I've written a driver successfully for full speed (USB1.1). It is for an FX2 based DAQ board. However, when I run it in high speed (USB 2.0) with the ehci driver the kernel crashes. This happens at the moment when I resubmit the urb within the interrupt context. The first transfer from the FX2 is fine. Only the resubmission crashes the kernel.
I'm still running 2.4.21. The new 2.6.0-test1 does not work at all and I need comedi which is not available for 2.6.x anyway. Has anybody an idea what makes it so different if I do a submit or a resubmit? The transfer mode is isochronous.
Related to this problem: how often is the completion routine called? Every microframe (125us) in high speed? Does urb->interval control this?
Re the FX2: when I look at the bInterval I get a realistic value for my ISO-out endpoint (1). However, for my ISO-in endpoint I get an arbitrary value. Looks like as if it is not initialised. Maybe the reason is that I'm using alternate setting 3 of the FX2 and only this is not inialised.
What I find also irritating that the ehci scheduler needs an interval for ISO transfer although this is only needed for int endpoints. At least this says the documentation when I look at usb.h.
Is there a driver out which does isochronous transfers under high speed?
/Bernd
-- http://www.cn.stir.ac.uk/~bp1 mailto:[EMAIL PROTECTED]
------------------------------------------------------- This SF.net email is sponsored by: VM Ware With VMware you can run multiple operating systems on a single machine. WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the same time. Free trial click here: http://www.vmware.com/wl/offer/345/0 _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
