Usb cpia webcam on assabet/neponset----------about /dev/video

[Jianlin Liang]

Thanks!  Yesterday I also thought about this problem. And now I've made the cpia webcam work in assabet/neponset with the OHCI driver.   The items I actived in the .config file are: CONFIG_VIDEO_DEV=y CONFIG_VIDEO_CPIA=y CONFIG_VIDEO_CPIA_USB=y CONFIG_USB=y CONFIG_USB_DEVICEFS=y CONFIG_USB_DEBUG=y CONFIG_USB_HID=y CONFIG_USB_OHCI=y CONFIG_SA1100_USB=y CONFIG_SA1100_USB_DEVICEFS=y   At the same time I disabled the following lines (I had no formal reason for doing this, but just thought that if the webcam is the only working USB device things would be better. ) CONFIG_SA1100_USB_NETLINK=m CONFIG_SA1100_USB_CHAR=m   Also you must make the above changes effective in include/linux/autoconf.h.  (As Russell King mentioned, I used an improper way to change this file. But at that time I just didn't know other methods that can do it. And thanks for Russell King to point out my error :)   )   And the following line in /drivers/usb/usb-ohci.c was commented: "#define DEBUG_STALL" (This was done during the debug process. The part between "#indef DEBUG_STALL" and "#endif" made the webcam unstable) Then I built zImage, loading it into assabet/neponset board. After it was booted up, I used "mknod" to create the video device, and then run Xwindow, then gqcam, the image was shown on the LCD, perfectly.   The kernel I used was linux-2.4.2-rmk2-np2. Now I am thinking about another problem: In arm-linux 2.4.4 there are some PCI parts in the ohci driver. But as far as I know, there's no PCI support in assabet/neponset. So I am afraid that the OHCI driver in the new version can not work in assabet/neponset. Maybe I should ask the person who designed this driver about that.   Anyway, now the webcam works in assabet/neponset. Thanks for all the persons who gave me so much help. And thanks for this maillist,too.   Jianlin Liang     >OHCI/UHCI defines what USB host chipset your unit has; the SA1111 has an >OHCI driver. Using a UHCI driver will do *nothing* as there is no UHCI >hardware present. >As the webcam is at least doing something, this proves that the OHCI driver >is doing something and the videodev is also doing something. There may well >be a bug in anything from cqcam downwards, though. >Hugo