Hello Florian,
On Tue, Dec 02, 2003 at 03:35:27PM +0100, florian wuechner wrote: > I'm just wondering what is the status for the cyberjack driver and the > upcoming new linux kernel 2.6.x? > > So far the driver included with 2.6.0-test11 does not work for me. > I'm using the uhci-hcd host controller on a Debian testing/unstable > system. I did add the locking to 2.6.0-t2. Newer kernels don't work, because calling copy_from_user inside a locked area. My current solution has not been accepted because it uses to much kernel stack. So I will rewrite that using kmalloc/kfree, which should be acceptable at the communication rate of the reader. > I guess I had the same problems, which you have described on the > linux-usb-devel-list: > http://sourceforge.net/mailarchive/message.php?msg_id=5831507 The above problem is not, what you probably have seen with test11. That is the problem you desribe at LABEL1. > I was playing around a little with the driver and after comparing it > with the kobil_sct.c driver. > I copied the function *_write_room to the cyberjack driver: > cyberjack_write_room which always returns 8. That might be another solution for kmalloc/kfree. I'll look into that. LABEL1: > However, after I did that the cyberjack test (cjtest) program runs > successfully. But only for the first time. If I unplug and replug the > cyberjack it works again (but only once). That is the problem and I still don't know why that is happening. The cyberJack driver worked allways with 2.4 and UHCI without problems, but it had the same problem as described in my message up to 2.4.19 with OHCI and it worked with OHCI starting with 2.4.20 without changing the cyberJack driver. I haven't yet tested it with OHCI, but I have now a dual-CPU Opteron with OHCI controller here. Other USB serial devices seem to have similar problems with 2.6.0-t*. There a queue flushing did solve the problem, but not with the cyberJack driver. Because of the history of the 2.4 kernel and the USB analyzer protocol I believe, that there is a problem in the UHCI HC driver or a problem with all UHCI and OHCI HCs which had been fixed in 2.4. Otherwise I do not know, why the PC gets the interrupt packet, but not the driver. > So far I have no idea, what this function is supposed to do or why it is > working now.... > > I'll attach the output from cjtest and my parts of kern.log. > > Thanks! > Florian > Dec 2 15:23:01 venus kernel: drivers/usb/core/usb.c: usb_hotplug Nice name for a computer. :) Mahlzeit endergone Zwiebeltuete ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
