Hi,
  > The ohci-platform and ehci-platform drivers already perform their own 
root-hub and bus resets.
  I don't understand the sentence clearly.

Synopsis control VERSION: DesignWare Cores USB2.0 Host-AHB Controller, Version 
2.98a
The synopsis has the three source clocks for EHCI Host Controller 
implementation:
A, System (AHB BIU) Clock
B, PHY Clock(BUS clock)
C, UTMI PHY Clock(per port)

A clock corresponding to a reset, so the ehci control should have the same 
number reset. We are ready to put the utmi phy clock and utmi reset into the 
phy module. The ehci-platform is on the lack of a reset number.
Clock and reset detail:
Signal                 Description
ehci_hclk_i                     EHCI AHB System Clock
                                        Function: The EHCI AHB System clock 
supports 30- to 166- MHz 
                                        operation and has been tested at 
40/50/60 MHz in FPGA technology. All 
                                        signal timings are related to the 
rising edge of ehci_hclk_i. The EHCI 
                                        System Clock runs off this clock. For 
correct synchronization between the 
                                        AHB and UTMI, the minimum required AHB 
frequency is 30 MHz.
                                        This clock is valid only if "Dedicated 
AHB interface for EHCI and OHCI" is 
                                        Selected

ehci_hreset_i_n                 EHCI System Reset
                                        Function: This active-low, synchronous 
or asynchronous (depending on the
                                        configuration) reset serves as power-on 
reset and resets all the EHCI 
                                        control signals.
                                        This signal is active only  if 
"Dedicated AHB interface for EHCI and OHCI" is 
                                        selected

Phy_clk_i              Local EHCI PHY Clock
                                        Function: This free-running clock 
provides an operating frequency in the 
                                        Root Hub for clocking receive and 
transmit parallel data.

phy_rst_i_n                     Power-on Reset
                                        Function: This signal is synchronous 
(asynchronous, if ASYNC reset option 
                                        selected) to phy_clk_i, and resets all 
registers and state machines in receive 
                                        and transmit logic.

Best regards,
Pengcheng Li

> -----Original Message-----
> From: Alan Stern [mailto:st...@rowland.harvard.edu]
> Sent: Friday, March 18, 2016 1:52 AM
> To: Lipengcheng
> Cc: ba...@ti.com; chasemetzge...@gmail.com; baolu...@linux.intel.com; 
> mj...@coreos.com; kbo...@gmail.com; jun...@freescale.com;
> robert.schlabb...@gmx.net; linux-usb@vger.kernel.org
> Subject: Re: ehci and ohci reset number doubt
> 
> On Thu, 17 Mar 2016, Lipengcheng wrote:
> 
> > Hi,
> >    In the files of ohci-platform.c and ehci-platform.c, they have only a 
> > control reset. Can the files have one more controller reset?
> 
> Currently only one reset is allowed.
> 
> >    Our usb2 controller using a synopsis, need bus reset, root hub reset, 
> > utmi reset. The usb controller of reset and clock have the same
> number.
> 
> If your platform has a Synopsis (now DesignWare) USB2 controller, why not use 
> the dwc2 driver that's already in the kernel?
> 
> The ohci-platform and ehci-platform drivers already perform their own 
> root-hub and bus resets.  They do not perform a UTMI reset, but they
> do call phy_init() and phy_power_on().  Will that be good enough for you?
> 
> Alan Stern

Reply via email to