Hello Oliver, > If you feel you need to emulate functionality the core should have, the > core should have that functionality. So go hence and expand the core :-)
Ideally, yes. But this is a huge job for my limited understanding of usb core internals. The auerswald driver has started as a seperate package, which can be compiled for any 2.4.x kernel. Changing the core was clearly no solution. (The code in question was inside the auerswald driver from the very beginning, drivers/usb/auerswald.c ). I was told that in 2.5.x, there is no problem for submitting more than one control message any more. So IMHO, an intermediate solution is fine. There are more workarounds in this driver: there was a requirement of one-shot INT OUT urbs, with individual byte counts. Manage this to work with 3 different core drivers was difficult. There is also a race between probe() and another process doing control messages at the same time (maybe the usb fs?). I am very happy that it works now on nearly every PC it was installed. Do you suggest that I replace these workarounds by changing the usb core in 2.4.x also? > For anything but usbcore to mess with the internals of an URB is > unacceptable. Hu? The status code of an URB is a "public" member. As long as the URB is not submitted, I may do what I want (may I?). > But in your case I am not so sure about the need. It seems to me that you > need an approach with a little more abstraction. Something like: [abstraction deleted] This is pretty much the same logic as provided by auerchain. Auerchain has the advantage, that it mimics usb_submit_urb(). So one time in 2.5.x, all I have to do is erasing auerchain and replace it by the original calls. Sounds like a good idea to me... Basicly, I don't think that this is a worthwhile discussion. For 2.4.x, the core will not change so much, and a good-working driver is essential for the user community. For 2.5.x, all these is different. I don't want this code go into 2.5.x. best regards Wolfgang Mües ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
