system, then plug in a cable? Is the cable connected to a
> > host PC, or to a device?
> 
> the cable is connected to a host PC. It is not shown when connecting after the
> system has booted. Unfortunately, I cannot generate the error reproducible, e.g.
> I just tried 10 times and nothing was shown. Maybe just forget it for now if
> no one has an obvious idea. I'll try to find some reproducable situation and will
> report back.

arr, just found it: it happens when you load an uboot image via tegrarcm first - so this
could be a kernel or uboot problem.

Marc

Re: unhandled interrupt from otg port

[Stephen Warren]

On 02/19/2013 03:17 PM, Marc Dietrich wrote:
> On Tuesday 19 February 2013 23:12:53 Marc Dietrich wrote:
>> Stephen,
>>
>> On Tuesday 19 February 2013 10:03:47 Stephen Warren wrote:
>>> On 02/17/2013 04:48 AM, Marc Dietrich wrote:
>>>> Hi,
>>>>
>>>> I sometimes see unhandled irq from the otg port on my AC100 (kernel 3.8.0-rc7).
>>>> This happens randomly when the a cable is pluged into the mini usb port. Seems
>>>> the port generates an interrupt on probe which isn't handled.
>>>
>>> Perhaps Venu can explain this; he is our USB expert.
>>>
>>>> Any ideas?
>>>
>>> I'm slightly confused by your description though; I assume this problem
>>> happens when you boot with a USB cable already plugged in, not when you
>>> first boot the system, then plug in a cable? Is the cable connected to a
>>> host PC, or to a device?
>>
>> the cable is connected to a host PC. It is not shown when connecting after the
>> system has booted. Unfortunately, I cannot generate the error reproducible, e.g.
>> I just tried 10 times and nothing was shown. Maybe just forget it for now if
>> no one has an obvious idea. I'll try to find some reproducable situation and will
>> report back.
> 
> arr, just found it: it happens when you load an uboot image via tegrarcm first - so this
> could be a kernel or uboot problem.

Ah, so that means that the Tegra boot ROM already initialized the USB
port in device mode to receive U-Boot. I assume you're not touching USB
in U-Boot at all, but loading the kernel/... from SD/eMMC. Then, the
kernel initializes/uses USB in host mode (since that's all we support so
far), and then you see the issue.

I reproduced this on Harmony. I'll file a bug and assign it to Venu.

Re: unhandled interrupt from otg port

[Marc Dietrich]

Am Dienstag, 19. Februar 2013, 15:24:44 schrieb Stephen Warren:
> On 02/19/2013 03:17 PM, Marc Dietrich wrote:
> > On Tuesday 19 February 2013 23:12:53 Marc Dietrich wrote:
> >> Stephen,
> >> 
> >> On Tuesday 19 February 2013 10:03:47 Stephen Warren wrote:
> >>> On 02/17/2013 04:48 AM, Marc Dietrich wrote:
> >>>> Hi,
> >>>> 
> >>>> I sometimes see unhandled irq from the otg port on my AC100 (kernel
> >>>> 3.8.0-rc7). This happens randomly when the a cable is pluged into the
> >>>> mini usb port. Seems the port generates an interrupt on probe which
> >>>> isn't handled.
> >>> 
> >>> Perhaps Venu can explain this; he is our USB expert.
> >>> 
> >>>> Any ideas?
> >>> 
> >>> I'm slightly confused by your description though; I assume this problem
> >>> happens when you boot with a USB cable already plugged in, not when you
> >>> first boot the system, then plug in a cable? Is the cable connected to a
> >>> host PC, or to a device?
> >> 
> >> the cable is connected to a host PC. It is not shown when connecting
> >> after the system has booted. Unfortunately, I cannot generate the error
> >> reproducible, e.g. I just tried 10 times and nothing was shown. Maybe
> >> just forget it for now if no one has an obvious idea. I'll try to find
> >> some reproducable situation and will report back.
> > 
> > arr, just found it: it happens when you load an uboot image via tegrarcm
> > first - so this could be a kernel or uboot problem.
> 
> Ah, so that means that the Tegra boot ROM already initialized the USB
> port in device mode to receive U-Boot. I assume you're not touching USB
> in U-Boot at all, but loading the kernel/... from SD/eMMC. Then, the
> kernel initializes/uses USB in host mode (since that's all we support so
> far), and then you see the issue.

yes, sounds reasonable.

> I reproduced this on Harmony. I'll file a bug and assign it to Venu.

Thanks!

Marc

RE: unhandled interrupt from otg port

[Venu Byravarasu]

> -----Original Message-----
> From: Marc Dietrich [mailto:marvin24-Mmb7MZpHnFY@public.gmane.org]
> Sent: Wednesday, February 20, 2013 2:21 PM
> To: Stephen Warren
> Cc: linux-tegra-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; Venu Byravarasu
> Subject: Re: unhandled interrupt from otg port
> 
> Am Dienstag, 19. Februar 2013, 15:24:44 schrieb Stephen Warren:
> > On 02/19/2013 03:17 PM, Marc Dietrich wrote:
> > > On Tuesday 19 February 2013 23:12:53 Marc Dietrich wrote:
> > >> Stephen,
> > >>
> > >> On Tuesday 19 February 2013 10:03:47 Stephen Warren wrote:
> > >>> On 02/17/2013 04:48 AM, Marc Dietrich wrote:
> > >>>> Hi,
> > >>>>
> > >>>> I sometimes see unhandled irq from the otg port on my AC100 (kernel
> > >>>> 3.8.0-rc7). This happens randomly when the a cable is pluged into the
> > >>>> mini usb port. Seems the port generates an interrupt on probe which
> > >>>> isn't handled.
> > >>>
> > >>> Perhaps Venu can explain this; he is our USB expert.
> > >>>
> > >>>> Any ideas?
> > >>>
> > >>> I'm slightly confused by your description though; I assume this problem
> > >>> happens when you boot with a USB cable already plugged in, not when
> you
> > >>> first boot the system, then plug in a cable? Is the cable connected to a
> > >>> host PC, or to a device?
> > >>
> > >> the cable is connected to a host PC. It is not shown when connecting
> > >> after the system has booted. Unfortunately, I cannot generate the error
> > >> reproducible, e.g. I just tried 10 times and nothing was shown. Maybe
> > >> just forget it for now if no one has an obvious idea. I'll try to find
> > >> some reproducable situation and will report back.
> > >
> > > arr, just found it: it happens when you load an uboot image via tegrarcm
> > > first - so this could be a kernel or uboot problem.
> >
> > Ah, so that means that the Tegra boot ROM already initialized the USB
> > port in device mode to receive U-Boot. I assume you're not touching USB
> > in U-Boot at all, but loading the kernel/... from SD/eMMC. Then, the
> > kernel initializes/uses USB in host mode (since that's all we support so
> > far), and then you see the issue.
> 
> yes, sounds reasonable.
> 
> > I reproduced this on Harmony. I'll file a bug and assign it to Venu.
> 

Reproduced the issue on Ventana.
Debug is in progress.

Thanks,
Venu
 
> Thanks!
> 
> Marc

RE: unhandled interrupt from otg port

[Venu Byravarasu]

> -----Original Message-----
> From: Venu Byravarasu
> Sent: Wednesday, February 20, 2013 2:30 PM
> To: 'Marc Dietrich'; Stephen Warren
> Cc: linux-tegra-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> Subject: RE: unhandled interrupt from otg port
> 
> 
> > -----Original Message-----
> > From: Marc Dietrich [mailto:marvin24-Mmb7MZpHnFY@public.gmane.org]
> > Sent: Wednesday, February 20, 2013 2:21 PM
> > To: Stephen Warren
> > Cc: linux-tegra-u79uwXL29TY76Z2rM5mHXA@public.gmane.org; Venu Byravarasu
> > Subject: Re: unhandled interrupt from otg port
> >
> > Am Dienstag, 19. Februar 2013, 15:24:44 schrieb Stephen Warren:
> > > On 02/19/2013 03:17 PM, Marc Dietrich wrote:
> > > > On Tuesday 19 February 2013 23:12:53 Marc Dietrich wrote:
> > > >> Stephen,
> > > >>
> > > >> On Tuesday 19 February 2013 10:03:47 Stephen Warren wrote:
> > > >>> On 02/17/2013 04:48 AM, Marc Dietrich wrote:
> > > >>>> Hi,
> > > >>>>
> > > >>>> I sometimes see unhandled irq from the otg port on my AC100
> (kernel
> > > >>>> 3.8.0-rc7). This happens randomly when the a cable is pluged into
> the
> > > >>>> mini usb port. Seems the port generates an interrupt on probe
> which
> > > >>>> isn't handled.
> > > >>>
> > > >>> Perhaps Venu can explain this; he is our USB expert.
> > > >>>
> > > >>>> Any ideas?
> > > >>>
> > > >>> I'm slightly confused by your description though; I assume this
> problem
> > > >>> happens when you boot with a USB cable already plugged in, not
> when
> > you
> > > >>> first boot the system, then plug in a cable? Is the cable connected to a
> > > >>> host PC, or to a device?
> > > >>
> > > >> the cable is connected to a host PC. It is not shown when connecting
> > > >> after the system has booted. Unfortunately, I cannot generate the
> error
> > > >> reproducible, e.g. I just tried 10 times and nothing was shown. Maybe
> > > >> just forget it for now if no one has an obvious idea. I'll try to find
> > > >> some reproducable situation and will report back.
> > > >
> > > > arr, just found it: it happens when you load an uboot image via
> tegrarcm
> > > > first - so this could be a kernel or uboot problem.
> > >
> > > Ah, so that means that the Tegra boot ROM already initialized the USB
> > > port in device mode to receive U-Boot. I assume you're not touching USB
> > > in U-Boot at all, but loading the kernel/... from SD/eMMC. Then, the
> > > kernel initializes/uses USB in host mode (since that's all we support so
> > > far), and then you see the issue.
> >
> > yes, sounds reasonable.
> >
> > > I reproduced this on Harmony. I'll file a bug and assign it to Venu.
> >
> 
> Reproduced the issue on Ventana.
> Debug is in progress.

Fix is pushed for external review with http://marc.info/?l=linux-usb&m=136203353205291&w=2  

> 
> Thanks,
> Venu
> 
> > Thanks!
> >
> > Marc