Date: Mon, 1 Aug 2016 22:53:29 +0000 (UTC) From: Alan Cox <alc@FreeBSD.org> To: src-committers@freebsd.org, svn-src-user@freebsd.org Subject: svn commit: r303649 - user/alc/PQ_LAUNDRY/sys/vm Message-ID: <201608012253.u71MrT3q029885@repo.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: alc Date: Mon Aug 1 22:53:28 2016 New Revision: 303649 URL: https://svnweb.freebsd.org/changeset/base/303649 Log: Update a couple comments. In particular, explain why we dequeue the page early in vm_pageout_scan(). Modified: user/alc/PQ_LAUNDRY/sys/vm/vm_pageout.c Modified: user/alc/PQ_LAUNDRY/sys/vm/vm_pageout.c ============================================================================== --- user/alc/PQ_LAUNDRY/sys/vm/vm_pageout.c Mon Aug 1 22:41:50 2016 (r303648) +++ user/alc/PQ_LAUNDRY/sys/vm/vm_pageout.c Mon Aug 1 22:53:28 2016 (r303649) @@ -948,9 +948,8 @@ vm_pageout_launder(struct vm_domain *vmd } /* - * We unlock the laundry queue, invalidating the - * 'next' pointer. Use our marker to remember our - * place. + * Unlock the laundry queue, invalidating the 'next' pointer. + * Use a marker to remember our place in the laundry queue. */ TAILQ_INSERT_AFTER(&pq->pq_pl, m, &vmd->vmd_laundry_marker, plinks.q); @@ -1333,9 +1332,12 @@ unlock_page: KASSERT(m->hold_count == 0, ("Held page %p", m)); /* - * We unlock the inactive page queue, invalidating the - * 'next' pointer. Use our marker to remember our - * place. + * Dequeue the inactive page and unlock the inactive page + * queue, invalidating the 'next' pointer. Dequeueing the + * page here avoids a later reacquisition (and release) of + * the inactive page queue lock when vm_page_activate(), + * vm_page_free(), or vm_page_launder() is called. Use a + * marker to remember our place in the inactive queue. */ TAILQ_INSERT_AFTER(&pq->pq_pl, m, &vmd->vmd_marker, plinks.q); vm_page_dequeue_locked(m);
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201608012253.u71MrT3q029885>