On Mon, May 03, 2004 at 12:02:25PM -0700, David Brownell wrote: > Zephaniah E. Hull wrote: > > > >At the moment, the 2.6.x USB core is /very/ noisy, giving KERN_WARNING > >level messages for all failed usbfs bulk transactions (in addition to > >returning the error), and for all usb transfer timeouts (in addition to > >returning the error). > > Is it noisier than 2.4 in such ways?
To be honest, it has been long enough since I've used a 2.4 kernel that
I don't know, but I could check the source if need be.
>
>
> >This results in two messages to syslog (and often the console) for all
> >bulk transfer timeouts, which seems rather excessive for information
> >which /may/ be useful when attempting to debug things, but which the
> >calling program has every chance to deal with itself.
> >
> >It also gives KERN_INFO level messages for all connections and
> >disconnections, and while this may have more value then the debugging
> >messages above, it can still get somewhat extensive for cases of heavy
> >hot-plugging.
> >
> >In both cases the messages are unconditional. Would it be reasonable
> >for these messages to be controlled by some of the various debugging
> >settings?
>
> I'd think most of those should be KERN_DEBUG message, which should
> already be turned on/off by CONFIG_USB_DEBUG.
Should be, but they are not at this time. At least not in the current
-mm kernels.
This is /exceedingly/ annoying when one is doing odd things with usb
programs which results in /tons/ of timed out reads. (Long story.)
--
1024D/E65A7801 Zephaniah E. Hull <[EMAIL PROTECTED]>
92ED 94E4 B1E6 3624 226D 5727 4453 008B E65A 7801
CCs of replies from mailing lists are requested.
This is commonly attributed to the lusers spending too much time talking
with their BOFH. They start thinking their name is "Moron" or "Dimwit"
because you keep calling them that. -- Toni Lassila <[EMAIL PROTECTED]>
in the Scary Devil Monastery about lusers forgetting their own names
signature.asc
Description: Digital signature
