Now that the upstream maintainer has abandoned this project without
establishing another maintainer, its difficult to choose an appropriate
version number. Going with 2.60 for now, as this only has incremental
changes to support 4B.
Based on this raspi-gpio commit[1] the 2711 chip uses a different method
for pull up/down. This change implements that same method.
With this change, wiringPi now works for the Pi4B with the subset of
functionality used by zynthian.
[1] 80fa7d04ea?diff=unified
This change make the following commands correctly detect the 4B
however there may be an issue with reading GPIO inputs on the 4B
as they are always returning zero.
The aim of this change is to be the source-equivalent of the binary
package release 2.52[1] which was the last release before upstream
development ceased.
[1] http://wiringpi.com/wiringpi-updated-to-2-52-for-the-raspberry-pi-4b/
Fixed some compiler warnings.
New wiringPi version mechanism.
Put 4800 baud (back) into the system.
Re-added piBoardRev (really an alias of piGpioLayout())
for kernels 4.8 onwards.
Some very old stuff might break.
Additional fixes for the ISR code and some tweaks here and
there. I've removed the checks for some operations that might
fail when using the gpiomem interface - which is now the default
way of doing things - if your program segfaults, then you
may need to use sudo on it.
Added support for the ADS1115 16-bit ADC
Updated the gpio readall command to correctly with with the Compute Module
and fixed a resulting bug in wiringPi...
Also affects the Pi v2 which also uses the new Revision
encoding scheme (It was slightly bodged before) This ought
to be relatively future proof, but who knows.
Fixed a minor thing in the examples Makefile
(not that anyone bothers to use it, judging by the emails I get )-:
It's not perfect - there is no symbol versioning, and I have had to
make some (perhaps false) assumptions about library ABI stability, but
it does produce something vaguely resembling what would be expected
from a library package found in Debian, and it does so using the
standard Debian source format.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
This makes it possible for the caller of ./build to add a version
number to the sonames of the shared libraries. (This is required for
Debian packages.)
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
This makes it possible for the caller of ./build to suppress the
chown/chmod of the gpio utility.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
This is not set in normal shell sessions.
Honouring an existing setting of LDCONFIG allows the caller of ./build
to specify an alternative program to run.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
These are not set in normal shell sessions. Honouring existing
settings allows the caller of ./build to specify alternative values.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
We are going to want to not run sudo sometimes. So replace literal
invocations of sudo with a variable, which is set to `sudo' by
default, but can be overridden from the environment.
No functional change if the environment variable is not set.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
We are going to want to introduce a real Debian source package build
system. But that needs to contain a lot of things in debian/ which
ought not to appear in the .deb made by the existing ad-hoc `./build
debian' arrangements, which we want to keep so as not to unduly
disturb existing users.
So rename the whole debian/ directory to debian-template/ and change
all references in the build system.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
The symlink libwiringPiDev.so was mistakenly created pointing to
libwiringPi.so.$(VERSION) (note lack of Dev), ie to the wrong library.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
We need to -I and -L the build tree directories, not the directories
in /usr/local or wheveever. Otherwise (a) the `./debian build' fails
if wiringPi is not installed (b) if it _is_ installed, the build picks
up the installed versions rather than the versions being built, which
is wrong.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
emacs complains about this line, because it thinks lines with just a
tab on in Makefiles are suspicious. Remove the needless tab.
No functional change.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
This patch is autogenerated, with:
git-ls-files | grep Makefile | xargs perl -i~ -pe 's/^\t\@/\t\$Q /'
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
We want to be able to disable the many @'s in the Makefiles so
that we can see what they are doing, when things go wrong. It is
conventional to arrange for setting V=1 to have this effect.
Here we supply a formulaic stanza for either setting Q?=@ or
doing nothing. (There is sadly no standard place which is
included in all the Makefiles so this is probably best).
In this patch we do not introduce any users of Q yet. This is
because the next patch, which introduces all the users of Q, can
be generated entirely automatically. (This is also convenient in
case something needs to be rebased across it.)
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
Mark Liffiton
Steve Baker
Steve Baker
Gordon Henderson
Philip Howard
Ian Jackson
IsaoNakamura