wacom-tools init.d script is broken

I should clarify that the real problem with the init script is not that it doesn't create the /dev/wacom symlink as I had discussed, but rather that it makes no attempt to configure the serial port at all.

exact same issue on my tc4200 too. The script is definitely broken:

$ sudo /etc/init.d/wacom-tools start
 * Doing Wacom setup... ln: target `/dev/wacom' is not a directory
                                                                         [ ok ]

Note that it doesn't even detect its own failure...

HTH

I have exaclty the saem problem. To be honest the script does nothing. it does not configure the serial port of the stylus or anything else. So nothing at all happens.

Here is the original script
. /lib/lsb/init-functions

if [ ! -d /sys/bus/pnp/devices ]; then exit; fi

cd /sys/bus/pnp/devices

case $1 in
    start|restart|reload|force-reload)
    log_begin_msg "Doing Wacom setup..."
    for x in *; do
        PORT=unknown;
        for y in `cat $x/id`; do
            case "$y" in
                WACf006*|WACf005*|WACf004*)
                PORT=/dev/`echo $x/tty:* | awk -F: '{print $3}'`
                echo "PORT is $PORT"
                ln -sf $PORT /dev/wacom
                ;;
            esac
        done
    done
    log_end_msg 0;
    ;;
    stop)
    exit 0;
    ;;
esac

and here is the replacement
. /lib/lsb/init-functions
cd /sys/bus/pnp/devices
DEVICENUM=0 # First serial port to try
WACOMNUM=0

case $1 in
    start|restart|reload|force-reload)
    log_begin_msg "Doing Wacom setup..."
    for x in *; do
        TYPE=none;
        PORT=unknown;
        CONFIGURED=0;
        for y in `cat $x/id`; do
            case "$y" in
                WACf006*|WACf005*|WACf004*)
                echo found device in $x and has id $y
            # Ensure that the device is enabled
                echo "activate" >$x/resources
            # And now figure out what port and IRQ it's on
                IOPORT=`grep io $x/resources | sed -e 's/io \(.*\)-.*/\1/'`
                echo "has IOPORT=$IOPORT"
                IRQ=`grep irq $x/resources | sed -e 's/irq \(.*\)/\1/'`
                echo "has IRQ=$IRQ"
                while [ "$CONFIGURED" = "0" ]; do
                    DEVICE="/dev/ttyS$DEVICENUM"
                    UART=`setserial -G $DEVICE | awk '{print $3}'`
                    echo "For device $DEVICE, UART=$UART"
                    if [ "$UART" = "unknown" ]; then
                    # Found an unused serial port
                        setserial $DEVICE port $IOPORT irq $IRQ autoconfig
                        if [ ! -e /dev/wacom$WACOMNUM ]; then
                            ln -s $DEVICE /dev/wacom$WACOMNUM
                        fi
                        let "WACOMNUM=$WACOMNUM+1";
                        CONFIGURED=1;
                    fi
                    let "DEVICENUM=$DEVICENUM+1";
                done;
            esac
        done
    done
    log_end_msg 0;
    ;;
    stop)
    exit 0;
    ;;
esac

I also changed the stylus device in xorg.conf from /dev/wacom to /dev/wacom0 to reflect this change.

I have not tested this script with USB based stylus. And also, it fails if I configure the stylus using setserial manually to another device. In that case it maps /dev/wacom0 to /dev/ttyS1 and things go haywire.

I forgot to add, please remove the echo statements, I used them for my own debugging and had left them there

Your script is broken. It's likely to re-introduce the same error that the old script had when it didn't detect any devices. This script is DANGEROUS (as in "an end user will only see lots of '/dev/ttySn: file or directory not found' errors").

Attaching the one I am using on my machine (which only checks existing devices).

both fixes this error and doesn't introduce an old one

You might want to add setting the CONFIGURED variable to one on success and check it for potential warnings.

The kernel should be doing the setup now. Can you test with the latest kernel and the stock init script?

works for me, thanks!

On 5/3/06, Matthew Garrett <email address hidden> wrote:
> The kernel should be doing the setup now. Can you test with the latest
> kernel and the stock init script?
>
> --
> wacom-tools init.d script is broken
> https://launchpad.net/bugs/39526
>

--
Thibaut VARENE
http://www.parisc-linux.org/~varenet/

In Flight 7 on my computer, it is detected, but it is not configured properly in xorg.conf . The xorg.conf entry for the wacom tablet points to /dev/wacom, but there is no /dev/wacom at all.

Then it's not detected properly. What sort of hardware are you talking about?

It's a Graphire usb. When using the Flight 7 live-cd, the symlink /dev/wacom is not available. This is my wacdump output:

MODEL=Wacom Graphire ROM=1.1-1
CLS=USB VNDR=Wacom DEV=Graphire SUB=ET_0405

I had the same problem as Wouter. Checking /etc/udev/rules.d/65-wacom.rules showed that udev creates a symlink /dev/input/wacom, not /dev/wacom. After changing /etc/X11/xorg.conf to use /dev/input/wacom instead of /dev/wacom, it works fine so far.

Wouter, Sasha,

This seems fixed now. The bug about /dev/input/wacom created instead of /dev/wacom is an old longstanding bug.

You should now check if /dev/input/wacom is created (should be). See Bug #40473 for news about that.

Since the initial bug seems closed, I'll close the bug but you are free to reopen it in case even /dev/input/wacom isn't created