ASPEED EHCI error: "controller refused to start: -110"

ASPEED EHCI error: "controller refused to start: -110"

[Tao Ren]

Hi,

I started to see below EHCI probe failure on my AST2500 and AST2600 BMC
platforms since commit 280a9045bb18 ("ehci: fix EHCI host controller
initialization sequence"):

> ehci-platform 1e6a3000.usb: USB 2.0, controller refused to start: -110

The error happens because USBSTS.HCHalted stays at 1. But if I ignore the
error at probe time, the downstream USB devices can still be enumerated
without problems (even though USBSTS.HCHalted is still 1).

My questions are: why USBSTS.HCHalted stays at 1 on ASPEED platforms? Do
we need extra configuration at EHCI probe time for ASPEED SoC?

CC Ryan @ASPEED for more comments.


Cheers,

Tao

RE: ASPEED EHCI error: "controller refused to start: -110"

[Ryan Chen]

Hello,
	ASPEED EHCI design for HALT bit effect is depend on Asynchronous/Periodic Schedule enable or not.
	If Asynchronous/Periodic Schedule not enable, but run bit is setting, Halt still = 1.

Ryan Chen

Tel : 886-3-5751185 ext:8857

> -----Original Message-----
> From: Tao Ren <rentao.bupt@gmail.com>
> Sent: Wednesday, August 25, 2021 2:24 PM
> To: openbmc@lists.ozlabs.org
> Cc: joel@jms.id.au; andrew@aj.id.au; Ryan Chen
> <ryan_chen@aspeedtech.com>
> Subject: ASPEED EHCI error: "controller refused to start: -110"
> 
> Hi,
> 
> I started to see below EHCI probe failure on my AST2500 and AST2600 BMC
> platforms since commit 280a9045bb18 ("ehci: fix EHCI host controller
> initialization sequence"):
> 
> > ehci-platform 1e6a3000.usb: USB 2.0, controller refused to start: -110
> 
> The error happens because USBSTS.HCHalted stays at 1. But if I ignore the
> error at probe time, the downstream USB devices can still be enumerated
> without problems (even though USBSTS.HCHalted is still 1).
> 
> My questions are: why USBSTS.HCHalted stays at 1 on ASPEED platforms? Do
> we need extra configuration at EHCI probe time for ASPEED SoC?
> 
> CC Ryan @ASPEED for more comments.
> 
> 
> Cheers,
> 
> Tao

Re: ASPEED EHCI error: "controller refused to start: -110"

[Tao Ren]

Hi Ryan,

On Wed, Aug 25, 2021 at 07:48:57AM +0000, Ryan Chen wrote:
> Hello,
> 	ASPEED EHCI design for HALT bit effect is depend on Asynchronous/Periodic Schedule enable or not.
> 	If Asynchronous/Periodic Schedule not enable, but run bit is setting, Halt still = 1.
> 
> Ryan Chen

Thanks for the inputs.

What is your recommendation? Enabling Async/Periodic schedule to make
USBSTS.HCHalted handshake happy? Or ignoring HCHalted check for ASPEED
platforms? Can you please send out a patch to fix the failure? The EHCI
driver stops working on ASPEED platforms since v5.10.


Cheers,

Tao
> 
> Tel : 886-3-5751185 ext:8857
> 
> > -----Original Message-----
> > From: Tao Ren <rentao.bupt@gmail.com>
> > Sent: Wednesday, August 25, 2021 2:24 PM
> > To: openbmc@lists.ozlabs.org
> > Cc: joel@jms.id.au; andrew@aj.id.au; Ryan Chen
> > <ryan_chen@aspeedtech.com>
> > Subject: ASPEED EHCI error: "controller refused to start: -110"
> > 
> > Hi,
> > 
> > I started to see below EHCI probe failure on my AST2500 and AST2600 BMC
> > platforms since commit 280a9045bb18 ("ehci: fix EHCI host controller
> > initialization sequence"):
> > 
> > > ehci-platform 1e6a3000.usb: USB 2.0, controller refused to start: -110
> > 
> > The error happens because USBSTS.HCHalted stays at 1. But if I ignore the
> > error at probe time, the downstream USB devices can still be enumerated
> > without problems (even though USBSTS.HCHalted is still 1).
> > 
> > My questions are: why USBSTS.HCHalted stays at 1 on ASPEED platforms? Do
> > we need extra configuration at EHCI probe time for ASPEED SoC?
> > 
> > CC Ryan @ASPEED for more comments.
> > 
> > 
> > Cheers,
> > 
> > Tao

RE: ASPEED EHCI error: "controller refused to start: -110"

[Ryan Chen]

> -----Original Message-----
> From: Tao Ren <rentao.bupt@gmail.com>
> Sent: Thursday, August 26, 2021 9:53 AM
> To: Ryan Chen <ryan_chen@aspeedtech.com>
> Cc: openbmc@lists.ozlabs.org; joel@jms.id.au; andrew@aj.id.au; Neal Liu
> <neal_liu@aspeedtech.com>; Joppa Tsao <joppa_tsao@aspeedtech.com>
> Subject: Re: ASPEED EHCI error: "controller refused to start: -110"
> 
> Hi Ryan,
> 
> On Wed, Aug 25, 2021 at 07:48:57AM +0000, Ryan Chen wrote:
> > Hello,
> > 	ASPEED EHCI design for HALT bit effect is depend on
> Asynchronous/Periodic Schedule enable or not.
> > 	If Asynchronous/Periodic Schedule not enable, but run bit is setting, Halt
> still = 1.
> >
> > Ryan Chen
> 
> Thanks for the inputs.
> 
> What is your recommendation? Enabling Async/Periodic schedule to make
> USBSTS.HCHalted handshake happy? Or ignoring HCHalted check for ASPEED
> platforms? Can you please send out a patch to fix the failure? The EHCI driver
> stops working on ASPEED platforms since v5.10.
>
I think it should ignoring the HCHalted check for ASPEED platform.
I need think about how to patch for fix this. 
 
> 
> Cheers,
> 
> Tao
> >
> > Tel : 886-3-5751185 ext:8857
> >
> > > -----Original Message-----
> > > From: Tao Ren <rentao.bupt@gmail.com>
> > > Sent: Wednesday, August 25, 2021 2:24 PM
> > > To: openbmc@lists.ozlabs.org
> > > Cc: joel@jms.id.au; andrew@aj.id.au; Ryan Chen
> > > <ryan_chen@aspeedtech.com>
> > > Subject: ASPEED EHCI error: "controller refused to start: -110"
> > >
> > > Hi,
> > >
> > > I started to see below EHCI probe failure on my AST2500 and AST2600
> > > BMC platforms since commit 280a9045bb18 ("ehci: fix EHCI host
> > > controller initialization sequence"):
> > >
> > > > ehci-platform 1e6a3000.usb: USB 2.0, controller refused to start:
> > > > -110
> > >
> > > The error happens because USBSTS.HCHalted stays at 1. But if I
> > > ignore the error at probe time, the downstream USB devices can still
> > > be enumerated without problems (even though USBSTS.HCHalted is still 1).
> > >
> > > My questions are: why USBSTS.HCHalted stays at 1 on ASPEED
> > > platforms? Do we need extra configuration at EHCI probe time for ASPEED
> SoC?
> > >
> > > CC Ryan @ASPEED for more comments.
> > >
> > >
> > > Cheers,
> > >
> > > Tao