Handle a "dma-noncoherent" property on the root node of the device tree.
Consistently use OF_getpropbool() for the "dma-coherent" and
"dma-noncoherent" properties.
ok patrick@, miod@, phessler@, jca@
x11/xfce4/xfce4.port.mk: add a git link to false in ${WRKDIR}/bin
for ports using meson, the meson.build files try to generate
xfce-revision.h from git rev-parse --short HEAD, which:
- is a hidden BDEP on devel/git
- is useless since we don't build from git checkouts
and if dpb junks devel/git after configure, the build fails.
so add the link to make 'git' always fail, instead of adding a useless
BDEP on devel/git.
found the hard way by claudio@ while fixing a metric ton of things on sparc64
looks good to tb@
Allow vmd(8) virtio hardware init to fail so we can clean up.
Remove a bunch of fatal() paths so we can return a failure when trying to
initialize virtio devices. This lets us clean up any child processes
created when we fork/exec things like virtio block and network devices.
This is part of my overall effort to make vmd(8) not just abort and instead
clean up its messes before exit.
ok mlarkin@
Switch to ports-gcc
cc1: error: -Werror=vla: No option -Wvla
cc1: error: unrecognized command line option "-Wimplicit-fallthrough=3"
cc1: error: unrecognized command line option "-Wcast-function-type"
cc1: error: unrecognized command line option "-Wshadow=compatible-local"
cc1: error: unrecognized command line option "-fexcess-precision=standard"
cc1: error: unrecognized command line option "-Wno-stringop-truncation"
OK jca@, go for it tb@
make rport_start more like ifiq_input/if_vinput.
this lets it call if_input_process, which in turn pushes packets
back into the stack with the current softnet threads netstack.