diff options
Diffstat (limited to 'docs/docs.html')
| -rw-r--r-- | docs/docs.html | 2217 |
1 files changed, 2217 insertions, 0 deletions
diff --git a/docs/docs.html b/docs/docs.html new file mode 100644 index 0000000..adbd29c --- /dev/null +++ b/docs/docs.html @@ -0,0 +1,2217 @@ +<HTML> +<HEAD> +<TITLE>John's Linux USB Wacom Page</TITLE> +</HEAD> +<STYLE> + .diff { background: #EEEEFF } + .nav { background: #EEEEEE } +</STYLE> +<BODY> +<CENTER><H1>John's Linux USB Wacom Page</H1>Last Modified: December 22, 2002 18:12<BR><A HREF="#changelog">Changelog</A></CENTER><BR><CENTER CLASS=nav>Navigation: +<A HREF="/wacom/index.php/id/book">MAIN</A> <A HREF="/wacom/index.php/id/intro">NEXT</A> <A HREF="/wacom/index.php/toc">INDEX</A> <A HREF="/wacom/index.php/all"><B>ALL</B></A> </CENTER><BR><BR> +<A NAME="intro"> +<H1>1.0 - Introduction</H1> +<P>This document was written for three reasons: 1) to remind myself +how I got my brand-new Wacom Intuos2 running on my Redhat 8.0 Linux box in +case I have to do it again from scratch, 2) to help other people do the same, +and 3) to help de-mystify the process by explaining why things work as they +do, and where to look when things seem to be going wrong. + +<P>If you find an error, have a question, or have something to add, please +contact me directly at +<A HREF="mailto:jej@j-arkadia.com?subject=WACOM">jej@j-arkadia.com</A>. +Furthermore, please add the word "WACOM" to the subject line. I get hundreds +of messages a day, so that'll help me see your message amidst the spam. +<A NAME="expect"> +<H2>1.1 - Expectations</H2> +<H2>What You Can Expect Today</H2> + +<P>Depending on whether you are using serial or USB tablets, your experience +will differ, but for the moment at least, the serial tablets have a generally +simpler configuration. This document currently addresses only the USB +tablets, but that will change in the future. Mr. Lepied's +<A HREF="http://people.mandrakesoft.com/~flepied/projects/wacom/">XInput driver for XFree86</A> web page +has a good description of the serial configuration process. The goal of +this website is to make configuration and use of both tablet types as +simple as possible. Here is what you can expect now: + +<H3>You Should Be Able to Get Your Tablet Working</H3> + +<P>Most all serial and USB tablets are working to a reasonably functional +degree. I have an old ArtPad (1997) that does not initialize with the +XFree86 driver, and USB Cintiq users are having a little trouble with the +kernel drivers. Everyone else should be up and running, namely: + +<UL> +<LI>PenPartner, +<LI>Graphire, +<LI>Graphire2, +<LI>Intuos, and +<LI>Intuos2 +</UL> + +<H3>All Your Fancy Styluses and Pointers Should Work</H3> + +<P>I have a collection of pointers, pens, and brushes to test; all of them +work completely as one would expect. Users have reported success with +their tools as well. I have not tested the airbrush yet, but presumably +it works just fine. Tools that are known to work include: + +<UL> +<LI>Graphire2 2D mouse (EC-120-0K), +<LI>Intuos Pen (GP-300E-01H), +<LI>Intuos2 2D mouse, +<LI>Intuos2 4D mouse, +<LI>Intuos2 Lens cursor, +<LI>Intuos2 grip pen (XP-501E-00A), +<LI>Intuos2 stroke pen (XP-120-00A), +<LI>Intuos2 ink pen (XP-110-00A), +</UL> + +<H3>You Will Need to Modify Your XF86Config File</H3> + +<P>This project is actively working to become a turn-key system. It will be +awhile before major distributions like Redhat include these settings +automatically for you. Until that time, you will need to make some changes +to your XFree86 configuration files by hand. + +<H3>You May Need to Update Your Kernel</H3> + +<P>This project assumes that you have a kernel version of at least 2.4.18. +While it may be possible to use this document with older kernels, the +likelihood of success diminishes the farther back you go. Redhat 8.0 +ships with this kernel, and most of the major distributions are already +past this point. + +<H3>You Will Need to Work With the Command Line</H3> + +<P>This document details a step-by-step process for setting up and diagnosing +the tablet. There are no GUI tools to help you through this process- at +least not today. Some familiarity with the command line is recommended +since I provide enough information to get the task accomplished, but I do +not explain how the command line itself operates. Any good book on UNIX +will do, and I use the bash shell for all my work. If you are not using +bash, you presumably know what you are doing already. + +<H2>What You Can Expect Soon</H2> + +<P>I have a near-term design goal to get all the tablets and their respective +tools functioning using the original driver design. This currently includes +the airbrush (which may work already), the Cintiq, and if possible and time +permits, the lowly ArtPad. If there are any outstanding problems with the +original driver, I will take care of them as they come up. + +<P>While the current way of configuring the tablet is sub-par in my opinion, +it is also the established way of doing things. Merging large-scale changes +back into either the Linux kernel or the XFree86 code tree will be difficult, +if for no other reason than insufficient multi-platform testing. To solve +this problem, the new functionality described below will be implemented in +a separate project and accessed through the original driver as an advanced +option. This will allow all the various platforms to use the new code or not +depending on their own constraints and circumstances. It also allows for +people to write their own tablet libraries without mucking around in the +XFree86 code. + +<H2>The Next Steps</H2> + +<P>First and foremost, the tablet needs to be configurable without restarting +X. By separating the tablet code from the low-level X code, it will be +possible to talk directly to the tablet engine without interfering +with the X server's operation. Additionally, it will be possible to use +either the XInput configuration API or the tablet engine's API to access +that functionality. The preferred way would probably be an X server +extension, but today at least, that is outside of the scope of my interest. +Ask me again in another month. + +<P>The next issue is plug and play. The USB tablets have the +unfriendly habit of being unaccessible when they are unplugged and replugged. +Since the X server holds the device open, the tablets cannot reconnect to +their original position and get bumped to another event slot. The new +tablet engine will address the problem, and can probably even auto-probe +for the USB tablets. + +<P>Third, a GUI-based configuration tool should cap things off nicely. I +intend to replicate the functionality of the Windows configuration tool. +This includes diagnostics, pressure mappings, button settings, keyboard +mappings, etc. + +<H2>Your Role</H2> + +I have some expectations as well. If you have a Wacom tablet and you +use this code, I would appreciate it if you could test the beta releases +whenever a feature or issue that might affect you comes up. The sooner +that problems are resolved, the less likely you will be impacted by a +buggy production release. If you only use the stable builds, none of +the new features will be available to you until after they have been +thoroughly hammered out which could be awhile. Of course, nothing is +stopping you from using the stable builds; they are purposefully +made available for people and companies who need the tablet drivers as +stable as possible. +<A NAME="updates"> +<H2>1.2 - Most Recent Updates</H2> + +<P>This is a list of the updates for the past month or so. The <A HREF="#changelog">Changelog</A> has a more complete +list. + +<P>December 22, 2002 - Added support for the serial Intuos2 under xf86Wacom.c. +Fixed the sign convention on the mouse wheel for USB Intuos mice. Graphire +should be fixed, too. Added preliminary support for Volito to kernel driver. +Cleaned up wacdump and hid-core to compile on Debian Woody without +warnings. Added build help for Debian Stable users. Added ability to read +serial devices using wacdump. New package is 0.3.3-beta. + +<P>December 15, 2002 - Added initial framework for new driver code. Fixed +the mouse wheel for 4D mice. Fairly responsive, almost too responsive, in +fact. (Updated: quick fix to xf86Wacom.c to remove some debugging code that +got left in accidentally; new driver version is 0.3.2a, but the package is +the same.) + +<P>December 13, 2002 - Added development page. New beta build has fix for +2D mouse wheel. Wacdump displays relative data now. Build environment for +wacom_drv.o set up for new code. + +<P>December 9, 2002 - Added more information about the parameters accepted +by the InputDevice section, including the man pages. + +<P>November 24, 2002 - Moved beta tree to version 0.2.0. Added complete +configure-based environment. wacom.o and wacdump are built by default; +wacom_drv.o is available in binary and source form. + +<A NAME="success"> +<H2>1.3 - Success Stories</H2> +<P>Originally, this document dealt only with Redhat 8.0. I will extend this +to cover as many different distributions as possible, but I can personally +only test on machines at my disposal. At present, that means Redhat 7.2 and +8.0. If you have success stories for a distribution or version not mentioned, +let me know so I can update the document. + +<UL> + <LI> Redhat + <UL> + <LI> Redhat 8.0 - tested on kernel 2.4.18-18.8.0 + <LI> Redhat 8.0 - tested on kernel 2.4.18-17.8.0 + <LI> Redhat 7.3 - tested on kernel 2.4.18-17.7.x + <LI> Redhat 7.2 - tested on kernel 2.4.18-17.7.x + </UL> + <LI> Mandrake + <UL> + <LI> Mandrake 9.0 - tested on ??? + </UL> + <LI> Gentoo + <UL> + <LI> Gentoo 1.4rc - tested on 2.4.19 kernel, gcc3.2, XFree86 4.2.0 + </UL> + <LI> Debian + <UL> + <LI>Debian Woody - XFree86 4.2 on 2.4.20 + <LI>Debian Stable - XFree86 4.1 (see + <A HREF="#debwcmdrv">Building wacom_drv.o On Debian Stable</A>) + </UL> + <LI> Slackware + <UL> + <LI>Slackware 8.1, (standard release) + </UL> + <LI> SuSE + <UL> + <LI>Suse Linux 8.0, kernel 2.4.18, XFree 4.2 + </UL> +</UL> +<A NAME="howto"> +<H2>1.4 - How To Use This Document</H2> +<BLOCKQUOTE><DIV CLASS=diff> +Where the guide differs between distributions or packages, I have added a +comment like this to help distinguish between versions. +</DIV></BLOCKQUOTE> + +<P>In terms of document organization, if you have not figured it out already, +you can browse the document one page at a time, or you can click on the +ALL link and view the entire thing in one long page. + +<P>Next, this document was written with the +assumption that you are starting from +scratch with a relatively recent distribution of the Linux kernel. Also, +if you have already added lines to your X11 configuration file (XF86Config), +you should comment them out and restart X. Since we'll be stepping through +the entire process, we need X to ignore the tablet until we're ready. +Otherwise, X will just get in the way. + +<P>Finally, if you know what you're doing, you can leave your X settings +intact, print this out, switch to runlevel 3, and follow along from the +console. The newer packages should have an HTML and text version of this +document in the documentation directory. And of course, there's always +<A HREF=http://lynx.isc.org/>lynx</A> for the gurus out there, which is +why this document is written out in plain vanilla HTML. +<A NAME="resources"> +<H2>1.5 - Resources and References</H2> + +<TABLE BORDER=0 CELLSPACING=5> +<TR><TD VALIGN=TOP><A HREF="http://people.mandrakesoft.com/~flepied/projects/wacom/">Lepied's Wacom XInput driver for XFree86</A></TD><TD>This page contains the original driver and help for serial tablet users.</TD></TR> +<TR><TD VALIGN=TOP><A HREF="http://www.lepied.com/archive_wacom.html">Lepied's Wacom Mailing List Archive</A></TD><TD>This is an archive of Lepied's driver mailing list. +<I>Warning: posting directly to the mailing list page does not mail +to the list. Many posts are missed this way.</I></TD></TR> +<TR><TD VALIGN=TOP><A HREF="mailto:wacom-subscribe@lepied.com">Subscribe to Lepied's Wacom Mailing List</A></TD><TD>If you want to post to the list and receive notifications about driver updates, this link will subscribe you.</TD></TR> +<TR><TD VALIGN=TOP><A HREF="http://www.runkel-it.de/wacom_tablet_howto.html">Stefan Runkel's Wacom Tablet HOWTO</A></TD><TD>This page is a bit old, but it covers a lot of details that I mention only briefly. Good reading, even if it is possibly outdated.</TD></TR> +<TR><TD VALIGN=TOP><A HREF="http://www.wacom.com">Wacom Webpage</A></TD><TD>The Wacom web site contains little information about the Linux drivers, but you +can look at their nifty new tablets.</TD></TR> +<TR><TD VALIGN=TOP><A HREF="http://www.pxh.de/fs/graphire/">pxh: Graphire USB</A></TD><TD>This page is devoted to a Graphire-specific XFree86 driver.</TD></TR> +<TR><TD VALIGN=TOP><A HREF="http://homepages.feis.herts.ac.uk/~bt7al/wcm/">A Webpage Dealing with Cintiq (Wacom PL-500 in particular) on Linux</A></TD><TD>This website deals with getting the Cintiq running; I will be merging this code into +the kernel soon.</TD></TR> +<TR><TD VALIGN=TOP><A HREF="http://sem.best.vwh.net/linux/wacom/">Wacom Intuos USB Micro-HOWTO</A></TD><TD>An older webpage which may not be relevent anymore, +but was useful to me when I got started.</TD></TR> +<TR><TD VALIGN=TOP><A HREF="http://www.ecn.wfu.edu/~cottrell/thinkpad/tpa21m/wacom.html">Wacom Intuos USB on Linux</A></TD><TD>Another page that was useful to +me when I started.</TD></TR> +<TR><TD><A HREF="http://www.wacomeng.com/devsupport/index.html">Wacom Software Developer Support</A></TD><TD>This page contains some documentation for the +older tablets.</TD></TR> +</TABLE> +<A NAME="faq"> +<H2>1.6 - Frequently Asked Questions</H2> + +<P>This FAQ is based on the questions that I receive in email. If you have +a question that is not answered here, <A HREF="mailto:jej@j-arkadia.com?subject=WACOM">send me email directly.</A> Please add +the word "WACOM" to the subject line to help me distinguish your message from +the spam I normally receive.</P> + +<H2>Index</H2> +<UL> +<LI><A HREF=#WHO>Who is responsible for this code?</A> +<LI><A HREF=#SUPPORT>Is my tablet supported? What about my special stylus?</A> +<LI><A HREF=#OTHERS>What about my Penpartner/Graphire/Cintiq?</A> +<LI><A HREF=#SPURIOUS>How do I get my stylus to stop clicking randomly whenever I use it?</A> +<LI><A HREF=#FILTER>The old drivers had filtering, how do I enable it now?</A> +<LI><A HREF=#FEATURE>What happened to feature XYZ?</A> +<LI><A HREF=#BUSTED>How do I get the special features of my stylus to work?</A> +</UL> + +<BR> +<BR> +<BR> + +<A NAME=WHO> +<H3>Who is responsible for this code?</H3> + +<P>We are. That is, Linux users with Wacom tablets are responsible for +the development of this code. The source code +was written by at least two dozen individuals, none of whom still seem to +be directly involved in the project. I, John Joganic, took up the code in +November of 2002, largely to satisfy my own needs, but also because I believe +Linux needs to be more useful in the graphic and video industries. What I +ask of you, is to use the drivers and report the problems you encounter. + +<A NAME=SUPPORT> +<H3>Is my tablet supported? What about my special stylus?</H3> + +<P>Most recent tablets are supported as well all the styluses. The +<A HREF="#expect">Expectations</A> page has a +list of all the tablets and styluses that are known to work as well +as what you can expect out of these drivers now and in the future. + +<A NAME=OTHERS> +<H3>What about my Penpartner/Graphire/Cintiq?</H3> + +The USB kernel driver currently seems to support Penpartner, Graphire, +Graphire2, Intuos, Intuos2, and PL400-800. Based on user reports, the +Cintiqs may identify themselves as PL series tablets. +There is no kernel driver necessary for serial tablets. +The XFree86 wacom driver supports both USB and serial devices. It has +support for all the USB devices as well as virtually all of the older +tablet models as well. Depending on how the Cintiq identifies itself, +it may potentially work with the XFree86 driver as a serial device. There +are no guarantees as of yet, so I would not buy one (currently over $3000 +USD) on the assumption that it will work without doing some actual research. + +<A NAME=SPURIOUS> +<H3>How do I get my stylus to stop clicking randomly whenever I use it?</H3> + +<P>The "Threshold" setting in the InputDevice section of the XFree86 +configuration file determines the minimum pressure value before a touch +event is detected. Wacdump can show you what pressure values are actually +being received from the tablet, and setting the threshold to a value higher +than that should solve the problem. My stylus oscillates between 0 and 6, +even if the stylus is not touching the tablet. I use a threshold of 10.</P> + +<A NAME=FILTER> +<H3>The old drivers had filtering, how do I enable it now?</H3> + +<P>At some point, the maintainers of the wacom kernel driver ripped out +all the good features and moved them into the xf86Wacom driver. Not +everything works as well as one might like. The "Suppress" setting in +the InputDevice section specifies the tablet sensitivity. + +<A NAME=FEATURE> +<H3>What happened to feature XYZ?</H3> + +<P>As mentioned above, many of the features in the original kernel driver +were removed and transferred to the xf86Wacom driver. Not all the features +made the trip. The goal of this website is to have all the original features +working again as well as all the features available for the windows driver. +If your favorite feature is not working, drop me a line. I prioritize my +work based on my perception of user need. + +<A NAME=BUSTED> +<H3>How do I get the special features of my stylus to work?</H3> + +<P>There are many different types of stylus, mice, and cursors. All of them +work to varying degrees, so if you are running into trouble, send me email. +There are presently known issues with the mouse wheel on 2D mice (fixed in +0.3.1-beta), and some of the extra buttons on the 4D mice. While the values +are generally reported correctly to the driver, the xf86Wacom driver does not +interpret the information properly. The 4D mouse wheel for instance behaves +completely contrary to expectation (fixed in 0.3.2-beta). These problems will +be addressed in future production releases (and are available now in the beta +releases). + +<P>I do not have an airbrush tool at my disposal. If you have one and it +does not work for you, let me know. You could be beta-testing the code that +gets this working. +<A NAME="dev"> +<H2>1.7 - Development</H2> +<H2>CURRENT DEVELOPMENT</H2> + +<TABLE BORDER=0> +<TR><TD ALIGN=RIGHT>Date: </TD> +<TD>2002-12-22 (Sunday)</TD></TR> +<TR><TD ALIGN=RIGHT>Status: </TD> +<TD>Moving over to SourceForge, and testing new serial driver. +</TD></TR> +<TR><TD ALIGN=RIGHT>Estimated Time: </TD> +<TD>Ready on 2002-12-29 (Sunday)</TD></TR> +</TABLE> + +<H3>Completed</H3> +<UL> +<LI>Added Intuos2 serial to xf86Wacom.c. +<LI>Added serial tablet code to wacdump.c. +<LI>Added experimental Volito code to kernel. +<LI>Split ncurses and Linux input code to avoid namespace collision on Debian. +<LI>Corrected wheel sign for USB mice. +</UL> + +<H3>Immediate</H3> +<UL> +<LI>Get code onto SourceForge. +<LI>Intuos2 2D mouse does not work correctly under xf86Wacom.c +</UL> + +<H3>Secondary</H3> +<UL> +<LI>Getting test application for XF86 driver together +<LI>Testing ability to create new pointer devices on the fly +<LI>Testing ability to fork out new server process +</UL> + +<BR><BR><BR> + +<H2>KNOWN ISSUES</H2> + +There are plenty of known issues. If you have one, and it is not listed, +check the FAQ, and if it's not there either, let me know. + +<H3>wacom.o</H3> + +<UL> +<LI>2002-12-13: Tablet does not reinitialize properly when replugged (report #1). +</UL> + +<H3>wacdump</H3> + +<UL> +<LI>Nothing outstanding. +</UL> + +<H3>wacom_drv.o</H3> + +<UL> +<LI>2002-12-22: Intuos2 2D mouse does not work correctly under xf86Wacom.c +</UL> +<A NAME="theory"> +<H2>1.8 - Wacom Driver Theory of Operation</H2> +<P>Initially at least, the Wacom tablet is an HID compliant device, and when +first connected to the computer, will identify itself as such. +Unfortunately, this is not what you want because in this mode, you will not +get any of the fancy features. The hid-core.c and usbmouse.c files contain +exceptions for the wacom; when the device is detected, they ignore the +tablet. In this way, the more sophisticated wacom driver +has the opportunity to assume control. + +<P>The first thing that the driver does is register itself with the USB +subsystem and wait for work to do. When the user plugs the device in, or +the device is first detected, the USB subsystem shops the vendor and +device identifier around, checking it against different drivers. The +wacom driver takes responsibility for the tablet and then notifies the +event system that it will be providing data. It then asks the tablet to +switch from HID-compliant mode to "mode 2", a wacom-specific protocol which +allows for values like pressure, Z rotation, +and tilt. As information arrives, the wacom driver dutifully converts +the data into real-world values and hands it on to the event system. + +<P>From here, any usermode application can get access to the event data +by opening /dev/input/event0. A stream of events including mouse movements, +clicks, and proximity updates can be read from the device. One such +application that would be interested in this data is the X system itself. +An XInput driver (wacom_drv.o) collects that information, +applies various forms of filtering, and offers it to GUI-based applications +throught the XInput API. A graphics program like gimp can then get access +to the positions of various tools by querying X directly. + +<P>By breaking the responsibility for the data into three distinct levels, +we keep the kernel code simple and robust, the applications generalized, +and the fancy features commonly accessible to all GUI applications in the +X window system itself. This document walks down the entire data path from +the kernel driver to the gimp application. +<A NAME="starthere"> +<H1>2.0 - Getting It Together</H1> +<P>This section is devoted to preparing your system for the installation. +In some cases, your distribution may have automatically loaded certain +modules which you will now need to unload. Incidentally, some distributions +detect new hardware on boot. If you allow Redhat's "anaconda" for instance +to automatically configure (or remove) your tablet, it may undo some of the +settings you will make here. Until Redhat's installation program recognizes +Wacom tablets as non-HID devices by default, you are best off in my opinion +to <i>not</i> allow it to configure the device. Just a warning. +<A NAME="beforestart"> +<H2>2.1 - Before We Start</H2> +<P>From the beginning, let's make certain that we are on the +same page. First, unplug the USB Wacom. If you have Wacom related lines +in your XF86Config file, you should comment them out or remove them; then +restart X.<BR> + +<P>Unload the old modules. +<BLOCKQUOTE><PRE> +[jej@ayukawa jej]$ su - +[root@ayukawa root]# lsmod | grep wacom +<B>wacom</B> 8020 0 (unused) +input 5920 0 [<B>evdev</B> <B>wacom</B> mousedev keybdev hid] +usbcore 77024 1 [<B>wacom</B> usb-storage snd printer hid usb-uhci] +[root@ayukawa root]# rmmod wacom evdev +</PRE></BLOCKQUOTE> + +The kernel now knows nothing about the Wacom or how to handle event-based +input devices. If you get an error unloading (busy driver perhaps), try +rebooting.<BR> + +<P>Check the presence of the input devices in the /dev/input directory. +You should find approximately 114 devices with names such as event*, mouse*, +keyboard, js*, and ttyACM*. We are interested in two devices, mouse0 and +event0. + +<BLOCKQUOTE><PRE> +[root@ayukawa root]# cd /dev/input +[root@ayukawa input]# ls -la mouse0 event0 +crw------- 1 root root 13, 64 Aug 30 16:31 event0 +crw------- 1 root root 13, 32 Aug 30 16:31 mouse0 +[root@ayukawa input]# xxd mouse0 +xxd: mouse0: No such device +[root@ayukawa input]# xxd event0 +xxd: event0: No such device +</PRE></BLOCKQUOTE> + +<P>On some distributions, Mandrake 9.0 for instance, the /dev directory is +managed by devfs; Redhat 8.0 does not use it. When devfs is running, the +/dev/input directory and/or its contents may be missing if the driver is not +loaded. If you are not using devfs and the /dev/input directory or +mouse0/event0 devices are not present, your distribution may be lacking these +devices. This could be an oversight, or it could be an indication that your +distribution is too old. + +<P>If after executing <i>xxd mouse0</i> you <B>did not</B> +receive the "xxd: mouse0: No such device" message, then the input device +is still loaded; go back and try running <i>rmmod evdev</i> again. +Once the new drivers are compiled and installed, this command will stream +information from the Wacom tablet. mouse0 behaves like a PS/2 style mouse. +event0 provides extensive information about the tablet including position, +tilt, pressure, and buttons. It can also track at least two different tools +simultaneously (Intuos2).<BR> +<A NAME="download"> +<H2>2.2 - Downloading the Code</H2> +The file <A HREF="/wacom/wacpack-0.2.0.tar.gz">wacpack-0.2.0.tar.gz</A> is the stable package and contains +files that you will need to get the tablet working. The current beta +package <A HREF="/wacom/wacpack-0.3.3-beta.tar.gz">wacpack-0.3.3-beta.tar.gz</A> is also available and may be +used by people who are willing to risk an occasional kernel panic to help +test new features. I will never put a beta package on this site that I am +not running myself on my primary development machine, so you can be certain +that if there are any obvious show stoppers, they will be fixed before you +get to see them. + + +<H3>Stable files included for wacpack-0.2.0:</H3> +<TABLE BORDER=0 CELLSPACING=5> +<TR><TH ALIGN=LEFT>File</TH><TH ALIGN=LEFT>Comment</TH></TR> +<TR><TD VALIGN=TOP>configure</TD><TD>- configure script for distribution independent builds</TD></TR> +<TR><TD VALIGN=TOP>prebuilt/wacom_drv.o</TD><TD>- binary X11 driver for the wacom (26-j0.2.0)</TD></TR> +<TR><TD VALIGN=TOP>src/wacom.c</TD><TD>- a stable, working kernel driver (1.30-j0.2.0)</TD></TR> +<TR><TD VALIGN=TOP>src/wacdump.c</TD><TD>- a simple program for displaying tablet event data directly using ncurses; current version is 0.2.0</TD></TR> +<TR><TD VALIGN=TOP>src/usbmouse.c</TD><TD>- replacement kernel driver (2.4.19), use only if needed</TD></TR> +<TR><TD VALIGN=TOP>src/hid-core.c</TD><TD>- replacement kernel driver (2.4.19), use only if needed</TD></TR> +<TR><TD VALIGN=TOP>src/hiddev.c</TD><TD>- replacement kernel driver (2.4.19), use only if needed</TD></TR> +<TR><TD VALIGN=TOP>src/hid-input.c</TD><TD>- replacement kernel driver (2.4.19), use only if needed</TD></TR> +<TR><TD VALIGN=TOP>src/xf86Wacom.c</TD><TD>- source for wacom_drv.o; requires XF86 build environment to compile. Binary available in prebuilt directory.</TD></TR> +<TR><TD VALIGN=TOP>GPL</TD><TD>- the GNU General Public License, in case you did not already have twenty or more lying around</TD></TR> +</TABLE> + +<H3>Beta files included for wacpack-0.3.2-beta:</H3> +<TABLE BORDER=0 CELLSPACING=5> +<TR><TH ALIGN=LEFT>File</TH><TH ALIGN=LEFT>Comment</TH></TR> +<TR><TD VALIGN=TOP>configure</TD><TD>- configure script for distribution independent builds</TD></TR> +<TR><TD VALIGN=TOP>prebuilt/wacom_drv.o</TD><TD>- binary X11 driver for the wacom (26-j0.3.2a)</TD></TR> +<TR><TD VALIGN=TOP>src/wacom.c</TD><TD>- a stable, working kernel driver (1.30-j0.3.1)</TD></TR> +<TR><TD VALIGN=TOP>src/wacdump.c</TD><TD>- a simple program for displaying tablet event data directly using ncurses; current version is 0.3.2</TD></TR> +<TR><TD VALIGN=TOP>src/usbmouse.c</TD><TD>- replacement kernel driver (2.4.19), use only if needed</TD></TR> +<TR><TD VALIGN=TOP>src/hid-core.c</TD><TD>- replacement kernel driver (2.4.19), use only if needed</TD></TR> +<TR><TD VALIGN=TOP>src/hiddev.c</TD><TD>- replacement kernel driver (2.4.19), use only if needed</TD></TR> +<TR><TD VALIGN=TOP>src/hid-input.c</TD><TD>- replacement kernel driver (2.4.19), use only if needed</TD></TR> +<TR><TD VALIGN=TOP>src/xf86Wacom.c</TD><TD>- legacy source for wacom_drv.o; requires XF86 build environment to compile.</TD></TR> +<TR><TD VALIGN=TOP>src/wcm-beta.c</TD><TD>- beta source for wacom_drv.o; requires XF86 build environment to compile.</TD></TR> +<TR><TD VALIGN=TOP>src/wcm-beta.h</TD><TD>- beta source for wacom_drv.o; requires XF86 build environment to compile.</TD></TR> +<TR><TD VALIGN=TOP>GPL</TD><TD>- the GNU General Public License, in case you did not already have twenty or more lying around</TD></TR> +</TABLE> + +<H3>Stable Packages by Version and Date:</H3> +<TABLE BORDER=0 CELLSPACING=5> +<TR><TH ALIGN=LEFT>File</TH><TH ALIGN=LEFT>Date</TH><TH ALIGN=LEFT>Comment</TH></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.2.0.tar.gz">wacpack-0.2.0.tar.gz</A></TD><TD> - 2002-11-24</TD><TD>Configure-based package</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.1b.tar.gz">wacpack-0.1b.tar.gz</A></TD><TD> - 2002-11-06</TD><TD></TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.1a.tar.gz">wacpack-0.1a.tar.gz</A></TD><TD> - 2002-11-05</TD><TD></TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.1.tar.gz">wacpack-0.1.tar.gz</A></TD><TD> - 2002-11-04</TD><TD></TD></TR> +</TABLE> +<H3>Beta Packages by Version and Date:</H3> +<TABLE BORDER=0 CELLSPACING=5> +<TR><TH ALIGN=LEFT>File</TH><TH ALIGN=LEFT>Date</TH><TH ALIGN=LEFT>Comment</TH></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.3.3-beta.tar.gz">wacpack-0.3.3-beta.tar.gz</A></TD><TD> - 2002-12-22</TD><TD>Serial Intuos2, serial wacdump, Volito added, USB mouse wheel direction fixed, kernel 2.4.20 friendly</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.3.2-beta.tar.gz">wacpack-0.3.2-beta.tar.gz</A></TD><TD> - 2002-12-15</TD><TD>Fix for 4D mice and working beta driver shell</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.3.1-beta.tar.gz">wacpack-0.3.1-beta.tar.gz</A></TD><TD> - 2002-12-13</TD><TD>Fix for 2D mice and relative events in wacdump</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.3.0-beta.tar.gz">wacpack-0.3.0-beta.tar.gz</A></TD><TD> - 2002-11-24</TD><TD>Configure-based package</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.1g-beta.tar.gz">wacpack-0.1g-beta.tar.gz</A></TD><TD> - 2002-11-14</TD><TD>wacdump updated; displays ranges, cleaner output</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.1f-beta.tar.gz">wacpack-0.1f-beta.tar.gz</A></TD><TD> - 2002-11-14</TD><TD>build environment for xf86Wacom added</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.1e-beta.tar.gz">wacpack-0.1e-beta.tar.gz</A></TD><TD> - 2002-11-11</TD><TD>fixed range for I2-6x8</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.1d-beta.tar.gz">wacpack-0.1d-beta.tar.gz</A></TD><TD> - 2002-11-10</TD><TD>updated wacdump, fixed range for I2-12x12</TD></TR> +<TR><TD VALIGN=TOP><A HREF="/wacom/wacpack-0.1c-beta.tar.gz">wacpack-0.1c-beta.tar.gz</A></TD><TD> - 2002-11-09</TD><TD>updates to range, pressure, and mouse events</TD></TR> +</TABLE> +<A NAME="config"> +<H2>2.3 - Configuring the Package</H2> + +Versions of wacpack-0.2.0 and greater are based on GNU's configure script +based build environment. The upside is that more details are guessed by +default. The downside is that the default configuration builds only the +kernel driver and wacdump programs (which should cover almost everybody), +but if you need a different module built, you will have to reconfigure the +package first. This is mostly because different modules have different +dependencies, some of which are difficult to resolve. +A list of the options currently available in the stable package is +presented below. Remember that for every --enable option, there is also +an equivalent --disable option. + +<BR><BR> + +<TABLE BORDER=0 CELLSPACING=5> +<TR><TH ALIGN=LEFT>Option</TH><TH ALIGN=LEFT>Default</TH> + <TH ALIGN=LEFT>Builds</TH></TR> +<TR><TD>--enable-wacom</TD><TD><B>yes</B></TD> + <TD>wacom.o kernel driver</TD></TR> +<TR><TD>--enable-wacdump</TD><TD><B>yes</B></TD> + <TD>wacdump event monitor</TD></TR> +<TR><TD>--enable-hid</TD><TD>no</TD> + <TD>hid.o replacement kernel driver (not normally necessary)</TD></TR> +<TR><TD>--enable-usbmouse</TD><TD>no</TD> + <TD>usbmouse.o replacement kernel driver (not normally necessary)</TD></TR> +<TR><TD>--enable-wacomdrv</TD><TD>no</TD> + <TD>wacom_drv.o XFree86 driver (binary is available in prebuilt directory)</TD></TR> +<TR><TD>--enable-modver</TD><TD>best guess</TD> + <TD>enables kernel module versioning; usually guesses correctly, but + can be enabled or disabled if not</TD></TR> +<TR><TD>--with-kernel=dir</TD><TD>best guess</TD> + <TD>Specifies the kernel source directory if configure cannot guess correctly.</TD></TR> +<TR><TD>--with-xf86=dir</TD><TD>none</TD> + <TD>Specifies the XFree86 build directory.</TD></TR> +<TR><TD>--with-arch=arch</TD><TD>best guess</TD> + <TD>Specifies the architecture if configure guesses incorrectly.</TD></TR> +</TABLE> + +<P>Chances are high that you will be able to run the configure script without +any options and have it execute correctly. Here is a sample output of the +script on a Redhat 8.0 system: + +<BLOCKQUOTE><PRE> +[jej@ayukawa wacpack]$ ./configure +checking for a BSD-compatible install... /usr/bin/install -c +checking whether build environment is sane... yes +... +checking for processor type... <B>i686</B> +checking for kernel module versioning... <B>yes</B> +checking for kernel sources... <B>/usr/src/linux-2.4</B> +... + +---------------------------------------- + BUILD ENVIRONMENT: + architecture - i686 + module versioning - yes + kernel - yes + XFree86 - no + + BUILD OPTIONS: + wacom.o - yes + wacdump - yes + hid.o - no + usbmouse.o - no + wacom_drv.o - no +---------------------------------------- +</PRE></BLOCKQUOTE> + +<P>The processor type is determined by the script and used to build the +kernel modules. Similarily, the script attempts to discover if the kernel +is using module versioning by looking for version numbers in the usb.o +module of the currently active kernel. This may not work well on all +distributions, so if you get a line like "checking for kernel module versioning... confused" or "maybe; assuming yes", you'll have to manually enable or +disable the option if it guesses incorrectly. + +<P>The kernel source directory is assumed to be either /usr/src/linux-2.4 +or /usr/src/linux. If your kernel sources are elsewhere, you will need to +specify the directory with the --with-kernel option. + +<P>Generally, you will not need to build wacom_drv.o since it ships in +binary form in the prebuilt directory. It is compiled with glibc with +gcc 3.2 on a Redhat 8.0 system; if this will not work for you, building +from source may be your only option. See the +<A HREF="#builddrv">Building wacom_drv.o From Scratch</A> page +for more information. + +<P>The configure script is pretty simple, so if you find any problems and +are inclined to fix them, please send me your patch so I can include the +change in future releases. + +<P>The following sample command-line will build everything but wacdump +while disabling module versioning: + +<BLOCKQUOTE><PRE> +[jej@ayukawa wacpack]$ ./configure --enable-hid --enable-usbmouse \ + --enable-wacomdrv --with-xf86=/usr/src/redhat/BUILD/XFree86-4.2.0 \ + --disable-modver --disable-wacdump +checking for a BSD-compatible install... /usr/bin/install -c +checking whether build environment is sane... yes +... +checking for processor type... i686 +checking for kernel module versioning... <B>yes</B> +checking for kernel sources... /usr/src/linux-2.4 +checking for valid XFree86 build environment... <B>ok</B> +... +---------------------------------------- + BUILD ENVIRONMENT: + architecture - i686 + module versioning - <B>no</B> + kernel - yes + XFree86 - <B>yes</B> + + BUILD OPTIONS: + wacom.o - yes + wacdump - <B>no</B> + hid.o - <B>yes</B> + usbmouse.o - <B>yes</B> + wacom_drv.o - <B>yes</B> +---------------------------------------- +</PRE></BLOCKQUOTE> + +<P>Notice that the configure script guessed module versioning was enabled +by default, but was disabled by the command-line option --disable-modver. +Similarly, the wacdump program which is enabled by default was also disabled. +All the kernel modules and the XFree86 wacom driver are enabled. + +<P>If the configure script fails to find something that it is looking for, +it may disable some options that you previously enabled on the command-line. +If this happens, check the output for a warning like the following: + +<BLOCKQUOTE><PRE> +*** +*** WARNING: +*** Unable to compile wacom_drv.o without XF86 build environment +*** wacom_drv.o will not be built +*** +</PRE></BLOCKQUOTE> + +<P>In this particular case, the XFree86 driver was enabled, but the +--with-xf86 option was not specified. Without the build environment, +the module cannot be compiled and was consequently disabled. +<A NAME="kernel"> +<H1>3.0 - The Kernel Driver</H1> +<P>The kernel driver for the Wacom Intuos that ships with the +Linux kernel through 2.4.19, and in particular Redhat 8.0, does +not work correctly. The bug is pretty simple and you can fix it yourself. +Optionally, you can compile the version that I provide directly and swap +the new module for the old one. This avoids having to rebuild the entire +kernel.</P> + +<P>Kernel modules must be recompiled for each new kernel so I can't +just provide binaries. By the time you read this, my present kernel +will be entirely out of date with yours. Additionally, this fix will +appear in 2.4.20, so until you upgrade your kernel to that point, each +time you upgrade to a version less than that, you will need to recompile this +driver and swap it for the old one. In many cases, distributions like +Redhat will "back-port" the fixes to older kernels, so keep an eye out +for the change. Your driver may already be fixed.</P> +<A NAME="newwacom"> +<H2>3.1 - Updated wacom.c</H2> +The wacom.c driver that is available in <A HREF="/wacom/wacpack-0.2.0.tar.gz">wacpack-0.2.0.tar.gz</A> has a +number of updates which cannot be found in many distributions yet. + +<P>First and foremost, it fixes an error that shipped with +the 2.4.19 kernel and was subsequently fixed in 2.4.20. This bug fix was +made available in the Redhat 8.0 back-ported kernel version 2.4.18-18.x.y. +Other distributions may not yet be updated. + +<P>This driver contains several changes provided by Wacom UNIX developer +Ping Cheng. Items include revised pressure code, better handling of 2D +mice and lens cursors. + +<P>The driver also contains additional debugging code which may be useful in +diagnosing data flow problems of which there are many, unfortunately. +<A NAME="buildwacom"> +<H2>3.2 - Building wacom.c</H2> +<P>Building the kernel driver requires having the kernel source. If you +are running on Redhat, you can get it by installing the kernel-source RPM. +At the time of this writing, I have the following RPMs installed: + +<BLOCKQUOTE> +kernel-2.4.18-18.8.0<BR> +kernel-2.4.18-14<BR> +kernel-source-2.4.18-18.8.0<BR> +</BLOCKQUOTE> + +On a Redhat 8.0 machine, the kernel source is installed to /usr/src/linux-2.4. You +will need to know where your source is installed to continue with the next +step. After unpacking the <A HREF="/wacom/wacpack-0.2.0.tar.gz">wacpack-0.2.0.tar.gz</A> file, configure the +package. If your kernel is not located at either /usr/src/linux-2.4 or +/usr/src/linux, use the --with-kernel option as shown below. Otherwise, +just run <I>configure</I> without any options. + +<BLOCKQUOTE><PRE> +[jej@ayukawa wacpack]$ ./configure --with-kernel=/home/jej/src/linux +... +checking for valid kernel source tree... <B>ok</B> +... +---------------------------------------- + BUILD ENVIRONMENT: + architecture - i686 + module versioning - yes + kernel - <B>yes</B> + XFree86 - no + + BUILD OPTIONS: + wacom.o - yes + wacdump - yes + hid.o - no + usbmouse.o - no + wacom_drv.o - no +---------------------------------------- +</PRE></BLOCKQUOTE> + +<P>If you did not receive an error or warning and the BUILD ENVIRONMENT +summary shows "kernel - yes" then your kernel source tree was detected. +If not, then you will need to look at the error or warning and determine +the correct source directory. Now is also a good time to verify that the +architecture was correctly detected. The --with-arch option allows you +to change the value. + +<P>Another detail that you need to be concerned about is module +versioning. If it reads "yes" or "no" then the configure script +has determined the correct setting. If it says "unknown," then it will +default disabled, and you may have to enable it manually with --enable-modver. +If module versioning is disable when it should be enabled, depmod will +complain about missing symbols. If it is enabled when it should be disabled, +the code may not compile, and it almost certainly will not load properly. +If in doubt, leave it disabled for now and enable it later if depmod complains. + +<P>To build the driver, just run <i>make</i>. The +output will be a file called wacom.o in the src directory. +This is your replacement driver. +<A NAME="testwacom"> +<H2>3.3 - Testing If wacom.o Will Load</H2> +<P>Before we install the wacom driver, we need to test that it will load +properly. We do this by loading the driver manually. WARNING: there is a +small chance that this will bomb your kernel, so we run <i>sync</i> to write +all the stale buffers to the disk. People using ext3 have little to worry +about, but it's always good to be prepared for the worst. At the very least, +save your work. + +<BLOCKQUOTE><PRE> +[jej@ayukawa wacom]# su - +[root@ayukawa root]# cd /home/jej/src/wacom/src +[root@ayukawa src]# sync +[root@ayukawa src]# rmmod wacom +[root@ayukawa src]# insmod wacom.o # for those about to rock, we salute you. +</PRE></BLOCKQUOTE> + +<P>Well, if you did not bomb, then good. And if you did, well, sorry. So +far, we have not had any reports of this happening, so please send in yours. + +<P>If you get errors inserting the module, then you may need to reconfigure +and build with module versioning disabled. If it loads without a hitch, +move on to the next part. +<A NAME="installwacom"> +<H2>3.4 - Installing wacom.o</H2> +<P>To install or not to install, that is the question. +Since the driver is in memory, you can pretty much use it this way +throughout the rest of this document. Anywhere you see +<i>modprobe wacom</i>, you'll instead need to <i>insmod wacom.o</i>. +You'll also need to be careful that you use the correct path to wacom.o +since insmod will load the driver from the kernel modules directory if it +fails to find the module you specified. The result is that you'll +be using the wrong driver. + +<P>Why would you not install the driver? Well, for one, you may be using +a beta driver, and if the system crashes (you get an Oops or things come +unglued in other ways), it would be nice to reboot and have the original +drivers load instead. + +<P>When should I install the driver? When you're comfortable that the +driver will not crash your system. The wacpack tarballs will be marked +as beta if I am not absolutely convinced that they are stable. On the +other hand, the new features will be in beta tarballs before they are +marked stable, so there you have it. For the first time through this document, +I would recommend installing the files found in the stable tarball. If you +really know what you're doing, just load the drivers manually like in the +previous section <A HREF="#testwacom">Testing If wacom.o Will Load</A>. + +<P CLASS=diff>On some distributions, Mandrake included, the wacom.o driver +that appears in the kernel modules directory appears to be compressed. If +you cannot find wacom.o using the method below, try locating wacom.o.gz +instead. People who encountered this problem were able to run gzip on the +module and copy that instead.</P> + +<P>Installing the driver requires knowing where it belongs. A +little research will help here. By using the <i>locate</i> command, you +can find all copies of the original driver on the computer.</P> + +<BLOCKQUOTE><PRE> +jej@ayukawa wacom]$ locate wacom.o +/lib/modules/2.4.18-14/kernel/drivers/usb/<B>wacom.o</B> +/lib/modules/2.4.18-18.8.0/kernel/drivers/usb/<B>wacom.o</B> + +[jej@ayukawa wacom]$ uname -r +<B>2.4.18-18.8.0</B> +</PRE></BLOCKQUOTE> + +On this computer, there are two kernels installed. <I>uname</i> identifies +the currently active kernel as 2.4.18-18.8.0. The correct driver to replace +is therefore at /lib/modules/2.4.18-18.8.0/kernel/drivers/usb/wacom.o. You +will need to be root to replace this file, and it is a very good idea to +make a backup copy. + +<BLOCKQUOTE><PRE> +[jej@ayukawa wacom]$ su - +[jej@ayukawa root]# cd /lib/modules/2.4.18-18.8.0/kernel/drivers/usb +[jej@ayukawa usb]# cp <B>wacom.o</B> /home/jej/src/wacom/<B>wacom_old.o</B> +[jej@ayukawa usb]# cp /home/jej/src/wacom/<B>wacom.o</B> <B>wacom.o</B> +</PRE></BLOCKQUOTE> + +Here, I've saved the original to wacom_old.o and copied my new driver over it. +Substitute directory names as appropriate. + +<P>Finally, it is always a good thing to update the module dependencies. +This is where you find out if the module was compiled without kernel module +versioning. The following command, even if it generates errors is relatively +benign. If it fails, then there is no harm done. It just means that you will +have to load modules in the correct order since the system will not be able to +guess for you. + +<BLOCKQUOTE><PRE> +[jej@ayukawa usb]# depmod -e +</PRE></BLOCKQUOTE> + +If you get no errors and no output, everything is fine, and the module was +compiled, linked, and installed properly. If you received unresolved symbols +like usb_set_idle or printk, then you need to reconfigure with module +versioning enabled and recompile. Or not. As I said, it is not critical. +<A NAME="loadwacom"> +<H2>3.5 - Loading the wacom Driver</H2> + +<P>If you have installed the driver, now is the time to test whether it will +load when needed. If you have not installed it, but are instead using insmod, +substitute <i>insmod mydir/wacom/wacom.o</i> where you see +<i>modprobe wacom</i> below. It is important that you use the correct +wacom.o file, the one you just built, since insmod may load the old driver +if it cannot find the one you have specified. + +<P>I am assuming +that you are using a USB device, so you will also need to modprobe for usb-uhci. +The important features of the stylus are available through the Linux event +drivers, so load those too. With some clever additions to /etc/modules.conf, +you can probably have these automatically loading for you whenever the tablet +is activated. So far, I haven't figured out how to make this work reliably, +so I've added the commands to my /etc/rc.d/rc.local file. + +<BLOCKQUOTE><DIV CLASS=diff> +On Redhat 7.2 and possibly on other older distributions, mousedev and input +are not loaded by default. Without them, the wacom driver may fail to load, +due to unresolved symbols, and the mouse driver may not receive wacom mouse +events, even if loaded afterwards. +</DIV></BLOCKQUOTE> + +<BLOCKQUOTE><PRE> +[jej@ayukawa usb]# rmmod wacom +[jej@ayukawa usb]# modprobe usb-uhci +<SPAN CLASS=diff>[jej@ayukawa usb]# modprobe input</SPAN> +<SPAN CLASS=diff>[jej@ayukawa usb]# modprobe mousedev</SPAN> +[jej@ayukawa usb]# modprobe wacom +[jej@ayukawa usb]# modprobe evdev +</PRE></BLOCKQUOTE> + +<P>Check the system log for status messages pertaining to the wacom. +Here's a copy of the messages from my version of the driver. + +<BLOCKQUOTE><PRE> +[jej@ayukawa usb]# grep -i wacom /var/log/messages | tail +Dec 22 21:23:35 ayukawa kernel: usb.c: registered new driver wacom +Dec 22 21:23:35 ayukawa kernel: wacom.c: <B>v1.30-j0.2.0</B> Vojtech Pavlik <vojtech@suse.cz> +Dec 22 21:23:35 ayukawa kernel: wacom.c: USB Wacom Graphire and Wacom Intuos tablet driver (MODIFIED) +</PRE></BLOCKQUOTE> + +<P>The original driver was version 1.21.3. This version number +is 1.30-j0.2.0 so the correct driver was loaded.</P> +<A NAME="testtablet"> +<H2>3.6 - Testing Tablet Detection</H2> +<P>In this section we will plug the tablet into the computer and determine +which driver, if any, claims control over the tablet. There are at least +three drivers that are interested: 1) hid.o which may think it is an HID +device, 2) usbmouse.o which may think it is an HID mouse, and 3) the wacom +driver which should identify the tablet as its own. Any number of problems +may be experienced here, so be prepared to spend some time looking at the +logs. A good way to do this, if you are running X, is to open a separate +xterm, su to root, and run <i>tail -f /var/log/messages</i>. The console +window will stream anything that is appended to the log. +<A NAME="plugin"> +<H3>3.6.1 - Plug It In</H3> + +<P>Test the tablet detection by plugging the wacom into the USB port and +checking /var/log/messages again. You should see a flurry of activity. The +exact output depends a lot on your particular kernel and distribution. + +<BLOCKQUOTE><PRE> +This is Redhat 8.0 (2.4.18-17.8.0):<BR> +[jej@ayukawa usb]# tail /var/log/messages +Dec 22 21:26:11 ayukawa kernel: hub.c: USB new device connect on bus2/2, assigned device number 2 +Dec 22 21:26:11 ayukawa kernel: <B>input0: Wacom Intuos2 12x12</B> on usb2:2.0 +Dec 22 21:26:14 ayukawa /etc/hotplug/usb.agent: Setup wacom hid for USB product 56a/44/115 +Dec 22 21:26:14 ayukawa /etc/hotplug/usb.agent: Setup <B>mousedev</B> for USB product 56a/44/115 +</PRE> +<PRE CLASS=diff> +And here it is again on Redhat 7.2 (2.4.18-17.7.x):<BR> +[jej@sasami root]# tail /var/log/messages +Dec 22 21:28:38 sasami kernel: hub.c: USB new device connect on bus1/1, assigned device number 2 +Dec 22 21:28:38 sasami kernel: <B>input0: Wacom Intuos2 12x12</B> on usb1:2.0 +Dec 22 21:28:39 sasami kernel: usb.c: registered new driver hiddev +Dec 22 21:28:39 sasami kernel: usb.c: registered new driver hid +Dec 22 21:28:39 sasami kernel: hid-core.c: v1.8.1 Andreas Gal, Vojtech Pavlik <vojtech@suse.cz> +Dec 22 21:28:39 sasami kernel: hid-core.c: USB HID support drivers +Dec 22 21:28:39 sasami kernel: mice: <B>PS/2 mouse</B> device common for all mice +</PRE> +</BLOCKQUOTE> + +<P>If all went well like above, the USB device was successfully detected and +handled by the wacom driver; secondly, it was assigned input0. This +presumably means that information like pressure and tilt will be received on +/dev/input/event0. Lastly, the tablet was setup with mousedev or the generic +PS/2 USB mouse driver, so that information should appear on /dev/input/mouse0. + +<P>If instead you got any of the following lines in your log, the wacom +driver <i>did not get control</i>. Either hid or usbmouse did. + +<BLOCKQUOTE><PRE> + <B>input0,hiddev0: USB HID v1.00 Mouse</B> [Tablet XD-0608-U] on usb1:5.0 + <B>input0: Tablet XD-1212-U</B> on usb2:2.0 +<PRE></BLOCKQUOTE> + +<P>The next section describes what to do if the wacom driver did not get +control. Otherwise, skip on to +<A HREF="#viewdata">Viewing the Raw Data</A>. + +<A NAME="noctrl"> +<H3>3.6.2 - Wacom Driver Does Not Get Control</H3> + +<P>If the wacom driver did not get control, and instead you see +"Tablet XD-1212-U" or similar in its place, then this section is for +you. Otherwise, consider yourself lucky and skip on to the next section, +<A HREF="#viewdata">Viewing the Raw Data</A>. For the unlucky, +there is a good possibility that your kernel is slightly out of date and +that your usbmouse.c and hid-core.c files need some help. + +<P>When the Tablet XD-1212-U driver comes up, either usbmouse.o or hid.o +has discovered your tablet, decided it was an HID device, and grabbed it. +This is bad news because HID tablets seem to only output relative +coordinates and nothing about pressure. You are in good company though +because a number of people have reported this, even on some fairly +up-to-date distributions. Now let's fix it. + + +<A NAME="buildhid"> +<H3>3.6.3 - Building usbmouse.o and hid.o</H3> + +<P>In the <A HREF="/wacom/wacpack-0.2.0.tar.gz">wacpack-0.2.0.tar.gz</A> file, you will find the 2.4.19 versions +of usbmouse.c and hid-core.c files which have special exceptions for wacom. +These files are not built by default, so you will need to reconfigure the +package and run make again. + +<DIV CLASS=diff> +At least one person has reported errors building these files on Debian Woody +running kernel version 2.4.20. This is not surprising since the +source code came from 2.4.19, but I will do what I can to get these drivers updated in the near future. +</DIV> + +<BLOCKQUOTE><PRE> +[jej@ayukawa wacom]$ ./configure --enable-hid --enable-usbmouse +... + BUILD OPTIONS: + hid.o - <B>yes</B> + usbmouse.o - <B>yes</B> +... +[jej@ayukawa wacom]$ make +[jej@ayukawa wacom]$ cd src +[jej@ayukawa src]$ ls -la hid.o usbmouse.o +-rw-rw-r-- 1 jej jej 26590 Dec 22 18:56 hid.o +-rw-rw-r-- 1 jej jej 4848 Dec 22 18:56 usbmouse.o +</PRE></BLOCKQUOTE> + +<P> First off, if the code does not build, and you cannot chase the problem +down yourself, send me a line. If your kernel is much older than 2.4.18, +then you may have to resign yourself to the fact that the 2.4.19 drivers +are not going to work for you. I would recommend upgrading your kernel. Some +people are already running kernel version 2.4.20 and have encountered +problems building as well. I will update these drivers to 2.4.20 as soon as possible. + +<P>Rather than copy these files to the kernel modules directory right +away, it would be wise to test whether they get the job done. We should also +verify that they load properly. + +<BLOCKQUOTE><PRE> +[jej@ayukawa src]$ su - +[root@ayukawa root]# cd /home/jej/src/wacom/src +[root@ayukawa src]# rmmod hid usbmouse wacom +[root@ayukawa src]# insmod hid.o +[root@ayukawa src]# insmod usbmouse.o +[root@ayukawa src]# insmod wacom.o +</PRE></BLOCKQUOTE> + +As for unloading the old drivers, none of them should be busy, but if you +run into trouble, you can use <i>lsmod</i> to check the dependencies +and try again. Now let's check how the wacom driver managed: + +<BLOCKQUOTE><PRE> +[jej@ayukawa wacom]# tail -15 /var/log/messages +Dec 22 21:07:38 ayukawa kernel: usbmouse.c: <B>mouse_init (MODIFIED DEBUG)</B> +Dec 22 21:07:38 ayukawa kernel: usb.c: registered new driver usb_mouse +Dec 22 21:07:39 ayukawa kernel: hid-core.c: <B>hid_init (MODIFIED DEBUG)</B> +Dec 22 21:07:39 ayukawa kernel: usb.c: registered new driver hiddev +Dec 22 21:07:39 ayukawa kernel: usb.c: registered new driver hid +Dec 22 21:07:44 ayukawa kernel: usb.c: registered new driver wacom +Dec 22 21:07:44 ayukawa kernel: usbmouse.c: <B>usb_mouse_probe: ignoring wacom</B> +Dec 22 21:07:44 ayukawa kernel: input0: <B>Wacom Intuos2 12x12</B> on usb2:8.0 +Dec 22 21:07:44 ayukawa kernel: wacom.c: v1.30-j0.2.0 Vojtech Pavlik <vojtech@suse.cz> +Dec 22 21:07:44 ayukawa kernel: wacom.c: USB Wacom Graphire and Wacom Intuos tablet driver <B>(MODIFIED-DEBUG)</B> +</PRE></BLOCKQUOTE> +<P>If everything is working correctly, the log should look similar to the +lines above. I removed some of the extraneous information, so if you have +more than this, you should be fine. The most important thing here is that +the usb_mouse_probe detected and ignored the wacom. It actually gets +several shots at it, but I removed the redundant lines. For the final test, +unplug and replug the tablet. Now check the log. The wacom driver should +be calling the shots from here on out. + +<P>If all is well, and you are installing modules rather than <i>insmod</i>'ing +them as you go, then now is the time to copy the usbmouse.o and hid.o +drivers to their proper places in the kernel modules directory. +<SPAN CLASS=diff>As before with some other distributions, if the files in the +kernel module directory are compressed, you'll need to run gzip on the .o +files to get .o.gz files.</SPAN> +Use <i>locate</i> as you did for the wacom.o file if necessary and don't +forget to backup the originals. +<A NAME="morehid"> +<H3>3.6.4 - More about hid.o</H3> + +<P>Incidentally, the hid.o file may not be necessary, but I've included the +source to build it anyway because hid-core.c contains a lot of wacom related +code. If the hid.o driver gets called upon to control the wacom, and it is +compiled in debug mode, you may see a line similar to the following in the +log: + +<BLOCKQUOTE><PRE> +Dec 22 21:11:44 ayukawa kernel: usb_hid_configure: configuring device 56A.44 (WACOM), quirk is 4 (IGNORE) +</PRE></BLOCKQUOTE> + +If you get this message, but with a different quirk, or if the wacom driver +still is not getting control by this point, there is a possibility that your +tablet is not listed in any of the driver files. If so, continue on to the +next section. Otherwise, if you are up and running, skip to +<A HREF="#viewdata">Viewing the Raw Data</A>. +<A NAME="newtablet"> +<H3>3.6.5 - Unknown Tablet?</H3> +To determine whether your device is listed in the driver, we need to determine +the device identifier. It can be discovered by doing the following: + +<BLOCKQUOTE><PRE> +[root@ayukawa root]# grep -i 56a /var/log/messages | tail -10 +Dec 22 21:03:09 ayukawa /etc/hotplug/usb.agent: Setup mousedev for USB product <B>56a/44/115</B> +Dec 22 21:27:48 ayukawa kernel: usb.c: USB device 2 (vend/prod <B>0x56a/0x44</B>) is not claimed by any active driver. +</PRE></BLOCKQUOTE> + +<P>In this case, the tablet identifier is 56a/44 (also written 0x56a/0x44). +The model is determined by the second number; the Intuos2 12x12 is 0x44 for +instance. + +<P>In the wacom.c file, you will find a table called "wacom_ids." Look for +your device identifier. Only the identifiers listed are handled by the +wacom driver, so if it is missing, it needs to be added. Also look in +hid-core.c. It has a table called "hid_blacklist" which identifies devices +to ignore. The code is a little strange, but PENPARTNER is 0x00, GRAPHIRE is +0x10, INTUOS is 0x20, PL is 0x30, and INTUOS2 is 0x40. A table entry like +INTUOS2 + 4 would therefore be 0x44. + +<P>If you've gotten this far, and still cannot get it to work, send me email +with your device identifier and as much of an explanation of where things did +and did not work as described. I'll see what I can do about at least finding +out why it did not work. Then we can go on to solutions. + +<P>The next section assumes you have things working up to this point. + +<A NAME="viewdata"> +<H2>3.7 - Viewing the Raw Data</H2> +<P>View the raw data from the tablet, +by returning to the /dev/input directory and streaming the data directly +from the device. Be patient here because this is where a lot of people +are getting stuck. + +<BLOCKQUOTE><PRE> +[root@ayukawa usb]# cd /dev/input +[root@ayukawa input]# xxd event0 +0000000: e65d c33d 597d 0100 0100 4101 0100 0000 .].=Y}....A..... +0000010: e65d c33d 5c7d 0100 0400 0000 b701 2800 .].=\}........(. +0000020: e65d c33d d9bb 0100 0100 4101 0000 0000 .].=......A..... +0000030: e65d c33d dcbb 0100 0400 0000 b701 2800 .].=..........(. +(Ctrl-C) +</PRE></BLOCKQUOTE> + +First off, you have to move the mouse or tap the pen to get any output. +Second, you might not get anything at all. Don't panic. This +seems to happen occasionally. Unfortunately, the work-around is neither +straightfoward nor foolproof. I am currently talking with Wacom developers +about this and hopefully we'll get it solved soon. + +<P>If absolutely no output occurs, you should check your messages +file for the following line: + +<BLOCKQUOTE><PRE> +[jej@sasami root]# tail /var/log/messages +Dec 22 17:31:06 sasami kernel: <B>wacom_intuos_irq: received unknown report #1</B> +Dec 22 17:31:09 sasami last message repeated 195 times +</PRE></BLOCKQUOTE> + +This is a common failure mode of the driver, and I am investigated the cause. +Each time this has happened to me though, I <I>have</I> been +able to get it to work. Simply unload and reload the driver. + +<BLOCKQUOTE><PRE> +[jej@sasami root]# rmmod wacom +[jej@sasami root]# modprobe wacom +[jej@sasami root]# tail /var/log/messages +Dec 22 17:31:31 sasami kernel: usb.c: deregistering driver wacom +Dec 22 17:31:34 sasami kernel: usb.c: registered new driver wacom +Dec 22 17:31:35 sasami kernel: input0: Wacom Intuos2 12x12 on usb1:2.0 +Dec 22 17:31:35 sasami kernel: wacom.c: v1.30-j0.2.0 Vojtech Pavlik <vojtech@suse.cz> +</PRE></BLOCKQUOTE> + +<P>Apparently, the device driver and the tablet occassionally get out of +sync with the tablet thinking it's still in HID mode when in fact it +should be in "mode 2." By unloading and reloading the driver manually, +the initialization code has another opportunity to get it right. Try the +<I>xxd /dev/input/event0</I> again. This time, it <I>should</I> work. If +not, send me some email. I am trying to create a repeatable test case +for the Wacom developers, and your problem may be just what we need to +get this bug fixed. + +<P>Incidentally, if you have a program running that is connected to +/dev/input/event0 (like X or wacdump for instance), it is possible that +the tablet will not reattach back to the same event. I have seen the +wacom reattach to /dev/input/event1 when unloading and reloading the +wacom driver with wacdump running for instance. + +<P>You should also try running <i>xxd</i> on /dev/input/mouse0. +You should get streams of data when the mouse and pen are moved +around the surface of the tablet. It is this device that X will +look at for mouse movement. Use Ctrl-C to exit xxd. +<A NAME="wacdump"> +<H1>4.0 - wacdump</H1> +<P>wacdump.c parses and displays the raw event output. You must +be root to run it unless you've set the permissions on /dev/input/event0 +such that you can read it. This program can run simultaneously with X, +but it's best if X has not be configured for the tablet yet. +wacdump is enabled by default in the configure script, so it should be +built automatically when you run <I>make</I>. + +<P>The command line usage of wacdump (wacpack-0.2.0) is pretty simple: + +<BLOCKQUOTE><PRE> +Usage: wacdump [-d device] + -?, -h, --help - usage + -d, --device device - use specified device +</PRE></BLOCKQUOTE> + +<P>If you know that the tablet is attached to a device other than +/dev/input/event0, you can override it with the -d option, but otherwise, +the default is correct. If you connect the /dev/input/event0 and get an +end-of-file error, then the wacom is probably attached to a different event. + +<P>When you run wacdump, nothing will happen until you place a mouse or +pen near the surface. After that, the screen will clear and you will see +a screen similar to the following output: + +<BLOCKQUOTE><PRE> +Wacom Intuos2 12x12 bus=3, vndr=56A, prd=44, ver=115 + + KEY=00000000.00000000 0000 0000 00000000 + ABS=00000000.00000000 0000 0000 00000000 + MSC=3DD488B5.0000F385 0004 0000 0028102E + + X=+17623 (+00000 .. +30480) Y=+10730 (+00000 .. +31680) + RZ=-00058 (-00900 .. +00899) THROTTLE=+00000 (-01023 .. +01023) + WHEEL=+00000 (+00000 .. +01023) PRESSURE=+00005 (+00000 .. +01023) +DISTANCE=+00007 (+00000 .. +00015) TILT_X=+00094 (+00000 .. +00127) + TILT_Y=+00122 (+00000 .. +00127) + + WHEEL=+00000 + SERIAL=0028102E + + LEFT= RIGHT= MIDDLE= SIDE= + EXTRA= PEN= RUBBER= BRUSH= + PENCIL= AIR= MOUSE= LENS= + TOUCH= STYLUS= STYLUS2= +</PRE></BLOCKQUOTE> + +<P>This output comes from version 0.3.2 of the wacdump program. + +<P>The first row displays the model number, vendor id, product id, and +revision number. + +<P>The next block shows events as they arrive. Only the +event types that the device claims to generate are displayed. +The quickly changing numbers are timestamps. + +<P>The middle block shows absolute positions of various device parameters. +It also shows the ranges that the device claims to provide. If your device +exceeds these ranges, please let me know so I can fix the +driver. All the available positions are displayed, even though some of +them are not available for certain pointers. Different tablets will have +different capabilities; these are capabilities registered specifically for +your tablet by the wacom.o kernel driver. + +<P>The second WHEEL value is from the relative events channel. Presently, +only the beta drivers report relative events. + +<P>The serial number of the current stylus is displayed next. When two +pointers are present, the serial number will alternate. This is especially +true of Intuos2 models and possibly less accurate of other models. + +<P>Lastly, the button positions are displayed. Some mice have more buttons +than others; the ones displayed are the ones that the tablet offers. Buttons +like "Pen" or "Airbrush" are displayed as "Down" when the tool is in proximity +to the surface. + +<P>Whenever you are curious if the device is working properly, wacdump is +a useful tool for determine exactly what is coming from the tablet. You can +run it in conjunction with X or without. The display is constrained to 80x25 +so it can also be used on the main console terminal. +<A NAME="x11"> +<H1>5.0 - Configuring X11</H1> +Two steps must be completed to get X to recognize the Wacom. First, +you need to add some lines to XF86Config to inform X of the tablet's +existence. Second, you need to update the XInput driver that pertains +to the Wacom since the one that ships with XFree86 is not very functional. +Neither driver holds a candle to the windows driver though, so you'll +have to take what you get for the time being. Updates to the XFree86 driver +are available in the stable and beta releases on the +<A HREF="#download">Downloading the Code</A> page. You may +also check <A HREF="http://people.mandrakesoft.com/~flepied/projects/wacom/">Lepied's Wacom Site</A> which +has source code up to version 26 as of this writing. +<A NAME="inputdev"> +<H2>5.1 - Adding the InputDevices</H2> +<P>The X Window system identifies the stylus and eraser on your tablet as +XInput devices. Applications that want to know the absolute position of +your stylus can request that information directly, and this generally +bypasses whatever the mouse happens to be doing at the time. This design +is not without problems though. For one, changing your setup or adding +a new tool requires making changes to the XF86Config file and restarting +X. This should be fixed in the future. + +<P>For now however, add the InputDevice sections to your XF86Config file. This +assumes you are running XFree86 4.x. <SPAN CLASS=diff>On some distributions, +this file is called XF86Config-4</SPAN>. + +<BLOCKQUOTE><PRE> +Section "InputDevice" + Driver "wacom" + Identifier "cursor" + Option "Device" "/dev/input/event0" + Option "Type" "cursor" + Option "Mode" "relative" + Option "USB" "on" + Option "Speed" "3.0" + <B>Option "Threshold" "10"</B> +EndSection + +Section "InputDevice" + Driver "wacom" + Identifier "stylus" + Option "Device" "/dev/input/event0" + Option "Type" "stylus" + Option "Mode" "absolute" + Option "USB" "on" + Option "Tilt" "on" + #Option "TiltInvert" "on" + <B>Option "Threshold" "10"</B> +EndSection + +Section "InputDevice" + Driver "wacom" + Identifier "eraser" + Option "Device" "/dev/input/event0" + Option "Type" "eraser" + Option "Mode" "absolute" + Option "USB" "on" + Option "Tilt" "on" + #Option "TiltInvert" "on" + <B>Option "Threshold" "10"</B> +EndSection + +#Section "InputDevice" +# Driver "wacom" +# Identifier "tablet" +# Option "Device" "/dev/input/event0" +# Option "Type" "beta" +#EndSection +</PRE></BLOCKQUOTE> + +<P>The first three sections identify the cursor, stylus, and eraser devices +to XInput. Notice that all three reference /dev/input/event0. Additionally, +the TiltInvert option has been commented out. You may want to experiment +with these values. + +<P>The fourth InputDevice section is for the new beta driver. Unless you +are actually testing this code, I do not recommend using this section, at +least not at present. There is no harm in leaving it in the file, commented +in or out, just so long as it is not used in the ServerLayout section +described later. + +<P><B>The threshold value is very important.</B> Various wacom tablets behave +differently in terms of how the "TOUCH" event is detected. On some tablets, +the device itself makes the determination, and on others, the driver detects +the event by examining the pressure value. To make matters more complex, +the ranges are all different so what is good for my Intous2 may not be good +for your Graphire. I am providing a default value of 10. If the pressure +exceeds that value, a "TOUCH" event is sent to the XInput system. If this +is too hard, you may want to reduce it. If the value is too low, you will +encounter the reviled "spurious touch" problem where the stylus clicks +randomly as you move it around the screen. + +<P>Also, people using serial tablets will need completely different settings. +Please see <A HREF="http://people.mandrakesoft.com/~flepied/projects/wacom/">Lepied's Wacom Site</A> for +more details. + +<P>I am including a copy of the man page for the driver. Some of the fields +pertain only to the serial device, and some of the fields may be deprecated +under XFree86 4.2.0. Try them out, see what happens.</P> + +<BLOCKQUOTE><PRE> +WACOM(4x) XFree86 WACOM(4x) + +NAME + wacom - Wacom input driver + +SYNOPSIS + Section "InputDevice" + Identifier "idevname" + Driver "wacom" + Option "Device" "devpath" + ... + EndSection + +DESCRIPTION + wacom is an XFree86 input driver for Wacom devices. + + The wacom driver functions as a pointer input device, and may be used + as the X server's core pointer. + +SUPPORTED HARDWARE + This driver supports the Wacom IV and Wacom V protocols. Preliminary + support is available for USB devices on some Linux platforms. + +CONFIGURATION DETAILS + Please refer to XF86Config(5x) for general configuration details and + for options that can be used with all input drivers. This section only + covers configuration details specific to this driver. + + Multiple instances of the Wacom devices can cohabit. It can be useful + to define multiple devices with different active zones. Each device + supports the following entries: + + Option "Type" "stylus"|"eraser"|"cursor" + sets the type of tool the device represent. This option is + mandatory. + + Option "Device" "path" + sets the path to the special file which represents serial + line where the tablet is plugged. You have to specify it + for each subsection with the same value if you want to have + multiple devices with the same tablet. This option is + mandatory. + + Option "USB" "on" + tells the driver to dialog with the tablet the USB way. + This option is only available on some Linux platforms. + + Option "DeviceName" "name" + sets the name of the X device. + + Option "Suppress" "Inumber" + sets the position increment under which not to transmit + coordinates. This entry must be specified only in the + first Wacom subsection if you have multiple devices for one + tablet. If you don’t specify this entry, the default value + is computed to + + Option "Mode" "Relative"|"Absolute" + sets the mode of the device. + + Option "Tilt" "on" + enables tilt report if your tablet supports it (ROM version + 1.4 and above). If this is enabled, multiple devices at + the same time will not be reported. + + Option "HistorySize" "number" + sets the motion history size. By default the value is zero. + + Option "AlwaysCore" "on" + enables the sharing of the core pointer. When this feature + is enabled, the device will take control of the core + pointer (and thus will emit core events) and at the same + time will be able, when asked so, to report extended + events. You can use the last available integer feedback to + control this feature. When the value of the feedback is + zero, the feature is disabled. The feature is enabled for + any other value. + + Option "TopX" "number" + X coordinate of the top corner of the active zone. + + Option "TopY" "number" + Y coordinate of the top corner of the active zone. + + Option "BottomX" "Inumber" + X coordinate of the bottom corner of the active zone. + + Option "BottomY" "number" + Y coordinate of the bottom corner of the active zone. + + Option "KeepShape" "on" + When this option is enabled, the active zone begins + according to TopX and TopY. The bottom corner is adjusted + to keep the ratio width/height of the active zone the same + as the screen while maximizing the area described by TopX, + TopY, BottomX, BottomY. + + + Option "DebugLevel" number + sets the level of debugging info reported. + + Option "BaudRate" "38400", "19200" or "9600" (default) + changes the serial link speed. This option is only avail‐ + able for wacom V models (Intuos). + + Option "Serial" "number" + sets the serial number associated with the physical device. + This allows to have multiple devices of the same type (i.e. + multiple pens). This option is only available on wacom V + devices (Intuos). To see which serial number belongs to a + device, you have to set the DebugLevel to 6 and watch the + output of the X server. + + Option "Threshold" "number" + sets the pressure threshold used to generate a button 1 + events of stylus devices for some models of tablets (Intuos + and Graphire). + +SEE ALSO + XFree86(1), XF86Config(5x), xf86config(1), Xserver(1), X(7x). + +AUTHORS + Frederic Lepied <lepied@xfree86.org> + +4.2.0 Version WACOM(4x) +</PRE></BLOCKQUOTE> +<A NAME="mouse1"> +<H2>5.2 - Mouse1</H2> +<P>Adding the Mouse1 device is <I>probably not something you want to do</I>, +and Redhat's Anaconda program will do it for you if you boot the machine with +the tablet plugged in. You'll need to be careful about this. + +<P>When you use the mouse1 input device, the data flows from the wacom +kernel driver, through the event subsystem, down into the mousedev driver, +out the /dev/input/mouse0 device, and finally into the XInput mouse driver. +You effectively lose all your absolute positioning information because the +mousedev driver converts it into relative data. Additionally, the XFree86 +wacom driver does not get control of the cursor because mouse1 is providing +those events. + +<P>Therefore, if you have a Mouse1 section, leave it. Redhat 8.0 at least, +expects it to be there; however, you will not be using it, so make certain +that it is commented out of the section covered next. +<A NAME="srvlayout"> +<H2>5.3 - ServerLayout</H2> +<P>The ServerLayout section describes what devices the X server will use. +Modify the ServerLayout section to reflect the new devices. Make certain +to comment out the Mouse1 device. + +<BLOCKQUOTE><PRE> +Section "ServerLayout" + Screen 0 "Screen0" 0 0 + InputDevice "Mouse0" "CorePointer" + InputDevice "Keyboard0" "CoreKeyboard" + <B>InputDevice "cursor" "SendCoreEvents" + InputDevice "stylus" "SendCoreEvents" + InputDevice "eraser" "SendCoreEvents"</B> + #InputDevice "Mouse1" "SendCoreEvents" + #InputDevice "tablet" "SendCoreEvents" +EndSection +</PRE></BLOCKQUOTE> + +<P>This section determines which devices are actually used by the server. In +the case above, the cursor, stylus, and eraser devices are selected while +the mouse1 and tablet devices are commented out. At present, this is the +correct configuration for proper operation of the tablet whether you are +using the beta package or the production package. + +<P>The new wacom_drv.o driver contains two independent implementations of the +tablet code. Using the "tablet" device instead of the cursor/stylus/eraser +devices will enable that code. Presently, that will do little for you since +the code is fairly volatile and is simply a testbed for various design ideas. +As it becomes more useful, I will update this section. + +<P>You have completed the XF86Config file changes. By aware that if you reboot +your computer with the Wacom plugged in and Redhat's Anaconda program notices, +it will treat the tablet as a USB mouse and reconfigure this file incorrectly. +You may need to go back and check the file to ensure that everything is +still correct afterwards. Rebooting with the device detached seems to +reverse the process, but again, you should check the file to be certain. +My recommendation is to tell Anaconda to ignore the tablet until the device +detection works properly. +<A NAME="wacomdrv"> +<H2>5.4 - Updating wacom_drv.o</H2> +<P>There are a number of wacom_drv.o files about. XFree86 is shipping +version 23. Redhat 8.0 is shipping version 25. Lepied's page has the +source code for 26, but the binary identifies itself as 23ss1. The +<A HREF="#download">Downloading the Code</A> page currently +offers source and binary for 26-j0.2.0 in the stable package +and 26-j0.3.2a in the beta version. In total, I recommend +using the drivers in the packages that I provide since they correspond with +what is shown in this document. Any new functionality that gets added will +be built on top of these files as well. + +<P>Note: the binary wacom_drv.o file is located in the prebuilt directory +of the stable and beta packages. Building this file from source (which is +provided) is a major hassle, and I do not recommend it unless you are +developer yourself or the binary is simply not compatible with your system. +Instructions for rebuilding from source are available in the Appendix +under <A HREF="#builddrv">Building wacom_drv.o From Scratch</A>. + +<P>With that said, locate and replace wacom_drv.o. Using <i>locate</i> again, we find +the driver in the X11 tree. This directory is pretty standard, so it is +unlikely that your file will be located elsewhere. Note the assumption is +that you are using XFree86 4.x. Anything from the 3.x version probably won't +work. + +<BLOCKQUOTE><PRE> +[root@ayukawa root]# locate wacom_drv.o +/usr/X11R6/lib/modules/input/<B>wacom_drv.o</B> +[root@ayukawa root]# cd /usr/X11R6/lib/modules/input +[root@ayukawa input]# cp <B>wacom_drv.o</B> /home/jej/src/wacom/<B>wacom_drv_old.o</B> +[root@ayukawa input]# cp /home/jej/src/wacom/prebuilt/<B>wacom_drv.o wacom_drv.o</B> +</PRE></BLOCKQUOTE> + +Again, the old file is copied away, and replaced with the newer version. The +binary modules are presently built on a Redhat 8.0 system using gcc 3.2 under +glibc. This should work for most, if not all, major distributions. If it +does not load on yours, please let me know. + +<A NAME="restartx"> +<H2>5.5 - Restart X</H2> +<P>Finally, restart X. You may wish to do this from runlevel 3 for +testing purposes. + +<BLOCKQUOTE><PRE> +[root@ayukawa root]# init 3 +...processes starting and stopping... +[root@ayukawa root]# startx +</PRE></BLOCKQUOTE> + +If the X server dies, you can always back-out the changes to the XF86Config +file and try again. Worse case, copy the wacom_drv_old.o file back to +its original place. But first, look at the XFree86 log file for clues. +You might want to do this even if everything works correctly. When things +are running right, the following lines appear in my log file. + +<BLOCKQUOTE><PRE> +[root@ayukawa root]# grep -i wacom /var/log/XFree86.0.log +(II) LoadModule: "wacom" +(II) Loading /usr/X11R6/lib/modules/input/wacom_drv.o +(II) Module wacom: vendor="The XFree86 Project" +(II) Wacom driver level: <B>26-j0.2.0</B> $ +(II) XINPUT: Adding extended input device "eraser" (type: Wacom Eraser) +(II) XINPUT: Adding extended input device "stylus" (type: Wacom Stylus) +(II) XINPUT: Adding extended input device "cursor" (type: Wacom Cursor) +(==) Wacom Kernel Input device name: "Wacom Intuos2 12x12" +(==) Wacom tablet maximum X=30480 maximum Y=30480 X resolution=0 Y resolution=0 +suppress=0 +(==) Wacom Cursor top X=0 top Y=0 bottom X=30480 bottom Y=30480 +(==) Wacom Stylus top X=0 top Y=0 bottom X=30480 bottom Y=30480 +(==) Wacom Eraser top X=0 top Y=0 bottom X=30480 bottom Y=30480 +</PRE></BLOCKQUOTE> + +Notice the driver version 26-j0.2.0 above. This is the new stable +wacom_drv.o driver. The beta driver is currently 26-j0.3.2a. + +<P>First things first, you should lift the mouse off the tablet and place it +back down. This seems to help reset things internally. When you replace the +mouse, the cursor should jump to that portion of the screen. If everything is +working correctly, the mouse should work in absolute mode with the four +corners of the tablet corresponding with the four corners of the screen. If +the cursor stops short of an edge, then the kernel driver limits are probably +incorrect. Please let me know so that I can update the driver. + +<P>Next, remove the mouse and try using the eraser end of the stylus. +Hovering over the surface of the tablet should move the cursor. Touching the +eraser tip to the surface should generate a click. If you invert the +pen and use the stylus tip, you should get a similar result. If the pen +tip generates spurious touch events, you will need to increase the threshold +value in the InputDevice section. Wacdump is useful for determining the +appropriate value. My pen pressure oscillates between 0 and 6 so I use +a threshold of 10 to compensate. +<A NAME="ptrstat"> +<H2>5.6 - Check the Pointer Status</H2> +<P>You can check the XInput pointer status by using <i>xsetpointer</i> +as below. The man page states that calling xsetpointer with the name of +a particular device will set it as the primary pointing device. So far, +that has not been the case in my experience; however, that is the subject +of current development, so try it, and if it doesn't work, stay tuned. + +<BLOCKQUOTE><PRE> +[root@ayukawa log]# xsetpointer -l +"eraser" [XExtensionDevice] +"stylus" [XExtensionDevice] +"cursor" [XExtensionDevice] +"Mouse0" [XPointer] +"keyboard" [XKeyboard] +</PRE></BLOCKQUOTE> + +<A NAME="gimp"> +<H1>6.0 - Working With Gimp</H1> +<DIV CLASS=diff>It has been suggested that gimp should be recompiled from +source (v1.2.3) on Mandrake 9.0. This does not seem to be true for Redhat +8.0. +</DIV> + +<P>Bring up <i>gimp</i> and select "File|Dialogs|Input Devices". +You will see a drop-down list with all three devices present. After enabling +them, you can see their respective statuses by calling up "File|Dialogs|Device +Status". It has been recommended that the devices be set to "Screen" mode +rather than "Window." + +<P>I have successfully been able to use gimp with several different pens, +including the tips and erasers. Tilt does not appear to be used by gimp +at present, but pressure works fine. +<A NAME="changelog"> +<H1>7.0 - Changelog</H1> +<UL> +<LI>2002-12-22 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Added serial code to wacdump<BR> +* Added serial protocols IV 1.0 - 1.4.<BR> +* Added serial protocol V 1.0 - 1.1.<BR> +* Intuos mouse wheels are now positive when rolled backwards per PS/2 + Intellimouse convention<BR> +* Added Intuos2 to xf86Wacom.c<BR> +* Added build help for Debian Stable users<BR> +</BLOCKQUOTE> + +<LI>2002-12-17 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Merged code from 2.4.20 hid-core.c<BR> +* Added configuration code to handle kernels still at 2.4.19<BR> +* Split wacdump curses code into wacscrn.c to solve namespace collision<BR> +* Added volito to wacom.c<BR> +</BLOCKQUOTE> + +<LI>2002-12-16 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* House cleaning on the website.<BR> +* Dropped the Intuos constraint from the document name.<BR> +* Added the expectations page.<BR> +* Changed the internal indexing to identifier rather than chapter. +</BLOCKQUOTE> +<LI>2002-12-15 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Removed some debugging code, redeployed package with wacom_drv.o version + 0.3.2a.<BR> +* New build has fix for 4D mouse wheel.<BR> +* wcm-beta.c fleshed out with new driver code.<BR> +</BLOCKQUOTE> +<LI>2002-12-13 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Added development page.<BR> +* New build has relative events and fix for 2D mouse wheel.<BR> +* wcm-beta.c added to build.<BR> +</BLOCKQUOTE> +<LI>2002-12-09 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Added detail on "Threshold" parameter.<BR> +* Included man page. +</BLOCKQUOTE> +<LI>2002-11-24 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Updated docs for version 0.2.0.<BR> +* Package uses configure-based scripts.<BR> +</BLOCKQUOTE> +<LI>2002-11-14 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Moved wacom_drv.o to prebuilt directory.<BR> +* Added xf86Wacom.c (version 26) to package along with changes to Makefile<BR> +* Extended documentation on X11 setup including caveat over Mouse1 device.<BR> +* Added xf86Wacom.c build instructions.<BR> +* Updated wacdump (0.2) to display identifiers, maximum values, and all + registered event types.<BR> +* Added highlighting in wacdump to display button down state.<BR> +</BLOCKQUOTE> + +<LI>2002-11-11 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* wacom.c - fixed ranges for intuos2 6x8<BR> +</BLOCKQUOTE> + +<LI>2002-11-10 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* wacdump.c - updated display to identify values<BR> +* wacom.c - fixed ranges for intuos2 12x12<BR> +</BLOCKQUOTE> + +<LI>2002-11-09 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* wacom.c - merged some of Cheng's changes, fixes to the +pressure calculation, device ranges.<BR> +* wacom.c - added an extra device (0x47) for M. T.<BR> +* wacdump.c - added text messages for event types<BR> +* wacdump.c - added usage and device override option<BR> +* wacdump.c - exits when channel is closed.<BR> +* Fixed module dependency problem in wacom.o and hid.o<BR> +* Added clarification for devfs and Mandrake 9.0<BR> +</BLOCKQUOTE> + +<LI>2002-11-06 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Busted page out into discrete chapters.<BR> +* Added some php magic to turn it into an active book.<BR> +* Added Mandrake 9.0<BR> +* Gimp may need to be compiled from source (1.2.3) +</BLOCKQUOTE> +<LI>2002-11-05 - John E. Joganic <jej@j-arkadia.com><BR> +<BLOCKQUOTE> +* Updated wacom driver with Christer Nilsson's 2.4.20 patch<BR> +* Added RedHat 7.3 success<BR> +* Added more discussion on the device driver election.<BR> +* Added more discussion on HID mode 1 and the correct mode 2.<BR> +* Added usbmouse.c and hid-core.c to wacpack-0.1b.<BR> +</BLOCKQUOTE> +<LI>2002-11-04 - John E. Joganic <jej@j-arkadia.com><BR> + +<BLOCKQUOTE> +* Added update section.<BR> +* Added success stories.<BR> +* Added RedHat 7.2 specific information<BR> +* Added more information about "unknown report #1"<BR> +</BLOCKQUOTE> +</UL> +<A NAME="contact"> +<H1>8.0 - Contacts</H1> +<P>John Joganic can be contacted at the following email address: +<A HREF="mailto: jej@j-arkadia.com?SUBJECT=WACOM">jej@j-arkadia.com</A>. +If at all possible, please include the word "WACOM" in the subject line. +I receive hundreds of messages a day, and this little addition will help +me find your message in the tons of spam. +<A NAME="appendix"> +<H1>9.0 - Appendix</H1> +<P>This section is for everything that is either not critical or simply +too complex to describe in the document above without becoming overly +distracting. +<A NAME="builddrv"> +<H2>9.1 - Building wacom_drv.o From Scratch</H2> +<H3>Introduction</H3> + +<P>I should tell you out-right that this is an ugly, time consuming process. +If you manage to get this working in fewer steps, by all means, let me know. + +<P>Why would you want to do this? Two reasons. One, you are a developer +and need to make changes to the source code directly. Two, your distribution +uses c libraries or configuration options that are not compatible with the +wacom_drv.o file that I provide. People running libc5 for instance, would +need to build their own driver. + +<BLOCKQUOTE><DIV CLASS=diff> +Timothy Klein has submitted a brief howto for compiling on Debian Stable +which is still running XFree86 4.1 as of this writing. It covers steps +one through four of this document, and a savvy developer should be able +to figure out step five on his own. If someone solves step five and +generates a patch to Makefile.am, I'll see what I can do about getting +it into the configuration script. That document is on the <A HREF="#debwcmdrv">Building wacom_drv.o On Debian Stable</A> page. +</DIV></BLOCKQUOTE> + +<P>You will need the X source code to rebuild the wacom_drv.o driver. +The build configuration for X generates a number of header files that are +necessary but not installed by default on most distributions. Consequently, +you will need to not only get the source, but build it, practically in its +entirety. Then, after all that, the configure script can be instructed to +hook into the X build tree and rebuild xf86Wacom.c at any time without +having to rebuild X again. + +<P>Since I am running Redhat 8.0 and cannot really pull down the original +XFree86 4.2.0 source code, compile it, and expect it to work on my +system, I need to instead use the source RPM provided by Redhat. If you +choose to go this route, I provide pretty detailed instructions for making +this work. If your distribution works differently, or you are using Gentoo +where most everything is source code by default, you'll need to handle this +as best as possible according to your particular situation. + +<P><H3>Step One: Get The Source</H3> + +<P>On Redhat 8.0, I discovered the version number for my currently installed +XFree86 packages by running <I>rpm -q XFree86</I>. This reported version +4.2.0-72, therefore the source package is XFree86-4.2.0-72.src.rpm. +I downloaded the package from Redhat directly and installed it to the system +as follows: + +<BLOCKQUOTE><PRE> +[root@sen src]# rpm -ivh XFree86-4.2.0-72.src.rpm + 1:XFree86 ########################################### [100%] +</PRE></BLOCKQUOTE> + +<P>This installs a number of files to the /usr/src/redhat directory, +particularly in the SOURCES and SPECS subdirectories. Other distributions +undoubtedly install elsewhere. Look for the XFree86.spec file which should +be located in the SPECS directory. This file contains all the information +necessary to patch the orginal XFree86-4.2.0 source code to the level that +Redhat is distributing in their regular binary package. The source code +and patch files are located in SOURCES. + +<H3>Step Two: Build the Source</H3> + +<P>This step describes how to build the source from the RPM itself. If +you are building from some other mechanism, I honestly cannot offer much +assistance since I generally don't build my X system from scratch. +If you'd like to write up a short section on building the +server for your particular distribution, I would be happy to include it here. + +<P>Next, you don't actually have to build the entire thing. The +point at which the xf86Wacom.c driver can be built however, +is not until somewhere in the middle of the build process. The driver +depends on a number of header files that are created dynamically so until +they are generated, wacom_drv.o cannot be compiled. My solution +was to open a separate terminal in the wacom driver directory and +periodically attempt to build it. When it successfully built, I stopped +the X build process. Here's how to build the source for an RPM that's +been exploded out into the SPECS and SOURCES directories. + +<BLOCKQUOTE><PRE> +[root@sen root]# cd /usr/src/redhat +[root@sen redhat]# rpmbuild -bc SPECS/XFree86.spec +</PRE></BLOCKQUOTE> + +<P>Not every distribution has <I>rpmbuild</I>; try using just <i>rpm</i> +instead. At some point, Redhat split the build functionality into separate +programs. If after looking through the <i>rpm</i> man page, you still +cannot get this to work, send me some email, and I'll look into it. + +<P>The important item is the "-bc" option of <I>rpmbuild</I> which unpacks, +patches, and builds the source without actually installing. While it is also +possible to simply unpack and patch using the "-bp" option, there does not +seem to be a way to just build. The "-bc" option simply deletes all the +files provided by "-bp" and recreates them again. The downside of this is +that if you wanted to simply unpack, patch, and then copy the new xf86Wacom.c +file over the old one, you'll find that the build step deletes it and starts +over again. I have gotten this to work by creating a new patch file, but +this requires a bit more effort, so I don't recommend it right off. + +<H3>Step Three: Build the Original Driver</H3> + +<P>The xf86Wacom.c file is buried pretty deep in the X build tree. If it +is in a different location than the one I have provided below, try using +<I>find . -name xf86Wacom.c</I> from the BUILD directory. + +<BLOCKQUOTE><PRE> +[root@sen redhat]# cd BUILD/XFree86-4.2.0/xc/programs/Xserver/hw/xfree86/input/wacom +[root@sen wacom]# ls +Imakefile wacom.man xf86Wacom.c.Wacom-USB-driver-a25-update +Makefile xf86Wacom.c +</PRE></BLOCKQUOTE> + +The "a25-update" file is the original xf86Wacom.c file before Redhat's patch. +If you open xf86Wacom.c, you'll find that it is version 25, at least as +of this writing and this distribution. The presence of the Makefile means +that the configuration has at least been run for this directory. If you +have built a sufficient portion of the X source files, then all the header +files that you need have been generated, and you can build xf86Wacom.c. Try +it, and if it does not build, wait a bit. The absence of xf86Version.h +for instance, is a good indication that the build process is not ready. + +<BLOCKQUOTE><PRE> +[root@sen wacom]# make +rm -f xf86Wacom.o +gcc -O2 -march=i386 ... -c xf86Wacom.c +rm -f wacom_drv.o +ld -r xf86Wacom.o -o wacom_drv.o +</PRE></BLOCKQUOTE> + +<H3>Step Four: Build the New Driver</H3> + +<P>With a valid build environment, it should be possible to copy the +updated driver source over the old one, run <i>make</i>, and get an updated +driver. + +<BLOCKQUOTE><DIV class=diff>The beta package contains a version xf86Wacom.c that depends +on wcm-beta.c and wcm-beta.h. If you are compiling this version, you will +need to copy all three files. Secondly, the driver will build and link, but +will not be usable since the current X build environment does not expect the +wcm-beta code to be compiled. This problem is addressed correctly in the +automated build process discussed below. +</DIV></BLOCKQUOTE> + +<BLOCKQUOTE><PRE> +[root@sen wacom]# cp xf86Wacom.c xf86Wacom_old.c +[root@sen wacom]# cp /home/jej/src/wacom/src/xf86Wacom.c xf86Wacom.c +[root@sen wacom]# make +rm -f xf86Wacom.o +gcc -O2 -march=i386 ... -c xf86Wacom.c +rm -f wacom_drv.o +ld -r xf86Wacom.o -o wacom_drv.o +</PRE></BLOCKQUOTE> + +If the source did not build correctly the second time, it is possible +that the Imake-based build environment has generated a makefile that +will not work correctly for the new code. Your options are to generate +a patch file that can be applied by the build process, hack the makefile +in the driver directory to work properly, or to find some +way to get Imake to reconfigure. This is probably less of a hassle for people +building from source, but to be fair, all the commands and information you +need to reconfigure from RPM are listed in the beginning of the build output. +You might also want to continue ahead to see if the default build rule does +not work for you. + +<H3> Step Five: Automating the Build Process</H3> + +<P>Copying the updated source file into the build tree and compiling from +there is not convenient. By configuring the package with the --with-xf86 +option set to the XFree86 build tree and enabling the driver using the +--enable-wacomdrv option, you can build the driver outside of the X build +tree. +<BLOCKQUOTE><PRE> +[jej@ayukawa wacom]$ ./configure \ + --with-xf86=/usr/src/redhat/BUILD/XFree86-4.2.0 \ + --enable-wacomdrv +... +BUILD ENVIRONMENT: + XFree86 - yes + BUILD OPTIONS: + wacom_drv.o - yes +[jej@ayukawa wacom]$ make +</PRE></BLOCKQUOTE> + +<P>The makefile rule which builds the driver is contained within src/Makefile.am +and is modified according to the configuration to generate a rule similar to +this in src/Makefile: + +<BLOCKQUOTE><PRE> +xf86Wacom.o: xf86Wacom.c + gcc -O2 -march=i386 -mcpu=$(ARCHITECTURE) -pipe -ansi \ + -pedantic -Wall -Wpointer-arith -fno-merge-constants \ + -I. -I$(XF86_DIR)/programs/Xserver/hw/xfree86/common \ + -I$(XF86_DIR)/programs/Xserver/hw/xfree86/loader \ + -I$(XF86_DIR)/programs/Xserver/hw/xfree86/os-support \ + -I$(XF86_DIR)/programs/Xserver/include \ + -I$(XF86_DIR)/programs/Xserver/mi \ + -I$(XF86_DIR)/exports/include/X11 \ + -I$(XF86_DIR)/include/extensions \ + -I$(XF86_DIR) \ + -I$(XF86_DIR)/exports/include \ + -Dlinux -D__i386__ -D_POSIX_C_SOURCE=199309L -D_POSIX_SOURCE \ + -D_XOPEN_SOURCE -D_BSD_SOURCE -D_SVID_SOURCE -D_GNU_SOURCE \ + -DSHAPE -DXINPUT -DXKB -DLBX -DXAPPGROUP -DXCSECURITY \ + -DTOGCUP -DXF86BIGFONT -DDPMSExtension -DPIXPRIV -DPANORAMIX \ + -DRENDER -DGCCUSESGAS -DAVOID_GLYPHBLT -DPIXPRIV \ + -DSINGLEDEPTH -DXFreeXDGA -DXvExtension -DXFree86LOADER \ + -DXFree86Server -DXF86VIDMODE -DXvMCExtension \ + -DSMART_SCHEDULE -DBUILDDEBUG -DXResExtension \ + -DX_BYTE_ORDER=X_LITTLE_ENDIAN -DNDEBUG -DFUNCPROTO=15 \ + -DNARROWPROTO -DIN_MODULE -DXFree86Module -DLINUX_INPUT \ + -o xf86Wacom.o -c xf86Wacom.c +</PRE></BLOCKQUOTE> + +<BLOCKQUOTE><DIV class=diff>In the beta package, a similar rule applies to wcm-beta.c. +</DIV></BLOCKQUOTE> + +<P>The options and directories specified come directly from the output of the +make command in the previous step. All the root and parent directories have +been replaced with the macro XF86_DIR which in this case is set by the +configuration script to /usr/src/redhat/BUILD/XFree86-4.2.0/xc. If the +options that you see in your build are identical to those above, then the +default rule will work for you now. If not, you'll need to make some +alterations. You can update the Makefile.am file and rerun automake, +update the Makefile.in and rerun configure, or just update the Makefile +directly. + +<P>So long as the X build tree exists, the include directories will point to +the correct locations, and the driver will build. If space is an issue, you +can probably remove all the non-essential directories, but be careful; the +dependency tree in X is huge. + +<H3>Conclusion</H3> + +<P>Although a bit of a hassle, it is possible to compile the xf86Wacom.c +file into the wacom_drv.o driver from outside the X build tree. If you +have any comments, questions, or suggestions regarding this process, +please send email to <A HREF="mailto:jej@j-arkadia.com?SUBJECT=WACOM_DRV">jej@j-arkadia.com</A> with a subject of "WACOM_DRV". +<A NAME="debwcmdrv"> +<H2>9.2 - Building wacom_drv.o On Debian Stable</H2> +The following documentation for building wacom_drv.o on Debian Stable +was written by Timothy Klein. If you have any problems or questions, +go ahead and post them to the list, or send email to me directly. If +I cannot provide a satisfactory answer, I'll forward your email on to +Timothy. + +<BR><BR><BR><BR> + +<PRE> +Just thought I would give you some brief notes on compiling wacom_drv.o +on Debian Stable. This must be done, as the Stable version of Debian +has XFree86 v. 4.1.0.1, and your pre-compiled module is for a later +version of X (4.2?). X complains about the server having ABI 2, and the +module having ABI 3, and refuses to load. So one must compile the +wacom_drv.o module for Debian Stable. + +The good news is that, while it requires a *huge* download, and takes a +bit of time, it is easy. + +------------------------------------------------------------- +Debian wacom_drv.o Compilation + +<B>1. Download the Debian source package for XFree86.</B> + +For Debian Stable this is three files: + +xfree86_4.1.0-16.diff.gz +xfree86_4.1.0-16.dsc +xfree86_4.1.0.orig.tar.gz + +Note that these are *big* files. The 'diff' file is 1.5 megabytes, and +the 'orig' file is 52 megabytes. So if you have a 28.8 modem +connection, this is going to be painful. + +These files can be obtained automatically by using the command: + +apt-get source xfree86 + +It is not necessary to be root to download the source package. After it +is downloaded, the package will be automatically extracted. Using this +method requires that you have a valid 'source' line in your +'/etc/apt/sources.list' file, though. + +You can also just download the three packages above from your favorite +Debian mirror. If you do that, you need to extract them yourself, which +is done with the following command: + +dpkg-source -x xfree86_4.1.0-16.dsc <only for manual download> + +Either method will give you a 'xfree86-4.1.0' directory in the download +directory. For both this step, and the later package creation, you are +going to need tools from the 'dpkg-dev' package. So install that with +'apt-get' if you don't already have it installed. + + + +<B>2. Make sure you have the requisite build dependencies.</B> + +XFree86 requires certain packages to compile. It seems that 'dpkg' is +pretty good about telling you what packages it needs if they are not +installed (probably a nice script included by the Debian X maintainer). +Thus, you can just try and compile it and see what it complains about +not having and, at that point, install the needed packages with apt-get. +The other, cleaner, option is to run: + +apt-get build-dep xfree86 + +This command will attempt to automatically install (or once in a blue +moon uninstall) all the necessary packages to compile XFree86. You need +to be 'root' to do this. + + +<B>3. Begin the Build Process for xfree86</B> + + +Become 'root' (or you could use fakeroot), and change into the +'xfree86-4.1.0' directory. Then run: + +dpkg-buildpackage + +The source will begin to compile. It will take a while. If you have a +really old machine, like a Pentium 166 or such, go on a vacation. It +may be done when you return. + + +<B>4. Recompile the wacom_drv.o module.</B> + +The source you need to replace and recompile will be in the following +directory: + +<download_dir>/xfree86-4.1.0/build-tree/xc/programs/Xserver/hw/xfree86/input/wacom + +Replace the 'xf86Wacom.c' file therein with the one provided. While X +does need to have some configuration done for this to compile, and that +configuration is accomplished automatically by the build process, it +does not necessarily need to finish compiling the whole package. So, +once a 'Makefile' is created automatically by the build scripts in this +directory, you can attempt to run 'make' from time to time. Once it +compiles without error, you can halt the very lengthy build process of +the XFree86 package. +</PRE> + +<BLOCKQUOTE><DIV CLASS=diff>This may not work with the current beta package +since it also contains the wcm-beta.c and wcm-beta.h files. You can +solve this by updating the Makefile, or proceed to step five +on the <A HREF="#builddrv">Building wacom_drv.o From Scratch</A> page. +The command line options for building the driver will undoubtably have +changed, so be prepared to update the Makefile in the wacpack/src directory. +</DIV></BLOCKQUOTE> + +<PRE> +<B>5. Move the wacom_drv.o module to the proper place.</B> + +That's it. Once the 'make' command compiles without error, you will +have several new files in the directory. The one we are interested in +is 'wacom_drv.o.' Now you can just move the 'wacom_drv.o' file to the +directory: + +/usr/X11R6/lib/modules/input/ + +This will also require 'root' privileges. +--------------------------------------------------------------------- + +You may want to add that to you HOWTO, for my fellow Debian users. And +thanks again for the very nice work on the HOWTO and the drivers. + +HTH, + +Tim +</PRE> + +<CENTER><B>Copyright (C) 2002 - John E. Joganic</B></CENTER> +</BODY> +</HTML> |