[linux-dvb] Re: Nova-T 500 (dvb_usb_dib0700) usb disconnects

Antti P Miettinen ananaza at iki.fi
Wed Feb 28 21:49:58 CET 2007


Antti P Miettinen <ananaza at iki.fi> writes:
> But anyway - I'll be power cycling the machine real soon now :-)

I'd say the dvb-usb-dib0700-02-rc2.fw makes the transaction errors
more frequent for me me. Here's the latest kernel log:

[17179744.028000] ehci_hcd 0000:03:0c.2: XactErr, ep3in, token=8000cd08
[17179788.896000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17179790.004000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=00004d08
[17179836.196000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=80004d08
[17179881.444000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=80004d08
[17179926.528000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17179927.636000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=0000cd08
[17179972.716000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17179973.836000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=80004d08
[17180018.896000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180020.004000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=8000cd08
[17180065.084000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180110.264000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180156.488000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180157.596000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=80004d08
[17180202.708000] ehci_hcd 0000:03:0c.2: XactErr, ep0in, token=00000d08
[17180203.816000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=8000cd08
[17180248.940000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180250.048000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=80004d08
[17180295.128000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180296.236000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=00004d08
[17180341.416000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=80004d08
[17180386.496000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180387.608000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=00004d08
[17180433.800000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=80004d08
[17180478.940000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180480.052000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=0000cd08
[17180525.116000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=00008d08
[17180526.228000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=00004d08
[17180571.420000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=00004d08
[17180616.484000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=80000e08
[17180617.596000] ehci_hcd 0000:03:0c.2: XactErr, ep2in, token=8000cd08
[17180662.676000] ehci_hcd 0000:03:0c.2: XactErr, ep3in, token=8038c948
[17180662.676000] ehci_hcd 0000:03:0c.2: XactErr, ep0out, token=80008148
[17180662.676000] ehci_hcd 0000:03:0c.2: devpath 1 ep0out 3strikes,t=80008148

About usbmon logs. Does anyone know of tools for parsing them?
Decoding by hand with the USB 2.0 spec is quite time consuming.

Anyway - one way to find the disconnect is to look for transactions
addressed to device one, endpoint zero, which is the control endpoint
of the hub on the bus. For example:

f7723ac0 30344798 S Ci:001:00 s a3 00 0000 0001 0004 4 <
f7723ac0 30344798 C Ci:001:00 0 4 = 01050000

is a port status query. What seems to be in my logs before hub access
is often a lot of (sometimes just a few) zero length callbacks with
status -71 to submits with request length 39480 (which seems to be the
dib0700 streaming buffer size) for endpoints 2 and 3. So is this -71 here
-EPROTO? Is that normal? The pattern seems pretty consistent. Soon
after starting to get -71 status callbacks, we will be talking to the
hub (and logs that do not contain a disconnect do not contain -71
status). Hmm.. looks like there is also always one callback with
status -32 (-EPIPE?) before those -71 status callbacks. And the data
length for those callbacks is 39424 which is 39480-56 and 56 is the
transfer length in the token preceding the hard failure.. hmm..

-- 
http://www.iki.fi/~ananaza/




More information about the linux-dvb mailing list