Hi Johannes,
> -----Original Message-----
> From: Johannes Erdfelt [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, April 06, 2000 4:52 PM
> To: Dunlap, Randy
> Cc: '[EMAIL PROTECTED]'
> Subject: Re: [linux-usb] local usb plugtest
>
>
> On Thu, Apr 06, 2000, Dunlap, Randy <[EMAIL PROTECTED]> wrote:
> > I did a small Linux-USB plugtest on my machine
> > with around 25 USB devices. Here are the config and
> > results.
> >
[snip]
> > 6 of 7 keyboards are working.
>
> What problem did the one keyboard have? Did it enumerate? etc
I'll get more info on this for you tomorrow. I'm aware
that it was missing some specifics.
> > All 3 are listed as diretories in /proc/scsi.
> > However, only 2 of them are listed in /proc/scsi/scsi
> > (the 2 Y-E DATA floppy drives are listed; the Zip-100
> > drive is not listed).
>
> > Could not unload usb-storage module (hangs system);
> > had to reboot system. Repeated and was able to unload
> > all modules OK.
>
> Interesting, Did you use the last patch I posted? (The most recent one
> you resent off to Linus)
Yes.
> I've had numerous problems with the usb-storage driver, all
> oops' in the
> usb-storage kernel threads and all NULL pointer deferences.
> I'm going to
> try the latest patch posted recently.
>
> > - cpia camera + HCD stability (I've seen some instability;
> > will report on it later.)
>
> I've seen a couple of problems myself recently. I was working with a
> coworker of mine on a demo and after a couple of minutes, the camera
> would start spewing STALL's on the control pipe.
>
> According the USB 1.1 specs, this mean a) The command was not
> recognized
> or b) Something really bad went wrong. Since the command
> doesn't change
> much, I don't think a) was the culprit.
>
> I only had this happen with 1 camera, another camera didn't
> produce this
> behaviour. I haven't tested this extensively yet.
I haven't either. I've also seen problems with both
UHCI HCDs. I'll try to gather more info for this also.
> I've also seen instability with usb-uhci. It locks my machine
> hard after
> a couple of minutes. Very reproducible. This is the same problem I
> reported a couple of weeks ago.
>
> > - increase maximum size of /proc/bus/usb/devices output
> > from 8 KB to 16 KB (comments ? make it dynamic ?)
>
> Making it dynamic may be tough since I don't think the code knows how
> large the buffer should be.
>
> I'd personally like to see /proc/bus/usb/devices removed and
> leave it up
> to userspace tools. Getting the ASCII stuff working right may
> be tough.
>
> But the whole ASCII vs binary flamewar is better left to
> linux-kernel :)
I didn't want to make it dynamic anyway. Moving it to
userspace is OK with me (one of these days).
Yeah, let's leave that discussion for lkml.
~Randy
>
> JE
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
