I have now got a test bed with 4 machines. 2 running the latest kernel 2.6.11.4-21.7-default and one running 2.6.4 and 2.6.8
I can clear up the first issue and confirm that all machines using my driver fail with the USB 2.0 hub so that has been thrown out and all tests are with the device plugged into the machine. the earlier kernels work fine with my device. The 2.6.11 machines fail with the following trace. Any hints would be most welcome Thanks Jonathan Jun 14 08:40:36 skydriver-driver.c: XXXXXX sky_open Jun 14 08:40:46 skydriver-driver.c: Sky Read 4096 Jun 14 08:40:46 skydriver-driver.c: pipesize 128 Jun 14 08:40:46 skydriver-driver.c: Sky Read 1 Jun 14 08:40:46 skydriver-driver.c: B4 Assignment Sky Read 4096 Jun 14 08:40:46 skydriver-driver.c: dev->read_urb[dev->readURB]->transfer_buffer_length 128 Jun 14 08:40:46 skydriver-driver.c: Sky Read set up urbs Jun 14 08:40:46 skydriver-driver.c: Sky Read submit 0, ret 0, URB No 0 Jun 14 08:40:46 skydriver-driver.c: B4 Assignment Sky Read 4096 Jun 14 08:40:46 skydriver-driver.c: dev->read_urb[dev->readURB]->transfer_buffer_length 128 Jun 14 08:40:46 skydriver-driver.c: Sky Read set up urbs Jun 14 08:40:46 skydriver-driver.c: Sky Read submit 1, ret 0, URB No 1 Jun 14 08:40:46 skydriver-driver.c: B4 Assignment Sky Read 4096 Jun 14 08:40:46 skydriver-driver.c: dev->read_urb[dev->readURB]->transfer_buffer_length 128 Jun 14 08:40:46 skydriver-driver.c: Sky Read set up urbs Jun 14 08:40:46 skydriver-driver.c: Sky Read submit 2, ret 0, URB No 2 Jun 14 08:40:46 skydriver-driver.c: B4 Assignment Sky Read 4096 Jun 14 08:40:46 skydriver-driver.c: dev->read_urb[dev->readURB]->transfer_buffer_length 128 Jun 14 08:40:46 skydriver-driver.c: Sky Read set up urbs Jun 14 08:40:46 skydriver-driver.c: Sky Read submit 3, ret 0, URB No 3 Jun 14 08:40:46 skydriver-driver.c: Sky Read B4 Sleep Count 4096 Jun 14 08:40:46 skydriver-driver.c: Sky Read After sleep Jun 14 08:40:46 skydriver-driver.c: Callback !!!! Jun 14 08:40:46 skydriver-driver.c: sky_read_loop: corrupted packet status: -84 Jun 14 08:40:46 skydriver-driver.c: packets 0 Jun 14 08:40:46 skydriver-driver.c: Sub Count 3 Jun 14 08:40:46 skydriver-driver.c: Callback !!!! Jun 14 08:40:46 skydriver-driver.c: sky_read_loop: corrupted packet status: -84 Jun 14 08:40:46 skydriver-driver.c: packets 0 Jun 14 08:40:46 skydriver-driver.c: Sub Count 2 Jun 14 08:40:46 skydriver-driver.c: Callback !!!! etc Jonathan Selby wrote: >This is quite interesting - the system is in production on SuSE 9.1 and Debian >unstable. > >The main production system uses a USB 1.1 motherboard and a second machine >uses a USB 2 motherboard. > >Pulling the 2.0 hub out and we have different problems but the 1st batch of >URBs are being submitted anyway and we are now failing on the callbacks. I >have a way forward to debug this so progress has been made but I can confirm >that we do have a working version on 2.6.6 and 2.6.8 kernels. > >Will backtrack and see if I come up with anything. > >I assume the EHCI driver is work in progress and will be happy to test any >improvements that may come about. > > >Thanks again > > >Jonathan > >*********** REPLY SEPARATOR *********** > >On 6/13/2005 at 09:19 David Brownell wrote: > > >>On Monday 13 June 2005 8:28 am, Alan Stern wrote: >> >> >>>I kind of expected something like this. The EHCI driver doesn't do a >>> >>very >> >>>good job of handling ISO transfers that involve a speed change (i.e., >>>full-speed device connected through a high-speed hub). There's some >>>pretty hairy scheduling involved and the driver just can't handle it yet. >>> >>Though I've been hearing success reports from folk with OUT transfers; >>it's the isochronous IN transfers that still seem like trouble. Folk >>have only recently started to try them. >> >>On the other hand ... did you say that this worked in 2.6.8 with the >>same configuration (through EHCI and a USB 2.0 hub)? That'd be odd. >> >>- Dave >> >> >> >>------------------------------------------------------- >>This SF.Net email is sponsored by: NEC IT Guy Games. How far can you >>shotput >>a projector? How fast can you ride your desk chair down the office luge >>track? >>If you want to score the big prize, get to know the little guy. >>Play to win an NEC 61" plasma display: http://www.necitguy.com/?r=20 >>_______________________________________________ >>[email protected] >>To unsubscribe, use the last form field at: >>https://lists.sourceforge.net/lists/listinfo/linux-usb-devel >> > > >Jonathan Selby >Director - Software Development >Xaxero Marine Software Engineering Ltd >at skype.com - xaxjon >Satellite Phone (00) 88163 142 9922 >Cell Argentina (54) 929 0160 2064 >(00) 64 (0)9 412 7580 fax (00) 64 (0)9 412 7579 >http://www.xaxero.com >Software for extending your horizon.... > > > > > >------------------------------------------------------- >This SF.Net email is sponsored by: NEC IT Guy Games. How far can you shotput >a projector? How fast can you ride your desk chair down the office luge track? >If you want to score the big prize, get to know the little guy. >Play to win an NEC 61" plasma display: http://www.necitguy.com/?r >_______________________________________________ >[email protected] >To unsubscribe, use the last form field at: >https://lists.sourceforge.net/lists/listinfo/linux-usb-devel > > > ------------------------------------------------------- This SF.Net email is sponsored by: NEC IT Guy Games. How far can you shotput a projector? How fast can you ride your desk chair down the office luge track? If you want to score the big prize, get to know the little guy. Play to win an NEC 61" plasma display: http://www.necitguy.com/?r=20 _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
