On Thu, Sep 15, 2005 at 02:10:41PM -0400, Lee Revell wrote: > On Thu, 2005-09-15 at 14:13 +0200, Clemens Ladisch wrote: > > Horms wrote: > > > On Wed, Sep 14, 2005 at 03:29:13PM +0200, Daniel Svensson wrote: > > > > CONFIG_USB_BANDWIDTH is enabled by default in the debian kernel. Could > > > > this be changed? According to the alsa developer who helped me debug the > > > > problem with my usb soundcard CONFIG_USB_BANDWIDTH should be disabled. > > > > The original email is forwarded here: > > > > > > That sounds fair enough to me, as long there are no objections > > > from the rest of the kernel-team. > > > > > > Clemens, could you provide a little more information on > > > the problems that CONFIG_USB_BANDWIDTH causes? > > > > The code that implements it computes wrong values for the bandwidth > > utilization. This doesn't seem to matter much for full speed stereo > > devices, but we have several bug reports for multichannel and/or high > > speed devices where usb_submit_urb() incorrectly returns -ENOSPC > > errors. > > > > The comment above the usb_check_bandwidth() function in hcd.c lists > > several of the known bugs. > > > > Perhaps CONFIG_USB_BANDWIDTH should have been marked BROKEN instead of > > just EXPERIMENTAL.
Thanks for the clarification. > Has this been reported upstream? Possibly the USB maintainers don't > know that this is causing real problems for users. That would be a good idea if it hasn't been done, but I know nought about this sistuation, so my comments are close to worthless. -- Horms -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
