I first noticed this today (17/12/14) although it is possible that it
happened yesterday. Checking /var/log/dpkg.log, no obvious (to me)
package update stood out.

If I have several xterms open, and I work in one, then updating the
that window (say executing ls -l) causes the pointer to vanish. I think
any repainting of the window does the same.

The pointer does not reappear until I remove a finger from the touch
pad (or tap). Just moving the (unseen) pointer around the window - by
moving a finger - does not cause tho pointer to appear. If I move the
(unseen) pointer out of the window - a guess because I cannot see it -
and then back into the window (my focus is set to follow pointer) then
the pointer reappears. This is quite new and disturbing behaviour.
The same problem is happening as I edit this report in vim.

Although I have given this example with xterm windows, something similar
seems to be happening in other windows. I can't be sure that this is
just the synaptic driver or something more generic.

However, I have just plugged in a trackball (~ mouse) and that does
*not* show the problem, so it does suggest a touchpad problem.

Just realised that the same is happening in firefox. I noticed this
for the first time (I think) yesterday. Presumably certain xevents
are being missed? I couldn't figure out a way to use xev to
investigate this problem.

ael

The problem reported is still present on the i386 machine used for
the report: current up to date debian testing.

On an amd64 laptop with a substantially larger Synaptics touchpad,
also Debian testing, the problem is not present. Rarely on both machines
the pointer freezes or vanishes, but reloading the psmouse driver
fixes that. Not sure that that is connected: unlikely, I think.

Both report the trackpad as "Vendor 0x2, Product 0x7". I am not sure how
to get further information on the hardware differences, although I guess
they might both be using the same chip.

ael

I spoke too soon. I am now seeing pointer freezes on the amd64 machine.
Different phenomena, but perhaps associated. I need to switch to a
pseudo terminal and "rmmod psmouse; modprobe psmouse" to get control
of the pointer back under X. gpm was still working in the pseudo
terminal before the reload.

Again, this has only started happening recently, so some sort of bug has
been introduced in the last month (testing, December 2014).