Date: Mon, 19 Oct 2009 21:43:56 +0200 From: John Hay <jhay@meraka.org.za> To: Hans Petter Selasky <hselasky@c2i.net> Cc: freebsd-usb@freebsd.org Subject: Re: [8.0] hplip3/ugen printer problem Message-ID: <20091019194356.GA4815@zibbi.meraka.csir.co.za> In-Reply-To: <200910192135.09875.hselasky@c2i.net> References: <20091018153910.375f8013@baby-jane.lamaiziere.net> <200910190219.49761.hselasky@c2i.net> <20091019185852.GA99922@zibbi.meraka.csir.co.za> <200910192135.09875.hselasky@c2i.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Oct 19, 2009 at 09:35:08PM +0200, Hans Petter Selasky wrote:
> On Monday 19 October 2009 20:58:52 John Hay wrote:
> > On Mon, Oct 19, 2009 at 02:19:49AM +0200, Hans Petter Selasky wrote:
> > > On Sunday 18 October 2009 20:40:31 John Hay wrote:
> > > > > > io/hpmud/musb.c 1028: bulk_write failed buf=0xbfbfc93c size=8192
> > > > > > len=-60:
> > >
> > > Error -60 means that the transfer timed out.
> > >
> > > Maybe the timeout given by the driver is too strict.
> > >
> > > Try increasing the timeout, or change the usb_bulk_write() code into
> > > temporarily using no timeout, timeout=0.
> >
>
> Hi,
>
> > With a timeout of 0 it is working properly.
>
> Great!
>
> The USB stack as a clamp on the timeout:
>
> xfer->timeout = fs_ep.timeout;
> if (xfer->timeout > 65535) {
> xfer->timeout = 65535;
> }
>
> The unit is milliseconds.
But then there should have been a 65 second timeout? I just saw a 5-8
second timeout.
Oct 19 19:57:41 angel hp[4016]: io/hpmud/musb.c 948: usb_bulk_write start
Oct 19 19:57:49 angel hp[4016]: io/hpmud/musb.c 951: usb_bulk_write stop -60
...
Oct 19 19:57:52 angel hp[4016]: io/hpmud/musb.c 948: usb_bulk_write start
Oct 19 19:58:00 angel hp[4016]: io/hpmud/musb.c 951: usb_bulk_write stop -60
...
Oct 19 20:21:38 angel hp[14189]: io/hpmud/musb.c 948: usb_bulk_write start
Oct 19 20:21:46 angel hp[14189]: io/hpmud/musb.c 951: usb_bulk_write stop -60
>
> > What unit is timeout? The comments in the code seems to think it is
> > in milliseconds and they set it to 72 hours (72*3600*1000) because
> > they actually have a timeout thread to handle timeouts. Both the
> > above value and that times 1000 seem to timeout in about 4-5 seconds.
> >
> > The file I have been editing:
> > work/hplip-3.9.8/io/hpmud/musb.c
>
> Can you report this problem to the authors of hplip-3.9.8 ?
>
> It seems FreeBSD behaviour is different that Linux behaviour!
John
--
John Hay -- jhay@meraka.csir.co.za / jhay@FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20091019194356.GA4815>