of cv_timedwait() on the pagedaemon condvar -- it's no use constantly
waking the pagedaemon up for new memory allocation attempts, as will
happen e.g. if new network connections are constantly pouring in.
Ran some tests but not enough. There is a deadlock against myself:
rndsink_request acquires rndsinks_lock
-> rnd_getmore
-> hyperentropy feedrandom (or any other rndsource callback)
-> rnd_add_data
-> rndsinks_distribute acquires rndsinks_lock
Need to break this cycle before rndsource callbacks can invoke
rnd_add_data.
While the rump kernel and local clients are by definition in the same
host vmspace, there are subtle differences in how in-kernel code works
in case accessing the kernel vmspace or a user process vmspace.
Problem discovered by riastradh's "read(fd, NULL, 1)" test.
While distributing processing all over the place is not relevant for
high-performance rump kernel I/O stacks (and downright counterproductive),
the mechanism is used e.g. to reach a quiescent state when detaching
an interface, and therefore a semantically correct implementation is
required.
Fixes at least an uncommon race in the ifconfig destroy case.
reported & patch tested by Justin Cormack.
instead of it being always provided by the rump kernel base. This
move accomplishes two things:
1) it is no longer necessary to provide sysproxy hypercall stubs for
platforms which do not want to use sysproxy
2) it is easier to reason about the security aspects, since configurations
not linking the sysproxy component simply do not support remote
system calls
discussed on rumpkernel-users
done in MODULE(). The sysctl tree nodes from all components which
were loaded when rump_init() was called now get created -- mimics the
monolithic kernel.
No change to the monolithic kernel.
mappings of file objects. move vnode-specific details of mmap()ing a vnode
from uvm_mmap() to the new vnode-specific vn_mmap(). add new uvm_mmap_dev()
and uvm_mmap_anon() convenience functions for mapping character devices
and anonymous memory, and replace all other calls to uvm_mmap() with those.
use the new fileop in drm2 so that libdrm can use mmap() to map things
like on other platforms (instead of the ioctl that we have used so far).
