Seems like right now only the UHCI code supports timeouts. I've
not been able to persuade myself that it really matters ...
There is one driver using the feature now (stv680.c) and I was trying to
get it into the usbnet.c driver when I noticed the fact it was disabled.
Most drivers do their own timeout handling it appears.
I didn't know about stv680, but I did reach that same conclusion.
It doesn't matter too much if we support URB timeouts or we don't in 2.4,
but I don't like the fact that 2 HCDs support it, but 1 doesn't.
Make that 2 vs. 2: OHCI and EHCI against the UHCI twins! ;)
We should make a decision to fix the support in OHCI (I've already
started) or yank it out of the UHCI drivers.
And fix it in EHCI too ... at which point (so my thinking went)
it should really have been in re-usable code like "hcd.c".
The reason I didn't bother doing that was adding another
mechanism would just carry forward more complication and
bug-friendliness to future completion models, while the
goal (hcd.c) was more to simplify. ... plus, it seemed to be
an un-necessary mechanism.
I the only argument either way that I can think of is in regards to the
API. It's there already in 2.4 and I'd rather not change the API for 2.4
further, but then again it's a broken feature right now (on OHCI
atleast) and only one driver in the kernel uses it, so maybe it's
something no one will miss?
Anyone have an opinion either way?
If someone is volunteering to resolve the issue (patches), I'd be all in
favor of eliminating urb->timeout support from 2.5 at least. And I'd not
object to doing the same to 2.4 either.
- Dave
-------------------------------------------------------
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
