The current code simply counts the number of directory entries and takes
that as the maximum event node given. That isn't true if we have a hole in the
event range after removing a device, making the last device in the list
unaccessible.
Fix this by actually looking at the event node numbers.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Upstream e2d3d44b9a6efe4f3968252d4b680397a9640268
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reported-by: Chris Healy <cphealy@gmail.com>
Previous code had repeat handling within the ifdef INPUT_PROP_SEMI_MT.
EV_REP pre-dates input properties.
Also rename the actual printing code. Repeat handling is sort-of a property
but let's not confuse the two.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
The various arrays that define the names are of the respective type size
(i.e. rel is of size REL_CNT, etc.). This means the size is encoded at
compile time, with no boundary checks at runtime for whether a type or
code exceeds those limits.
Wrap the accesses into those arrays into helper functions that check for the
maximum allowed sizes for types and for the respective event codes.
Reported-by: Stéphane Chatty <chatty@enac.fr>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
EV_SYN is 0, so !type means type == EV_SYN
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
This is a bump from the previous one since current kernels have SYN_DROPPED
of 3 - might as well update to this value.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
SYN_MT_REPORT is always defined (see top of evtest.c)
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Having /dev/input/event10 sort after /dev/input/event9 almost makes me think
the future is here.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>
Needs a bit of special treatment, EV_REP bits are not set by the kernel on
EVIOCGBIT as properties between 0 and REP_MAX are always set. Hack around
that so the output is preventing EV_REP to appear as an event.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
evtest-capture has been dropped, but the references are still there in
Makefile.am and configure.ac. In a system having libxml, the make command
returns without being able to compile evtest-capture.o
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Badly maintained and it only provided rudimentary functionality. Several
bugs were never fixed, so recordings would sometimes skip events.
Use evemu instead:
http://www.freedesktop.org/wiki/Evemu
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The default behavior (autodetection) is unchanged.
This adds a knob for distros to control directly.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
With newer automake, it throws an error:
configure.ac:5: error: 'AM_CONFIG_HEADER': this macro is obsolete.
You should use the 'AC_CONFIG_HEADERS' macro instead.
The newer macro has been around a long time, so there's no need to
worry about backwards compat here.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This declaration references INPUT_PROP_MAX, and was introduced in
commit ca8b12d651.
Signed-off-by: Daniel Beer <dlbeer@gmail.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
fe76f7cd2a got inadvertently tagged as 1.28
without bumping configure. Do so now and jump to 1.29.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
evtest versions are different enough now that it does matter what version a
user uses. Add a --version flag to ease triaging of bug reports.
"gcc -o evtest evtest.c" should continue to work, so add an extra ifdef for
the case PACKAGE_VERSION is undefined.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
in query mode, the zero/non-zero exit code matches the zero/non-zero
state of the key bit. so the command "succeeds" in the shell
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Add functionality to query evdev state of a specific key, switch, button,
LED or sound event. This is useful in programs such as powerd
(http://wiki.laptop.org/go/Powerd) which need to query things like the
state of the laptop lid switch from shell code.
Original capture-mode functionality is left unchanged and is still
activated by default. New usage modes are explained in the man page.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This is in preparation for a followup patch which adds a one-shot
query mode.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
As suggested by Dmitry Torokhov, drop the textual name of the individual
state bits. We already have the name of the constant - use that instead.
Also mark the name lookup tables as static const data.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
EV_SW is the event type for switches.
The EV_SW event codes (SW_*) are defined in the linux kernel in linux/input.h.
However, leave the definition of switches in ifdef blocks according to
when they were committed to the linux kernel.
Signed-off-by: Daniel Kurtz <djkurtz@google.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Peter Hutterer
Benjamin Tissoires
Benjamin Tissoires
Mike Frysinger
Daniel Beer
Paul Fox
Daniel Drake
David Gravereaux
Daniel Kurtz