Re: libusb usb_interrupt_read hangs under FreeBSD

["Xiaofan Chen" <xiaofanc_(_at_)_gmail_(_dot_)_com>]

On 7/9/07, Hans Petter Selasky <hselasky_(_at_)_c2i_(_dot_)_net> wrote:
> Perhaps what happens is that the "*pDst.bRam = _UCPU;"
> command clears the FIFO
> contents of the USB interrupt endpoint in addition to clearing the stall!?
>
> If the sequence is like this:
>
> Write to interrupt endpoint.
> Reply command is written to FIFO.
> Clear interrupt endpoint stall.
> There is no data to read, because the FIFO has been emptied as a part of the
> stall command.
>
> Xiaofan Chen: Could you check the datasheet for the chip that is used, what
> the stall command actually does?

I need to look further but what you mentioned seem to make sense.
"*pDst.bRam = _UCPU" seems to clear the buffer.

The following are the definitions for for the Buffer Descriptor Status
Register.

/* Buffer Descriptor Status Register Initialization Parameters */
#define _BSTALL     0x04                //Buffer Stall enable
#define _DTSEN      0x08                //Data Toggle Synch enable
#define _INCDIS     0x10                //Address increment disable
#define _KEN        0x20                //SIE keeps buff descriptors enable
#define _DAT0       0x00                //DATA0 packet expected next
#define _DAT1       0x40                //DATA1 packet expected next
#define _DTSMASK    0x40                //DTS Mask
#define _USIE       0x80                //SIE owns buffer
#define _UCPU       0x00                //CPU owns buffer

FYI:
Datasheet of PIC18F2550 USB MCU:
http://ww1.microchip.com/downloads/en/DeviceDoc/39632D.pdf

bit 7 UOWN: USB Own bit(1)
0 = The microcontroller core owns the BD and its corresponding buffer
1 = SIE owns the BD and its corresponding buffer
Note: This bit must be initialized by the user to the desired value prior
to enabling the USB module.

bit 6 DTS: Data Toggle Synchronization bit(2)
1 = Data 1 packet
0 = Data 0 packet
Note: This bit is ignored unless DTSEN = 1

bit 5 KEN: BD Keep Enable bit
1 = USB will keep the BD indefinitely once UOWN is set (required for
SPP endpoint configuration)
0 = USB will hand back the BD once a token has been processed

bit 4 INCDIS: Address Increment Disable bit
1 = Address increment disabled (required for SPP endpoint configuration)
0 = Address increment enabled

bit 3 DTSEN: Data Toggle Synchronization Enable bit
1 = Data toggle synchronization is enabled; data packets with incorrect
Sync value will be ignored
except for a SETUP transaction, which is accepted even if the data toggle
bits do not match
0 = No data toggle synchronization is performed

bit 2 BSTALL: Buffer Stall Enable bit
1 = Buffer stall enabled; STALL handshake issued if a token is received
that would use the BD in the given location (UOWN bit remains
set, BD value is unchanged)
0 = Buffer stall disabled

bit 1-0 BC9:BC8: Byte Count 9 and 8 bits
The byte count bits represent the number of bytes that will be
transmitted for an IN token or received
during an OUT token. Together with BC<7:0>, the valid byte counts are 0-1023.


What Warner mentioned before might also be correct.

On 7/8/07, M. Warner Losh <imp_(_at_)_bsdimp_(_dot_)_com> wrote:
> : > > I never learned the details, but a client of mine was able to get
> : > > fixes from Microchip for their product.  The exact problem was that
> : > > endpoint stall clearing didn't work for these devices and it was a
> : > > firmware bug.
> I need to look it up, but I believe that a clear endpoint stall also
> resets the toggle, and that was the bug that was tracked down.

FYI: the Firmware source code for PICkit 2, an HID device is here.
http://ww1.microchip.com/downloads/en/DeviceDoc/FirmwareV2.10.00.zip

I am still in the process of understanding the firmware so I may
misunderstand something. But I will try to dig further.

Regards,
Xiaofan
_______________________________________________
freebsd-usb_(_at_)_freebsd_(_dot_)_org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-usb
To unsubscribe, send any mail to "freebsd-usb-unsubscribe_(_at_)_freebsd_(_dot_)_org"