On Thu, Mar 10, 2011 at 2:26 PM, Favux ... <favux...@gmail.com> wrote: > Hi Chris, > > BackBONE7 reports the patch didn't work for his stylus issue. Sounds > more like jumps than jitters. > > He sees: > (WW) Serial Wacom Tablet stylus: bad data at 7 v=91 l=9 > (WW) Serial Wacom Tablet stylus: bad data at 7 v=91 l=9 > (WW) Serial Wacom Tablet stylus: bad data at 8 v=91 l=9 > With both ToolDebugLevel for stylus and TabletDebugLevel.
In what context are these messages seen? I recall something about right when stylus is hovering between in and out of proximity? The above says it received a packet of length 7 and 8 when it was expecting a packet of length 9. There is some logic in wcmISDV4.c related to packet length size and its based on status of proximity. That might be get confused by hovering. Length of 7 is valid packet size for pen+touch tablets. Is this one such device? And the v=91 value also looks touch-ish. Also, I'm wondering if wcmSerialValidate should return a value of "n-1" instead of "n" so that it attempts to restart at what looks like a new header? If it becomes confused on 1 packet length, it looks like it may feed on itself and ignore several? I could be misreading here. If the user wants to try it, shouldn't hurt much. Try editing wcmISDV4.c line #182 and change "return n;" to "return n-1;". > > I don't know if I'm seeing jumps in his debug Xorg.0.logs or not. What > I do notice is that it'll often hold the same x & y locations for 2 to > 5 times and then go to the next one. One thing I notice during the hangs from snippet you emailed is that it hangs until this happens a couple of times: (II) /dev/ttyS0 (10:wcmReadPacket): buffer has 8 bytes Packets of 9 and 7 are OK but not 8. Something is going wrong but eventually cleans up after itself. The above "n-1" change may allow it to correct faster and less jumps. > Since he updated to > xf86-input-wacom-0.10.11+ I'm wondering if the filter is holding onto > old values too long. Maybe the serial tablets don't stream in as much > data as the usb ones do? So the filtering window or average is wrong > for them? Or maybe it's just a few serial digitizers like his. A > sample of what I'm talking about is attached. He supplied a complete > Xorg.0.log with both ToolDebugLevel for stylus and TabletDebugLevel > turned up. I'm currently guessing its because of discarded data but I could be totally wrong of course. I think it may take a developer with tablet to figure this out totally. Chris ------------------------------------------------------------------------------ Colocation vs. Managed Hosting A question and answer guide to determining the best fit for your organization - today and in the future. http://p.sf.net/sfu/internap-sfd2d _______________________________________________ Linuxwacom-devel mailing list Linuxwacom-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel
