Re: false positive -ENOSPC in ohci-q

Andrea Arcangeli <[email protected]>
Newsgroups gmane.linux.usb.devel
Message-ID <[email protected]>
On Mon, Feb 26, 2007 at 08:07:59AM -0800, David Brownell wrote:
> On Monday 26 February 2007 7:11 am, Andrea Arcangeli wrote:
> > Hello,
> > 
> > until I applied the below patch I couldn't use the usb camera at the
> > same time of the usb modem (both usb1 devices). The camera returned
> > -ENOSPC. Now they both work fine at the same time (camera managed by
> > zoneminder and running all the time and the modem giving me the
> > internet link, both without any apparent problem).
> 
> NAK unless you can demonstrate a bug in the logic you disabled.

demonstrate a bug? I thought I already did. Things works perfectly
with that code disabled: zoneminder analyzes 8 frames per second while
my RTT latency to the other end of the modem is still 10msec and I get
zero errors and everything works flawlessy. With that broken code
compiled in, as soon as I connect to the internet the camera returns
-ENOSPC. What else could I ever need to do in order to demonstrate
that there is a bug in the logic that I disabled? I'm afraid I can't
provide a mathematical proof, cause I don't even understand what that
code is doing and what a worthless number 90% can ever be when things
may work perfectly at 95% load too (can you instead demonstrate that
90% is a correct max load instead of 95% or 100%?). Code like that
must be overoptimistic! it should catch over-exaggerate bandwidth
loads that are absolutely guaranteed to fail, it must not be
overstrict like it seems to be now, in turn generating false
positives.

Now if this is a hardware bug in the modem or the camera, I still
would need a runtime tweak to disable that heuristic that clearly
can choke in practice with real life usage (I can also imagine
hardware advertising a peak value, but if run in certain modes, like
with low res, generating lower max-values, so a runtime tweak to tell
the heuristic to go away still sounds useful even if this wouldn't be
a hardware issue and if the heuristic would be legitimate at only 90%).

> Looks more like your config is one that's disallowed.

I tried all possibile config and tweaks, so I doubt you can blame on me.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.