The base VMCache class changed to the generic_ types with their
introduction in in *2011* (435c43f5912b109e7d5cf682865d2061e62fad8c),
but these classes were never properly adapted. These functions should not
be called here (they panic() -- but the base class only returns B_ERROR,
so that is a difference at least.)
Found by Clang's -Woverloaded-virtual.
- Don't remove .eh_frame in the linker scripts, as this saves little
space in practice, and results in debug binaries of the runtime loader
not being properly debuggable due to missing canonical frame
information.
* When SystemProfiler::_MaybeNotifyProfilerThreadLocked() is called
and the conditions are right, it will lock the thread's scheduler
spinlock and unblock it. Internally, the unblock will enqueue the
thread into the run queue, which causes a ThreadEnqueuedInRunQueue
event for SystemProfiler. Since the conditions haven't changed, it
now went into _MaybeNotifyProfilerThreadLocked again (this time
from the profiler thread context). In there, it will try to lock
the profiler thread's scheduling spinlock, which is already locked
by the other thread (which is firmly sleeping). Deadlock, KDL.
* Before unblocking the profiler thread, unset fWaitingProfilerThread
so that further events will not try to unblock it again.
Under normal conditions this will behave identically. But when using LD_PRELOAD
or the like, some shared objects (and dependencies) are loaded before the
app image is, and on the secondary arch on hybrid builds, will fail to load
any of their dependencies due to sSearchPathSubDir not being set.
Fixes#12214, which had plagued users of libroot_debug for a long time now.
This should have been done along with the time_t change, but I forgot
to check this then.
Technically this breaks ABI against BeOS, but:
1. BeOS used an int32, so we'd already slightly broken ABI here
2. Only one thing at HaikuArchives (VMwareAddons) and one recipe at HaikuPorts
(samba) uses this function at all.
If it turns out some critical BeOS app uses this, then I guess we can enclose
GCC2 guards around it, but since I can't find any evidence of that, I'm
pushing it without them for now.
This reverts commit 2960780faa7e9c2e5e7a80cb8925be5ee59ae5ef.
Sorry for the noise; I thought I'd run a compile but apparently it was
only a partial one.
It seems that the netboot core is linked into the main bootloader
no matter the build configuration, which means that since this one
function was called, the linker tried to resolve all the other symbols
from that static library. Which of course failed.
The NetStack.h header is currently not usable from C code. So while
net_stack_init() is called from platform code, we cannot call
net_stack_cleanup() from OpenFirmware's platform_start_kernel().
Thus call it directly from main()'s cleanup TODO, having assured
that the function is a no-op when no network stack was initialized.
Signed-off-by: Augustin Cavalier <waddlesplash@gmail.com>
The UDP service does not own the UDP sockets. When shutting down,
inform the bound sockets that the service is no longer available.
This allows subsequent method calls to error out cleanly.
Signed-off-by: Augustin Cavalier <waddlesplash@gmail.com>
Add a cleanup function net_stack_cleanup() that calls a new NetStack::ShutDown() method.
Make sure this method works even if the network stack was never initialized.
Signed-off-by: Augustin Cavalier <waddlesplash@gmail.com>
It is currently done in both ~EthernetService() and ~NetStack().
Since NetStack is where it's added and where an explicit accessor function is provided,
choose that location.
Signed-off-by: Augustin Cavalier <waddlesplash@gmail.com>
* With the amd64 architecture, the stack fault exception got a new
meaning: it is raised when an instruction tries to access a
non-canonical address, and the stack is referenced in the
instruction (e.g. by its addressing mode).
So unlike on x86, this is not a fatal exception and shouldn't
trigger a KDL -- instead, it is to be treated like a general
protection fault, terminating the team which caused it.
* Fixes#13744
For readers waiting in select() or simply in read(),
B_SELECT_DISCONNECTED notification didn't wake them anymore.
Now, cascade notifications to be sure to wake them, either
on POLLHUP, POOLERR/B_SELECT_ERROR or, at worst, from read().
See #7859 for details.
