Reverts commits 6e65547 and 479df8e.
These commits caused the build to fail on DragonFlyBSD.
mesa-libs 21.3.9 is not by any means ancient, it is only 3 years old and
it is still receiving critical bug fixes, and most importantly it is
still being used on DragonFlyBSD, also these are relatively small and
harmless checks.
Signed-off-by: b-aaz <b-aazbsd.proton.me>
If HAVE_LIBUDEV isn't set, this function isn't needed, so also should
be ifdef'ed out.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
only used inside the same source file, so better make it static,
so we also get rid of missing prototype warning.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
only used inside the same source file, so better make it static,
so we also get rid of missing prototype warning.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
only used inside the same source file, so better make it static,
so we also get rid of missing prototype warning.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This header is always present in xfree86 SDK for aeons now and already
manadatory for this driver, so there's no need for extra check whose
result isn't used anyways.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
C standard forbids pointer arithmetics on `void*`, only a GNU extension
allows that. On FreeBSD we're getting a warning on that.
Better adhere to the standard and use `char*` instead of `void*` here.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
only used inside the same source file, so better make them static,
so we also get rid of missing prototype warning.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
it's only used inside the same source file, so better make it static,
so we also get rid of missing prototype warning.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
those fields are already declared at the function scope and used there,
so there's no need to have the same one a second time at branch scope.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
a) instead of duplicated extern declarations of gAMDGPUEntityIndex in
several source files, just move this into amdgpu_drv.h header.
b) no need to _X_EXPORT this field (no other modul using it ever).
c) no need for extra getter function, just use the field directly.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
The name ClientStateCallback is already taken for a global field, so we
shouldn't also use it for local one.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This symbol is always present for over a decade ago, so no need
to check for it. And so we also don't need any support for old
external libglamor anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
It's always present, no matter whether DRI3 is actually enabled.
No need to care about ancient - pre-DRI3 - Xserver versions.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
It's present since at least mesa 22.3. No need to care about ancient
mesa versions anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
It's present since at least mesa 22.3. No need to care about ancient
mesa versions anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This function is always present since about a decade ago, so no need to
check for it and having a fallback.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This function is always present for about a decade now, so no need
to check for it (and having a fallback) anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This is always defined in the Xserver SDK for about a decade now.
No need to keep around workarounds for ancient Xserver versions.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This header is always present, no matter whether present extension
actually enabled. Thus no need to explicitly check for it.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
It's always defined in the Xserver SDK since over a decade now.
No need to keep workarounds for ancien Xserver versions.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This function is present in the Xserver for way over a decade, so no
need to carry around our own copy anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
No need to support old Xserver versions with XF86_CRTC_VERSION < 8
or even XF86_CRTC_VERSION < 7.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
It's always defined for about 7 years now. No need to keep special
hacks for ancient Xservers.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
These are always present for over a decade now. No need to keep
special hacks for really ancient Xservers.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This function has been removed over a decade ago - no need to support
such ancient Xserver versions anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
It's always present, no matter whether MIT sync fencing is actually
enabled in the Xserver (we're checking that at rnutime).
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This symbol is always set since about a decade ago - no need to support
ancient Xserver versions anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
All source files should include xorg-config at the very top: this is
necessary to define all the necessary (potentially platform specific)
symbols consumed by other xserver SDK headers. We really shouldn't rely
on this one being included "by accident" somewhere else - include order
really matters here.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Only required for really ancient Xservers versions below 1.18.99.1.0,
so no need to carry it around anymore.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
