On Thu, Feb 13, 2003 at 12:11:34PM -0500, Dan Parks wrote:
> >> "iso_find_start: gap in seamless isochronous scheduling"
> I moved this conversation over to the developer list, since I think it
> is more apropos. We have our own custom driver that has been around for
> about 3 years. In the beginning of the development we had gotten this
> error, but after a complete rewrite over a year ago, we never got the
> error anymore. Our driver has been tested with many different computers
> and the only computer that doesn't work is this one specific Dell
> computer. Furthermore, the RedHat 8.0 (2.4.18-14) kernel with this
> computer does not give us any problem, but vanilla 2.4.18/2.4.19 kernels
> as well as our modified kernel which is a (2.4.19 with O(1), preemptive,
> and low-latency) all give us this error. All the other computers we use
> run this modified kernel, and we have never experienced this problem
> since the rewrite. I don't expect people to go through the whole source
> of our driver and find the problem, I'm just asking for more information
> about this error: what does it mean exactly, and in what ways can it be
> caused.
It means that the currently submitted URB is submitted too late to guarantee
a continous data flow. It works like this: When the end frame of the last
queued URB is found, it is checked whether the current frame counter lies
between that end and (end+number_of_wanted_frames). In that case, a
unchanged scheduling would result race between the HCD and the descriptor
insertion and in a totally scrambled stream. So the new start is moved a bit
into the future (currently 5 frames from "now").
The causes for this message can be:
- Too much scheduled URBs (i.e. with the new URB, there would be more than
1024 frames "running")
- Not enough URBs running (eg. only 2). At the completion of the first URB,
it's absolutely too late to submit the second one...
--
Georg Acher, [EMAIL PROTECTED]
http://wwwbode.in.tum.de/~acher
"Oh no, not again !" The bowl of petunias
-------------------------------------------------------
This SF.NET email is sponsored by: FREE SSL Guide from Thawte
are you planning your Web Server Security? Click here to get a FREE
Thawte SSL guide and find the answers to all your SSL security issues.
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
