Synaptics touchpad detection (regression)

True, it is detected as a normal mouse, which would be fine if it let me click and grab, but it does not.

Same on a Compaq Presario 700 and Acer Travelmate 8100, using Dapper Beta 2 live installer. If I run dpkg-reconfigure xserver-xorg, the resulting xorg.conf gets the correct synaptics section.

this is hutterly weird bug.. that code didn't change in ages... Tormod can you please attach /var/log/casper.log and run these few lines on a shell? (you can just copy paste all of them in one shot)

LAPTOP=""
if [ -n "$(which laptop-detect)" ]; then
    if laptop-detect >/dev/null; then
        LAPTOP=true
    fi
fi
echo $LAPTOP

Thanks
Fabio

That laptop-test eturns true for me...

And doing "dpkg-reconfigure xserver-xorg" fixes it for me too.

I don't have any casper.log on any of the laptops!

The script returns "true". Some funny way of writing "laptop-detect && echo true" I would say ;)

I am puzzled... because when that code return true the synaptic instance in xorg.conf is always written.

Yes i know it looks funny but on install we also need to make sure laptop-detect is installed, available and so on.. also LAPTOP gets used a few times, so once it's set.. it's done.. no more to think about it.

you get casper.log only if you are running from the LiveCD.

Fabio

I used the Dapper Beta 2 live CD for both installations. There is no casper.log on the disk, and in /var/log/installer there is only syslog.

No casper.log here, since I can't run the LiveCD at all (display driver problems).

Could this be in any way related to xorg failing to start during the installation after first boot? My laptop can't boot to x with any current ubuntu live or installation cd, I have to install fglrx drivers manually and restart gdm in first boot.

> Could this be in any way related to xorg failing to start

Not likely. I had no problems except the missing synaptics section.

(I thought the log would have been copied to the installed system.) Never mind, here is the log from running the live CD again.

BTW, the casper.log was from the Presario. The laptop-detect returns true also when running from the live CD.

Could you please do the following:

- boot the live cd, but add "break=bottom" to the kernel command line.
- once you get a shell, type:
 chroot /root mount /sys
 chroot /root mount /proc
 chroot /root laptop-detect
 echo $?
- if the last line shows a 1, please do chroot /root sh -x /usr/sbin/laptop-detect and post the output (you might want to run it as chroot /root sh -x /usr/sbin/laptop-detect > /root/var/tmp/laptop.txt 2>&1 to capture the output to a file)

Just a comment: "chroot /root laptop-detect" does not work since /usr/sbin is not in the path. I used the full path and it returned 1.

The previous output was on the Acer. The full dmidecode output is here: http://librarian.launchpad.net/1853317/dmidecode.8103

dmidecode output from the Compaq is here: http://librarian.launchpad.net/2341493/dmidecode.out
I can't try Tollef's test on it right now.

Got the exact same result on the Compaq. I also noticed it was very little in /root/proc/acpi. Then I did "chroot /root /sbin/modprobe battery" and got the attached output, where the battery slot is detected and the script returns 0.

For me this says TRUE

Installed Flight6 to MSI S250 and the Synaptics section got written to xorg.conf.

Any more info needed? The problem persists in 6.06.1 and Edgy Knot-1. That laptop-detect fails to identify laptops during early boot also breaks some other things like screen resolution detection.

Should this bug be reassigned to laptop-detect? Or should other programs make sure the battery module is loaded before calling laptop-detect?

Reassigning to laptop-detect. Please reassign to xorg if you feel this is not correct.

Works for me now in egdy final, upgraded from dapper

Travelmate 8100

I am running Ubuntu 6.10 Desktop cd on the TravelMate - still same problem, and no synaptics in xorg.conf.

I've been having this same problem occur with Feisty, from herd 4 thru today's daily build (20070401) I'm running Kubuntu AMD_64.
laptop-detect returns true.
on herd-5,the last one I installed, dpkg --reconfigure xserver-xorg did correctly detect both the touchpad and the monitor, giving me the proper resolution.

Richard, can you try Tollef's test from comment #12?

Is there any chance this will get fixed before feisty final release?
On my Toshiba (very recent) A110 laptop, as well as on a much older Acer laptop, the synaptics driver is not added to the xorg file by default. As this is on two laptops, and suse detects it by default I think this is more likely to be a problem with the code than with this being weird hardware or something.
Are there any tests I can perform to help?

Mathijs, try Tollef's test from comment #12 (booting from the Desktop CD), and report back here. Then try also my variation in comment #16. That might help, thanks.

Thanks

"dpkg-reconfigure xserver-xorg" worked for me as well.

Travelmate 8100
Feisty

dpkg-reconfigure works for me as well. The diff between the original and the
new xorg.conf shows the section with synaptics touchpad, as well as that the
dpkg-reconfigure fills out values as "Generic Video Card", etc. whereas the
original file had more specific names as "Intel Corporation Mobile
945GM/GMS/940GML Express Integrated Graphics Controller".

So the question remains: why is the default setup procedure creating the
xorg.conf file in some different way?

On 7/8/07, Kobus Prinsloo <email address hidden> wrote:
>
> Thanks
>
> "dpkg-reconfigure xserver-xorg" worked for me as well.
>
> Travelmate 8100
> Feisty
>
> --
> Synaptics touchpad detection (regression)
> https://bugs.launchpad.net/bugs/40503
> You received this bug notification because you are a direct subscriber
> of the bug.
>

On the Festy I didn't meet that problem. So I guess it has been fixed.

I did meet this problem on feisty! The standard install process did not add
synaptics to the xorg.conf, whereas the dpkg-reconfigure method later does.

On 7/8/07, Ulfzart <email address hidden> wrote:
>
> On the Festy I didn't meet that problem. So I guess it has been fixed.
>
> --
> Synaptics touchpad detection (regression)
> https://bugs.launchpad.net/bugs/40503
> You received this bug notification because you are a direct subscriber
> of the bug.
>

> So the question remains: why is the default setup procedure creating the xorg.conf file in some different way?

Because when the installer CD starts and configures X, the "battery" kernel module is not loaded, and therefore the laptop-detect script does not see the battery and does not recognize the computer as a laptop and then does not write the "synaptics" section.

For the specific card names turning into "Generic": the mapping between PCI IDs and names is in the libdiscover1-data package, which is deinstalled at the end of the installation.

If should check for the laptop... I mean, it would be much better, if it
check for that.

2007/7/8, Tormod Volden <email address hidden>:
>
> > So the question remains: why is the default setup procedure creating
> the xorg.conf file in some different way?
>
> Because when the installer CD starts and configures X, the "battery"
> kernel module is not loaded, and therefore the laptop-detect script does
> not see the battery and does not recognize the computer as a laptop and
> then does not write the "synaptics" section.
>
> For the specific card names turning into "Generic": the mapping between
> PCI IDs and names is in the libdiscover1-data package, which is
> deinstalled at the end of the installation.
>
> --
> Synaptics touchpad detection (regression)
> https://bugs.launchpad.net/bugs/40503
> You received this bug notification because you are a member of Ubuntu
> Laptop Team, which is a bug contact for laptop-detect in ubuntu.
>

--
Best regards,
Ulfzart Bjornsson

Ulfzart, I am not sure what you mean, but the point is that finding a battery is a fall-back way to detect a laptop, if there's no more direct indication in for instance the dmidecode output. (BTW, please don't quote complete messages including addresses when you answer bug mail.)

I mean, that if there is another methods to check for the laptop platform, they should be used, but still, we need to check, whether the computer uses battery or not. If the hardware is old, there is no way to hotplug the battery (correct me if I'm wrong), so some people will use the laptop without it, and they don't need some battery-modules in the memory, that will just slow down the system. Also they won't be able to hotplug the battery (even if thay do... then the modules should be loaded). That was my idea.

Thank you.

This patch just runs "modprobe battery || true" before probing for acpi batteries. I made a debdiff against the new upstream version, since it has all the previous Ubuntu patches incorporated.

This patch also fixes bug #91447.

Matthew: can you maybe take a look at this one?

They accepted my patch upstream, so now a sync would be fine.

I filed a sync request in bug #126503. I am not able to unsubscribe ubuntu-main-sponsors from this bug though.

Marking as Fix Committed then (bug 126503) - thanks Tormod.

 laptop-detect (0.13.2) unstable; urgency=low
 .
   [ Otavio Salvador ]
   * Drop the hack used during the package building and use dependencies
     architeture specific handling for it.
   * Add versioned dependency to dmidecode 2.8-2 since we use chassis-type.
   * Add XS-VCS-{Git,Browser} fields.
 .
   [ Tormod Volden ]
   * laptop-detect.in: Load battery module before probing for acpi
     batteries. Closes: #432407