Using CUDA with NVidia requires those two programs if one wants to use
more than one program doing CUDA at the same time.
This is only available on x86_64.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Al West <al.west@v-nova.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, nvidia-driver always installs the X.org driver, assuming this
is a requirement for all functionalities provided by nvidia-driver, thus
pulling in quite a bunch of X.org packages.
However, it is perfectly legit to be doing CUDA (and OpenCL) work
without the full X.org stack, and indeed the NVidia CUDA and OpenCL
libraries do not require the X.org stack.
Split the configuration so that it is posible to install the different
parts independently from each others, so that CUDA can be installed all
on its own.
Reported-by: Al West <al.west@v-nova.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Al West <al.west@v-nova.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add option to build the nvidia.ko module. If CUDA is enabled on x86_64,
also build the nvidia-uvm.ko kernel module (for Unified Memory access),
which is required by the CUDA user-land library.
Substancially inspired by the corresponding Gentoo ebuild:
http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/x11-drivers/nvidia-drivers/nvidia-drivers-340.32.ebuild?revision=1.2&view=markup
[Thomas:
- add quotes when using $(TARGET_CC) and other variables, since they
can have spaces in their values
- remove space after opening parenthesis and before closing parenthesis.]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In order to compile xserver, libgl provider have to provide gl.pc file.
Signed-off-by: Jérôme Pouiller <jezz@sysmic.org>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch only adds the userland part. Unless other such other
packages (which we named like: rpi-userland), we do not replicate this
naming scheme with this package, as a future patch will also enable
building the kernel part of the driver. So, it is better to just name
that package with -driver, rather than with -userland and renaming it
afterwards.
[Thomas:
- Rewrap Config.in help text.
- Add a comment to explain why mesa3d-headers, xlib_libX11 and
xlib_libXext are part of the dependencies.
- Fix typo in comment about library installation: s/The/Then/
- Use 'addsuffix' instead of 'patsubst' to calculate the final
filename of libraries to install.
- Use more temporary variables to make the library installation loop
clearer: 'libpath' is the relative path of the library in
nvidia-driver sources, 'libname' the base name of the library,
'libsoname' the soname of the library, and 'baseso' the base .so
symlink name.]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
