Author: kib Date: Sun Aug 16 20:57:02 2020 New Revision: 364286 URL: https://svnweb.freebsd.org/changeset/base/364286
Log: vm_object: allow paging_in_progress to be acquired after object termination. The vm objects are type-stable, and can be accessed even after the last reference is dropped, or in case of vnode objects, after vgone() destroyed it as well. Stop asserting that pip == 0 after vm_object_terminate() waited for existing owners to drop it, we only want to drain them before setting OBJ_DEAD flag. Also stop asserting pip == 0 in object destructor. Update comments explaining the interaction between paging_in_progress and termination. Reviewed by: markj Tested by: pho Sponsored by: The FreeBSD Foundation Differential revision: https://reviews.freebsd.org/D25968 Modified: head/sys/vm/vm_object.c Modified: head/sys/vm/vm_object.c ============================================================================== --- head/sys/vm/vm_object.c Sun Aug 16 20:52:24 2020 (r364285) +++ head/sys/vm/vm_object.c Sun Aug 16 20:57:02 2020 (r364286) @@ -192,9 +192,6 @@ vm_object_zdtor(void *mem, int size, void *arg) ("object %p has reservations", object)); #endif - KASSERT(blockcount_read(&object->paging_in_progress) == 0, - ("object %p paging_in_progress = %d", - object, blockcount_read(&object->paging_in_progress))); KASSERT(!vm_object_busied(object), ("object %p busy = %d", object, blockcount_read(&object->busy))); KASSERT(object->resident_page_count == 0, @@ -294,6 +291,9 @@ vm_object_init(void) * The lock portion of struct vm_object must be type stable due * to vm_pageout_fallback_object_lock locking a vm object * without holding any references to it. + * + * paging_in_progress is valid always. Lockless references to + * the objects may acquire pip and then check OBJ_DEAD. */ obj_zone = uma_zcreate("VM OBJECT", sizeof (struct vm_object), NULL, #ifdef INVARIANTS @@ -936,12 +936,13 @@ vm_object_terminate(vm_object_t object) ("terminating shadow obj %p", object)); /* - * wait for the pageout daemon to be done with the object + * Wait for the pageout daemon and other current users to be + * done with the object. Note that new paging_in_progress + * users can come after this wait, but they must check + * OBJ_DEAD flag set (without unlocking the object), and avoid + * the object being terminated. */ vm_object_pip_wait(object, "objtrm"); - - KASSERT(!blockcount_read(&object->paging_in_progress), - ("vm_object_terminate: pageout in progress")); KASSERT(object->ref_count == 0, ("vm_object_terminate: object with references, ref_count=%d", _______________________________________________ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
