From owner-svn-src-user@freebsd.org Mon May 1 01:58:21 2017 Return-Path: Delivered-To: svn-src-user@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 5AAF3D583E9 for ; Mon, 1 May 2017 01:58:21 +0000 (UTC) (envelope-from markj@FreeBSD.org) Received: from repo.freebsd.org (repo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 246C3E3C; Mon, 1 May 2017 01:58:21 +0000 (UTC) (envelope-from markj@FreeBSD.org) Received: from repo.freebsd.org ([127.0.1.37]) by repo.freebsd.org (8.15.2/8.15.2) with ESMTP id v411wKfo087804; Mon, 1 May 2017 01:58:20 GMT (envelope-from markj@FreeBSD.org) Received: (from markj@localhost) by repo.freebsd.org (8.15.2/8.15.2/Submit) id v411wJml087798; Mon, 1 May 2017 01:58:19 GMT (envelope-from markj@FreeBSD.org) Message-Id: <201705010158.v411wJml087798@repo.freebsd.org> X-Authentication-Warning: repo.freebsd.org: markj set sender to markj@FreeBSD.org using -f From: Mark Johnston Date: Mon, 1 May 2017 01:58:19 +0000 (UTC) To: src-committers@freebsd.org, svn-src-user@freebsd.org Subject: svn commit: r317628 - user/markj/PQ_LAUNDRY_11/sys/vm X-SVN-Group: user MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-user@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: "SVN commit messages for the experimental " user" src tree" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 01 May 2017 01:58:21 -0000 Author: markj Date: Mon May 1 01:58:19 2017 New Revision: 317628 URL: https://svnweb.freebsd.org/changeset/base/317628 Log: MFC r309898 (by alc): Purge mentions of PG_CACHED pages from sys/vm. Modified: user/markj/PQ_LAUNDRY_11/sys/vm/vm_map.c user/markj/PQ_LAUNDRY_11/sys/vm/vm_object.c user/markj/PQ_LAUNDRY_11/sys/vm/vm_object.h user/markj/PQ_LAUNDRY_11/sys/vm/vm_page.c user/markj/PQ_LAUNDRY_11/sys/vm/vm_page.h user/markj/PQ_LAUNDRY_11/sys/vm/vm_reserv.c Directory Properties: user/markj/PQ_LAUNDRY_11/ (props changed) Modified: user/markj/PQ_LAUNDRY_11/sys/vm/vm_map.c ============================================================================== --- user/markj/PQ_LAUNDRY_11/sys/vm/vm_map.c Mon May 1 01:56:13 2017 (r317627) +++ user/markj/PQ_LAUNDRY_11/sys/vm/vm_map.c Mon May 1 01:58:19 2017 (r317628) @@ -1858,9 +1858,7 @@ vm_map_submap( * limited number of page mappings are created at the low-end of the * specified address range. (For this purpose, a superpage mapping * counts as one page mapping.) Otherwise, all resident pages within - * the specified address range are mapped. Because these mappings are - * being created speculatively, cached pages are not reactivated and - * mapped. + * the specified address range are mapped. */ static void vm_map_pmap_enter(vm_map_t map, vm_offset_t addr, vm_prot_t prot, Modified: user/markj/PQ_LAUNDRY_11/sys/vm/vm_object.c ============================================================================== --- user/markj/PQ_LAUNDRY_11/sys/vm/vm_object.c Mon May 1 01:56:13 2017 (r317627) +++ user/markj/PQ_LAUNDRY_11/sys/vm/vm_object.c Mon May 1 01:58:19 2017 (r317628) @@ -1356,7 +1356,7 @@ retry: goto retry; } - /* vm_page_rename() will handle dirty and cache. */ + /* vm_page_rename() will dirty the page. */ if (vm_page_rename(m, new_object, idx)) { VM_OBJECT_WUNLOCK(new_object); VM_OBJECT_WUNLOCK(orig_object); @@ -1443,6 +1443,13 @@ vm_object_scan_all_shadowed(vm_object_t backing_object = object->backing_object; + /* + * Initial conditions: + * + * We do not want to have to test for the existence of swap + * pages in the backing object. XXX but with the new swapper this + * would be pretty easy to do. + */ if (backing_object->type != OBJT_DEFAULT && backing_object->type != OBJT_SWAP) return (false); @@ -1594,8 +1601,7 @@ vm_object_collapse_scan(vm_object_t obje * backing object to the main object. * * If the page was mapped to a process, it can remain mapped - * through the rename. vm_page_rename() will handle dirty and - * cache. + * through the rename. vm_page_rename() will dirty the page. */ if (vm_page_rename(p, object, new_pindex)) { next = vm_object_collapse_scan_wait(object, NULL, next, Modified: user/markj/PQ_LAUNDRY_11/sys/vm/vm_object.h ============================================================================== --- user/markj/PQ_LAUNDRY_11/sys/vm/vm_object.h Mon May 1 01:56:13 2017 (r317627) +++ user/markj/PQ_LAUNDRY_11/sys/vm/vm_object.h Mon May 1 01:58:19 2017 (r317628) @@ -79,17 +79,6 @@ * * vm_object_t Virtual memory object. * - * The root of cached pages pool is protected by both the per-object lock - * and the free pages queue mutex. - * On insert in the cache radix trie, the per-object lock is expected - * to be already held and the free pages queue mutex will be - * acquired during the operation too. - * On remove and lookup from the cache radix trie, only the free - * pages queue mutex is expected to be locked. - * These rules allow for reliably checking for the presence of cached - * pages with only the per-object lock held, thereby reducing contention - * for the free pages queue mutex. - * * List of locks * (c) const until freed * (o) per-object lock Modified: user/markj/PQ_LAUNDRY_11/sys/vm/vm_page.c ============================================================================== --- user/markj/PQ_LAUNDRY_11/sys/vm/vm_page.c Mon May 1 01:56:13 2017 (r317627) +++ user/markj/PQ_LAUNDRY_11/sys/vm/vm_page.c Mon May 1 01:58:19 2017 (r317628) @@ -1435,9 +1435,7 @@ vm_page_replace(vm_page_t mnew, vm_objec * * Note: we *always* dirty the page. It is necessary both for the * fact that we moved it, and because we may be invalidating - * swap. If the page is on the cache, we have to deactivate it - * or vm_page_dirty() will panic. Dirty pages are not allowed - * on the cache. + * swap. * * The objects must be locked. */ @@ -2075,18 +2073,18 @@ unlock: } else if (level >= 0) { /* * The page is reserved but not yet allocated. In - * other words, it is still cached or free. Extend - * the current run by one page. + * other words, it is still free. Extend the current + * run by one page. */ run_ext = 1; #endif } else if ((order = m->order) < VM_NFREEORDER) { /* * The page is enqueued in the physical memory - * allocator's cache/free page queues. Moreover, it - * is the first page in a power-of-two-sized run of - * contiguous cache/free pages. Add these pages to - * the end of the current run, and jump ahead. + * allocator's free page queues. Moreover, it is the + * first page in a power-of-two-sized run of + * contiguous free pages. Add these pages to the end + * of the current run, and jump ahead. */ run_ext = 1 << order; m_inc = 1 << order; @@ -2094,16 +2092,15 @@ unlock: /* * Skip the page for one of the following reasons: (1) * It is enqueued in the physical memory allocator's - * cache/free page queues. However, it is not the - * first page in a run of contiguous cache/free pages. - * (This case rarely occurs because the scan is - * performed in ascending order.) (2) It is not - * reserved, and it is transitioning from free to - * allocated. (Conversely, the transition from - * allocated to free for managed pages is blocked by - * the page lock.) (3) It is allocated but not - * contained by an object and not wired, e.g., - * allocated by Xen's balloon driver. + * free page queues. However, it is not the first + * page in a run of contiguous free pages. (This case + * rarely occurs because the scan is performed in + * ascending order.) (2) It is not reserved, and it is + * transitioning from free to allocated. (Conversely, + * the transition from allocated to free for managed + * pages is blocked by the page lock.) (3) It is + * allocated but not contained by an object and not + * wired, e.g., allocated by Xen's balloon driver. */ run_ext = 0; } @@ -2315,11 +2312,11 @@ unlock: if (order < VM_NFREEORDER) { /* * The page is enqueued in the physical memory - * allocator's cache/free page queues. - * Moreover, it is the first page in a power- - * of-two-sized run of contiguous cache/free - * pages. Jump ahead to the last page within - * that run, and continue from there. + * allocator's free page queues. Moreover, it + * is the first page in a power-of-two-sized + * run of contiguous free pages. Jump ahead + * to the last page within that run, and + * continue from there. */ m += (1 << order) - 1; } @@ -2368,9 +2365,9 @@ CTASSERT(powerof2(NRUNS)); * conditions by relocating the virtual pages using that physical memory. * Returns true if reclamation is successful and false otherwise. Since * relocation requires the allocation of physical pages, reclamation may - * fail due to a shortage of cache/free pages. When reclamation fails, - * callers are expected to perform VM_WAIT before retrying a failed - * allocation operation, e.g., vm_page_alloc_contig(). + * fail due to a shortage of free pages. When reclamation fails, callers + * are expected to perform VM_WAIT before retrying a failed allocation + * operation, e.g., vm_page_alloc_contig(). * * The caller must always specify an allocation class through "req". * @@ -2405,8 +2402,8 @@ vm_page_reclaim_contig(int req, u_long n req_class = VM_ALLOC_SYSTEM; /* - * Return if the number of cached and free pages cannot satisfy the - * requested allocation. + * Return if the number of free pages cannot satisfy the requested + * allocation. */ count = vm_cnt.v_free_count + vm_cnt.v_cache_count; if (count < npages + vm_cnt.v_free_reserved || (count < npages + @@ -2676,9 +2673,8 @@ vm_page_activate(vm_page_t m) /* * vm_page_free_wakeup: * - * Helper routine for vm_page_free_toq() and vm_page_cache(). This - * routine is called when a page has been added to the cache or free - * queues. + * Helper routine for vm_page_free_toq(). This routine is called + * when a page is added to the free queues. * * The page queues must be locked. */ @@ -2766,8 +2762,8 @@ vm_page_free_toq(vm_page_t m) pmap_page_set_memattr(m, VM_MEMATTR_DEFAULT); /* - * Insert the page into the physical memory allocator's - * cache/free page queues. + * Insert the page into the physical memory allocator's free + * page queues. */ mtx_lock(&vm_page_queue_free_mtx); vm_phys_freecnt_adj(m, 1); @@ -2871,21 +2867,10 @@ vm_page_unwire(vm_page_t m, uint8_t queu /* * Move the specified page to the inactive queue. * - * Many pages placed on the inactive queue should actually go - * into the cache, but it is difficult to figure out which. What - * we do instead, if the inactive target is well met, is to put - * clean pages at the head of the inactive queue instead of the tail. - * This will cause them to be moved to the cache more quickly and - * if not actively re-referenced, reclaimed more quickly. If we just - * stick these pages at the end of the inactive queue, heavy filesystem - * meta-data accesses can cause an unnecessary paging load on memory bound - * processes. This optimization causes one-time-use metadata to be - * reused more quickly. - * - * Normally noreuse is FALSE, resulting in LRU operation. noreuse is set - * to TRUE if we want this page to be 'as if it were placed in the cache', - * except without unmapping it from the process address space. In - * practice this is implemented by inserting the page at the head of the + * Normally, "noreuse" is FALSE, resulting in LRU ordering of the inactive + * queue. However, setting "noreuse" to TRUE will accelerate the specified + * page's reclamation, but it will not unmap the page from any address space. + * This is implemented by inserting the page near the head of the inactive * queue, using a marker page to guide FIFO insertion ordering. * * The page must be locked. @@ -3012,16 +2997,9 @@ vm_page_advise(vm_page_t m, int advice) if (advice == MADV_FREE) /* * Mark the page clean. This will allow the page to be freed - * up by the system. However, such pages are often reused - * quickly by malloc() so we do not do anything that would - * cause a page fault if we can help it. - * - * Specifically, we do not try to actually free the page now - * nor do we try to put it in the cache (which would cause a - * page fault on reuse). - * - * But we do make the page as freeable as we can without - * actually taking the step of unmapping it. + * without first paging it out. MADV_FREE pages are often + * quickly reused by malloc(3), so we do not do anything that + * would result in a page fault on a later access. */ vm_page_undirty(m); else if (advice != MADV_DONTNEED) Modified: user/markj/PQ_LAUNDRY_11/sys/vm/vm_page.h ============================================================================== --- user/markj/PQ_LAUNDRY_11/sys/vm/vm_page.h Mon May 1 01:56:13 2017 (r317627) +++ user/markj/PQ_LAUNDRY_11/sys/vm/vm_page.h Mon May 1 01:58:19 2017 (r317628) @@ -352,19 +352,16 @@ extern struct mtx_padalign pa_lock[]; * free * Available for allocation now. * - * cache - * Almost available for allocation. Still associated with - * an object, but clean and immediately freeable. - * - * The following lists are LRU sorted: - * * inactive * Low activity, candidates for reclamation. + * This list is approximately LRU ordered. + * + * laundry * This is the list of pages that should be * paged out next. * * active - * Pages that are "active" i.e. they have been + * Pages that are "active", i.e., they have been * recently referenced. * */ Modified: user/markj/PQ_LAUNDRY_11/sys/vm/vm_reserv.c ============================================================================== --- user/markj/PQ_LAUNDRY_11/sys/vm/vm_reserv.c Mon May 1 01:56:13 2017 (r317627) +++ user/markj/PQ_LAUNDRY_11/sys/vm/vm_reserv.c Mon May 1 01:58:19 2017 (r317628) @@ -62,7 +62,7 @@ __FBSDID("$FreeBSD$"); /* * The reservation system supports the speculative allocation of large physical - * pages ("superpages"). Speculative allocation enables the fully-automatic + * pages ("superpages"). Speculative allocation enables the fully automatic * utilization of superpages by the virtual memory system. In other words, no * programmatic directives are required to use superpages. */ @@ -155,11 +155,11 @@ popmap_is_set(popmap_t popmap[], int i) * physical pages for the range [pindex, pindex + VM_LEVEL_0_NPAGES) of offsets * within that object. The reservation's "popcnt" tracks the number of these * small physical pages that are in use at any given time. When and if the - * reservation is not fully utilized, it appears in the queue of partially- + * reservation is not fully utilized, it appears in the queue of partially * populated reservations. The reservation always appears on the containing * object's list of reservations. * - * A partially-populated reservation can be broken and reclaimed at any time. + * A partially populated reservation can be broken and reclaimed at any time. */ struct vm_reserv { TAILQ_ENTRY(vm_reserv) partpopq; @@ -196,11 +196,11 @@ struct vm_reserv { static vm_reserv_t vm_reserv_array; /* - * The partially-populated reservation queue + * The partially populated reservation queue * - * This queue enables the fast recovery of an unused cached or free small page - * from a partially-populated reservation. The reservation at the head of - * this queue is the least-recently-changed, partially-populated reservation. + * This queue enables the fast recovery of an unused free small page from a + * partially populated reservation. The reservation at the head of this queue + * is the least recently changed, partially populated reservation. * * Access to this queue is synchronized by the free page queue lock. */ @@ -225,7 +225,7 @@ SYSCTL_PROC(_vm_reserv, OID_AUTO, fullpo static int sysctl_vm_reserv_partpopq(SYSCTL_HANDLER_ARGS); SYSCTL_OID(_vm_reserv, OID_AUTO, partpopq, CTLTYPE_STRING | CTLFLAG_RD, NULL, 0, - sysctl_vm_reserv_partpopq, "A", "Partially-populated reservation queues"); + sysctl_vm_reserv_partpopq, "A", "Partially populated reservation queues"); static long vm_reserv_reclaimed; SYSCTL_LONG(_vm_reserv, OID_AUTO, reclaimed, CTLFLAG_RD, @@ -267,7 +267,7 @@ sysctl_vm_reserv_fullpop(SYSCTL_HANDLER_ } /* - * Describes the current state of the partially-populated reservation queue. + * Describes the current state of the partially populated reservation queue. */ static int sysctl_vm_reserv_partpopq(SYSCTL_HANDLER_ARGS) @@ -301,7 +301,7 @@ sysctl_vm_reserv_partpopq(SYSCTL_HANDLER /* * Reduces the given reservation's population count. If the population count * becomes zero, the reservation is destroyed. Additionally, moves the - * reservation to the tail of the partially-populated reservation queue if the + * reservation to the tail of the partially populated reservation queue if the * population count is non-zero. * * The free page queue lock must be held. @@ -363,7 +363,7 @@ vm_reserv_has_pindex(vm_reserv_t rv, vm_ /* * Increases the given reservation's population count. Moves the reservation - * to the tail of the partially-populated reservation queue. + * to the tail of the partially populated reservation queue. * * The free page queue must be locked. */ @@ -597,7 +597,7 @@ found: } /* - * Allocates a page from an existing or newly-created reservation. + * Allocates a page from an existing or newly created reservation. * * The page "mpred" must immediately precede the offset "pindex" within the * specified object. @@ -721,12 +721,12 @@ found: } /* - * Breaks the given reservation. Except for the specified cached or free - * page, all cached and free pages in the reservation are returned to the - * physical memory allocator. The reservation's population count and map are - * reset to their initial state. + * Breaks the given reservation. Except for the specified free page, all free + * pages in the reservation are returned to the physical memory allocator. + * The reservation's population count and map are reset to their initial + * state. * - * The given reservation must not be in the partially-populated reservation + * The given reservation must not be in the partially populated reservation * queue. The free page queue lock must be held. */ static void @@ -895,7 +895,7 @@ vm_reserv_level(vm_page_t m) } /* - * Returns a reservation level if the given page belongs to a fully-populated + * Returns a reservation level if the given page belongs to a fully populated * reservation and -1 otherwise. */ int @@ -908,8 +908,8 @@ vm_reserv_level_iffullpop(vm_page_t m) } /* - * Breaks the given partially-populated reservation, releasing its cached and - * free pages to the physical memory allocator. + * Breaks the given partially populated reservation, releasing its free pages + * to the physical memory allocator. * * The free page queue lock must be held. */ @@ -927,9 +927,9 @@ vm_reserv_reclaim(vm_reserv_t rv) } /* - * Breaks the reservation at the head of the partially-populated reservation - * queue, releasing its cached and free pages to the physical memory - * allocator. Returns TRUE if a reservation is broken and FALSE otherwise. + * Breaks the reservation at the head of the partially populated reservation + * queue, releasing its free pages to the physical memory allocator. Returns + * TRUE if a reservation is broken and FALSE otherwise. * * The free page queue lock must be held. */ @@ -947,11 +947,10 @@ vm_reserv_reclaim_inactive(void) } /* - * Searches the partially-populated reservation queue for the least recently - * active reservation with unused pages, i.e., cached or free, that satisfy the - * given request for contiguous physical memory. If a satisfactory reservation - * is found, it is broken. Returns TRUE if a reservation is broken and FALSE - * otherwise. + * Searches the partially populated reservation queue for the least recently + * changed reservation with free pages that satisfy the given request for + * contiguous physical memory. If a satisfactory reservation is found, it is + * broken. Returns TRUE if a reservation is broken and FALSE otherwise. * * The free page queue lock must be held. */