On Wed, Apr 17, 2013 at 09:56:43AM +0200, Karsten Malcher wrote: > Am 15.04.2013 18:16, schrieb Greg KH: > >> When i look at the debug output of the opened working port i would > >> say there is a polling for reading data about every 100 µs. > >> http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=42;filename=usingport.log;att=3;bug=704242 > >> > >> [ 1749.073740] /build/buildd-linux-2.6_2.6.32-38-amd64-bk66e4/linux-2.6-2.6.32/debian/build/source_amd64_none/drivers/usb/serial/pl2303.c: pl2303_chars_in_buffer - returns 0 > >> [ 1749.073833] /build/buildd-linux-2.6_2.6.32-38-amd64-bk66e4/linux-2.6-2.6.32/debian/build/source_amd64_none/drivers/usb/serial/pl2303.c: pl2303_chars_in_buffer - returns 0 > >> > >> I didn't see something like this in kernel 3.2.0 ? > > You should, but note, that's userspace, or the line discipline asking > > for this, but we might have tty changes that have cleaned this up so we > > aren't so "chatty" anymore. > > I attached the sources of the driver from the Debian kernel sources for both versions. > In the screenshot of K3Diff you can see that there are some > subroutines missing in the new driver? > It's not only missing debug output. As Greg mentioned, a lot of things have changed since 2.6.32, and the pl2303 is using a lot more of the generic usb-serial implementation than it used to, hence your diff. If you can't run a bisect, what about testing some intermediate kernels? 2.6.35, 2.6.38 and 3.1 are good candidates. Also could your provide logs for a successful read on 2.6.32 and failed read on 3.2 when you have enabled debugging for usb-serial (e.g. modprobe usbserial debug=1) as well as pl2303? Is write also broken by the way? Thanks, Johan -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
