Date: Thu, 6 Dec 2001 01:27:35 +0000 From: Tony <tony@ubik.demon.co.uk> To: freebsd-mobile@freebsd.org Subject: Re: USB Problems Message-ID: <TYj1nQAHmsD8IwKU@ubik.demon.co.uk> In-Reply-To: <200112051002.LAA21436@bonn.rtg.consol.de> References: <200112051002.LAA21436@bonn.rtg.consol.de>
next in thread | previous in thread | raw e-mail | index | archive | help
In article <200112051002.LAA21436@bonn.rtg.consol.de>, Thorsten Greiner
<Thorsten.Greiner@consol.de> writes
>Since I got no response to my first posting, I will try again with a more
>detailed description. I'm trying to get USB working on my ASUS L7300 laptop.
>I am running 4.4-STABLE, my kernel config looks like this (USB part only):
I am not familiar with the workings of the USB stack on FreeBSD.
However I have worked on programming a USB host driver and testing this
with a commercial USB stack included in a Set-Top Box embedded OS. As
this was at a client site a year ago I can't remember all the details.
[snip]
>So the USB controller is recognized and shows up in the usbdevs output. Now I
>start usbd manually and connect my Logitech Mini Wheel Mouse. This happens:
A Logitech wheel mouse, and 'no name' versions with the same
microcontroller in various different mouldings, that caused me the most
hassle when testing my USB driver.
>Dec 5 09:45:40 tybalt /kernel: uhci_timeout: ii=0xc0cbcb00
>Dec 5 09:45:51 tybalt last message repeated 2 times
>Dec 5 09:45:51 tybalt /kernel: usbd_new_device: addr=2, getting first desc
>failed
>Dec 5 09:45:51 tybalt /kernel: uhub_explore: usb_new_device failed,
>error=TIMEOUT
>Dec 5 09:45:51 tybalt /kernel: uhub0: device problem, disabling port 1
Before integrating my USB driver to the STB I wrote a test harness that
read a device's information when a device was plugged in.
The Logitech mouse behaved weirdly in this test. If device
configuration was not completed in a short time, (10's of milliseconds I
think), the mouse removed itself from the USB bus and reset it's
microcontroller. My test harness was too slow, and the appearance was
that the mouse was being plugged & unplugged 2 times a second.
USB devices should not behave like this, it is certainly not specified
anywhere. The person writing the device code felt there was a good
reason, perhaps to work with a buggy USB stack. <shrug> It caused me
some hassle, and eventually an hour of combat with a digital 'scope.
On a running system this particular weirdness in the mouse should not be
a problem, and if this what you are hitting it indicates the UHCI or
mouse drivers have too much latency during device configuration.
During my testing I also saw some other problems with devices
misbehaving if enumeration did not exactly follow the order of actions
the Windows USB stack uses. I cannot remember if the Logitech mouse had
any of these.
>I have tried the same thing with the usb debug level raised to 10, but this
>produces a lot of output - don't want to spam everybody with this.
If you want to do this and email the log to me I'll see whether I
recognise the symptoms.
Regards,
Tony
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-mobile" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?TYj1nQAHmsD8IwKU>