[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [usb] Resume in OHCI
** Proprietary **
one part of the OHCI will be PCI compliant and other part will be towards USB. PCI clock can't be stopped for USB purposes. as u said, the particular bit has to be set by PCI. so clock is a must for PCI operations unless otherwise there is suspending protocol on PCI side also. then how do u detect whether that bit is set or not? probably u can seperately implement that bit in h/w(through a FF) instead of using RAM bit such that it will act as enable signal to your combinational circuit.
stopping the OHCI does mean stopping clock to most of the OHCI except the part involved in resume detection etc.
i didn't understand why clock restarting is taking 1.14ms. probably in that data sheet, clock restarting is done by s/w(i can't estimate how s/w can do it). that's why time taken is so much. h/w implementation will not take a few micro-seconds also.
....vikas.
>>> ak56_n_ak47@rediffmail.com 12/27/02 05:20PM >>>
Thanks for the reply sir.
I have another related question from the point U raised about the host
initiated resume. Please try to answer it also to the best of UR ability.
Host initiated resume will start when HCFS field is written by HCD
(driver) to value corresponding to USBRESUME. This write is possible if
host bus clock (e.g. PCI) is running (which will ofcourse be). But if OHCI
clock has stopped how dual ff synchronizers work?
Does stopping the OHCI clock means stopping clock to most of the OHCI
except the part involved in resume detection etc?
Moreover I was going through datasheet of one HC which stated that
clock can be restarted in 1.14ms. (I am using this as reference because
I am not well versed with the analog circuitry involved with
stopping/starting clock) If remote-resume is detected the clock will
start in 1.14ms , but HC has to reflect the resume signalling in 900us.
Does this again mean that part of OHCI responsible for resume signalling
doesnot have its clock stopped?
I know I am asking implementation specific things. But please answer as
much as U can.
Thanx in advance
amit kumar
----- Original Message -----
From: "Vikas T Rao " <vikasraot@m... >
To: <usb@o... >
Date: Fri, 27 Dec 2002 10:43:19 +0530
Subject: Re: [usb] Resume in OHCI
>
>
> ** Proprietary **
>
> though i am not working on host controllers, i will try to explain
> ur second question. whatever u are asking is all implementation
> specific. i think spec will not say, u have to do like this only
> for acheiving such and such things.
>
> for ur second question, probably u can implement like this.
>
> it's better that OHCI goes into power saving mode. let the internal
> clock be switched off. any remote wake-up from device is an
> asynchonous one. u can always detect a 'K' state through
> combinational circuit(aux power is required for it definitely).
> then that circuit should enable the clock generation circuit and
> further processing can be done once clock is available. i suppose
> analog transceiver will convert between the voltage levels, match
> impedances and not much other things. so clock will be generated
> internal to OHCI taking external clock from oscillator/crystal or
> micro-processor.
>
> now one more question arises. if the OHCI wants to wake-up
> irrespective of the device's remote wake-up signal, how it should
> detect it? the same reasoning applies in this case also. probably
> OHCI wants to wake-up due to some interrupt from system itself.
> this interrupt can be detected through combinational circuit as
> well.
>
> hope i am correct and explained you according to ur needs.
>
> ...vikas.
>
> >>> ak56_n_ak47@r... 12/27/02 12:46PM
> >>>
> Hi
>
> I am a sort of newbie involved in OHCI hardware development.
> I have some doubt about OHCI.
>
> First:
> In EHCI spec, there are two power wells (auxilliary and core). No
> such
> power wells are defined in OHCI spec. Why? Any specific reason they
> are needed in EHCI only or is it because OHCI spec are written
> first and
> this is improvement in EHCI?
>
> Second:
> When entire USB bus is suspended, should OHCI go to power saving
> mode? At this time does clock gets turned off?
> I guess answers to above questions is yes (correct me if i am
> wrong).
> But main question is how does OHCI detect remote wakeup? Through
a
> combinational circuit? Is it the function of OHCI to restart the
> clock or
> the analog transceiver does it? I think the analog transceiver
> provides
> the clock to OHCI (correct me if I am wrong). Should I place the
> remote
> wakeup detecting circuit in OHCI with aux power well for it?
>
> Any answers on this will help.
>
> Thanx in advance
> ak
>
--
To unsubscribe from usb mailing list please visit http://www.opencores.org/mailinglists.shtml
--
To unsubscribe from usb mailing list please visit http://www.opencores.org/mailinglists.shtml