usb, davinci: usb 2.0 problem on an am1808 based board

[Heiko Schocher]

Hello Manjunathappa,

Manjunathappa, Prakash wrote:
> Hi Heiko,
> 
> On Mon, May 21, 2012 at 11:40:33, Heiko Schocher wrote:
>> Hello Manjunathappa,
>>
>> Manjunathappa, Prakash wrote:
>>> Hi Heiko,
>>>
>>> I do not know how putting delay helped MSC device detection.
>>> Can you please check if MUSB is coming up in "b_idle" state(by
>>> $cat /sys/devices/platform/musb-da8xx/musb-hdrc/mode)?
>>> State should move to b_peripheral on connecting gadget cable.
>> We have connected the USB0_ID pin to ground -> host only mode, so
>> I could not try this.
>>
> 
> Ok, So it is coming up in which state, "$cat /sys/devices/platform/musb-da8xx/musb-hdrc/mode"?

without the delay in da8xx_musb_interrupt() and the
+       musb->a_wait_bcon = OTG_TIME_A_WAIT_BCON;
 change:

booting Linux with USB stick attached shows:
root at armv5te:~# cat /sys/devices/platform/musb-da8xx/musb-hdrc/mode
a_host
root at armv5te:~#

booting Linux without USB stick attached:

root at armv5te:~# cat /sys/devices/platform/musb-da8xx/musb-hdrc/mode
a_wait_vrise
root at armv5te:~#

and connecting an usb stick:

root at armv5te:~# cat /sys/devices/platform/musb-da8xx/musb-hdrc/mode
a_host
root at armv5te:~#

>>> If you connect MSC device via mini-A connector, state should change to "a_host".
>>> OTG timer is responsible for above state changes, can you please check if below
>>> changes are present?
>>> diff --git a/drivers/usb/musb/da8xx.c b/drivers/usb/musb/da8xx.c index 4da7492..a1a692e 100644
>>> --- a/drivers/usb/musb/da8xx.c
>>> +++ b/drivers/usb/musb/da8xx.c
>>> @@ -76,6 +76,7 @@
>>>  #define DA8XX_INTR_TX_SHIFT    0
>>>  #define DA8XX_INTR_TX_MASK     (DA8XX_USB_TX_EP_MASK << DA8XX_INTR_TX_SHIFT)
>>>
>>> +#define A_WAIT_BCON_TIMEOUT     1100            /* in ms */
>>>  #define DA8XX_MENTOR_CORE_OFFSET 0x400
>>>
>>>  #define CFGCHIP2       IO_ADDRESS(DA8XX_SYSCFG0_BASE + DA8XX_CFGCHIP2_REG)
>>> @@ -443,6 +444,7 @@ static int da8xx_musb_init(struct musb *musb)
>>>                  rev, __raw_readl(CFGCHIP2),
>>>                  musb_readb(reg_base, DA8XX_USB_CTRL_REG));
>>>
>>> +       musb->a_wait_bcon = A_WAIT_BCON_TIMEOUT;
>>>         musb->isr = da8xx_musb_interrupt;
>>>         return 0;
>>>  fail:
>> No, this change is not in current mainline ...
>>
>> Added this patch, and I get:
>>
>> $ make -s uImage
>> drivers/usb/musb/da8xx.c: In function 'da8xx_musb_init':
>> drivers/usb/musb/da8xx.c:448:22: error: 'A_WAIT_BCON_TIMEOUT' undeclared (first use in this function)
>> drivers/usb/musb/da8xx.c:448:22: note: each undeclared identifier is reported only once for each function it appears in
>> make[3]: *** [drivers/usb/musb/da8xx.o] Error 1
>> make[2]: *** [drivers/usb/musb] Error 2
>> make[1]: *** [drivers/usb] Error 2
>> make: *** [drivers] Error 2
>> $
>>
>> Do you mean
>>
>> musb->a_wait_bcon = OTG_TIME_A_WAIT_BCON;
>>
>> ? This is setup in musb_core.c::allocate_instance() as Sergei Shtylyov
>> already commented ... nevertheless tried with this define your patch,
>> and my USB stick gets detected too...
>>
> 
> Does that mean issue got fixed or Do you still have delay in da8xx_musb_interrupt handler?

Without the delay in the da8xx_musb_interrupt() ...

bye,
Heiko
-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de