| Age | Commit message (Collapse) | Author |
|---|
|
make it maxusers adaptive
note, that NTEXT approximation is probably not the best idea
|
|
<mycroft@netbsd.org> reorganization of the vm_collapse logic, although not
used verbatim. We no longer collapse objects from the pagedaemon as that
is not necessary anymore with the more aggressive collapses that gets done.
This also increases performance of loaded systems. Much KNF too.
|
|
|
|
the object's shadow linkage. Also note that the object is going to die
so that a possible collapse can finish early.
|
|
paging in.
|
|
|
|
simultaneously without interfering. Also maintain PG_BUSY at pagein time.
Removed some statics so gdb will be happier.
|
|
Sanitize "fake" page handling once more.
|
|
|
|
when collapsing two objects a formerly COWed page might lose it's backing store
causing stuff like lost ld.so fixups of shared lib images. Also handle fake
pages in all cases. In short: UPDATE!
|
|
|
|
|
|
|
|
|
|
Extend use of vm_object_prefer() to vm_allocate_with_pager().
Make vm_object_prefer() call MD aligner for "pageless" objects too,
so we can have more control over the virtual address to be used.
Implementation could be simpler if we by-pass the object to mapped, but
we'd loose the ability to adapt alignment to objects that were previously
mmap'ed with MAP_FIXED on.
Only expect vm_fork() to return if __FORK_BRAINDAMAGE is defined.
Eliminate unused third arg to vm_fork().
|
|
|
