1710 and USB host

1710 and USB host

[Kevin Hilman]

Anyone else having success using the 1710 in USB host mode?  Using the latest 
git kernel, I'm getting the boot messages below.  I have a USB key plugged in 
and it's lighting up, so presumably it's getting powered properly.

Maybe the jumpers aren't set quite right, I don't quite grok the comments 
around the 'hmc_mode' value in. board-h3.c.  In any case, my SW1501 settings 
are 1=on, and the rest are off.

Thanks for any pointers,

Kevin

isp1301_omap 0-002d: chiprev 2.00, driver 24 August 2004
MUX: initialized N21_1710_GPIO14
usbmon: debugfs is not available
isp1301_omap 0-002d: A-Host sessions ok
ohci ohci: OMAP OHCI
ohci ohci: new USB bus registered, assigned bus number 1
ohci ohci: irq 38, io mem 0xfffba000
ohci ohci: init err (00000000 0000)
ohci ohci: can't start
ohci ohci: startup error -75
ohci ohci: USB bus 1 deregistered
ohci: probe of ohci failed with error -75

Re: 1710 and USB host

[David Brownell]

On Monday 01 May 2006 10:54 am, Kevin Hilman wrote:
> Anyone else having success using the 1710 in USB host mode? 

Last time I tried it was on a platform where the transceiver wasn't
wired up usably, so it was kind of pointless.  I expect that's not
what you're finding ...


> Using the latest  
> git kernel, I'm getting the boot messages below.  I have a USB key plugged in 
> and it's lighting up, so presumably it's getting powered properly.
> 
> Maybe the jumpers aren't set quite right, I don't quite grok the comments 
> around the 'hmc_mode' value in. board-h3.c.  In any case, my SW1501 settings 
> are 1=on, and the rest are off.

I have no H3, so I can't say whether any of this is sane or not.
As I recall, there are two transceivers on that board:  OTG, and non-OTG.
Ideally both should work.  I seem to recall positive reports about using
the OTG transciever.

> Thanks for any pointers,
> 
> Kevin
> 
> isp1301_omap 0-002d: chiprev 2.00, driver 24 August 2004
> MUX: initialized N21_1710_GPIO14
> usbmon: debugfs is not available
> isp1301_omap 0-002d: A-Host sessions ok
> ohci ohci: OMAP OHCI
> ohci ohci: new USB bus registered, assigned bus number 1
> ohci ohci: irq 38, io mem 0xfffba000
> ohci ohci: init err (00000000 0000)

That "init err" means the controlle wasn't responding to basic stuff.
Like maybe it wasn't clocked correctly ... you could turn on USB_DEBUG
in Kconfig and see if the register dump looks even vaguely sane (like,
it should report three ports etc).

- Dave


> ohci ohci: can't start
> ohci ohci: startup error -75
> ohci ohci: USB bus 1 deregistered
> ohci: probe of ohci failed with error -75
> 

Re: 1710 and USB host

[ml]

Hi Kevin,

On Wednesday 03 May 2006 07:11, David Brownell wrote:
> > isp1301_omap 0-002d: chiprev 2.00, driver 24 August 2004
> > MUX: initialized N21_1710_GPIO14
> > usbmon: debugfs is not available
> > isp1301_omap 0-002d: A-Host sessions ok
> > ohci ohci: OMAP OHCI
> > ohci ohci: new USB bus registered, assigned bus number 1
> > ohci ohci: irq 38, io mem 0xfffba000
> > ohci ohci: init err (00000000 0000)
>
> That "init err" means the controlle wasn't responding to basic stuff.
> Like maybe it wasn't clocked correctly ... you could turn on USB_DEBUG
> in Kconfig and see if the register dump looks even vaguely sane (like,
> it should report three ports etc).

We have a custom 1710 based board and I see exactly the same error message. 
Did you solve your problems with the "init err" on the 1710 platform? If yes, 
what was the solution. Thanks for any input here.

Best regards,
Stefan

1710 and USB Host

[Stefano Babic]

Hi,

sometimes ago someone has already reported troubles to get USB Host working on 
OMAP 1710. I have now on a custom board the same problem and the same error 
messages that were previously reported:

ohci ohci: OMAP OHCI
ohci ohci: new USB bus registered, assigned bus number 1
ohci ohci: irq 38, io mem 0xfffba000
ohci ohci: init err (00000000 0000)
ohci ohci: can't start
ohci ohci: startup error -75
ohci ohci: USB bus 1 deregistered
ohci: probe of ohci failed with error -75

I have seen that the ohci-omap module cannot access the USB registers. A dump 
of all registers reports only zeros.

After some investigation I have found a couple of problems in the function 
otg_init (arch/arm/plat-omap/usb.c). At the start of the function it is still 
possible to access the registers, at the end even the revision register 
(0xFFFBA000) is zero. And this is not possible...

I found that the behavior is related to the values of two registers 
OTG_SYSCON_1 and OTG_SYSCON_2. 
In OTG_SYSCON_2 the bit UHOST_EN is set to zero and this disables the Host 
controller. The behavior of the processor seems correct because according to 
the manual it is not possible to access the USB memory area if UHOST_EN is 
disabled.

A second problem is related to OTG_SYSCON_1. In this register the bit 
OTG_IDLE_EN is set and this produces the same effect disallowing memory 
accesses.

I changed the two registers setting UHOST_EN and disabling OTG_IDLE_EN and now 
it is working. A USB stick is detected by the controller and I can read the 
filesystem, so it seems this is the solution.

Has someone seen the same problems ? In this case, how do you fixed the 
problem ? Do you agree with my conclusions ?

It seems this is a problem related only to the OMAP 1710 and not affecting 
other processors, or it didn't ever worked before. 

Thanks,
stefano

-- 
stefano <stefano.babic@babic.homelinux.org>
GPG Key: 0x55814DDE 
Fingerprint 4E85 2A66 4CBA 497A 2A7B D3BF 5973 F216 5581 4DDE