- 04 Jun, 2014 40 commits
-
-
Matthew Wilcox authored
By calling the device driver to write the page directly, we avoid allocating a BIO, which allows us to free memory without allocating memory. [akpm@linux-foundation.org: fix used-uninitialized bug] Signed-off-by:
Matthew Wilcox <matthew.r.wilcox@intel.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Dheeraj Reddy <dheeraj.reddy@intel.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Matthew Wilcox authored
A block device driver may choose to provide a rw_page operation. These will be called when the filesystem is attempting to do page sized I/O to page cache pages (ie not for direct I/O). This does preclude I/Os that are larger than page size, so this may only be a performance gain for some devices. Signed-off-by:
Dheeraj Reddy <dheeraj.reddy@intel.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Hugh Dickins <hughd@google.com> Signed-off-by:
-
Matthew Wilcox authored
page_endio() takes care of updating all the appropriate page flags once I/O has finished to a page. Switch to using mapping_set_error() instead of setting AS_EIO directly; this will handle thin-provisioned devices correctly. Signed-off-by:
-
Matthew Wilcox authored
__mpage_writepage() is over 200 lines long, has 20 local variables, four goto labels and could desperately use simplification. Splitting clean_buffers() into a helper function improves matters a little, removing 20+ lines from it. Signed-off-by:
-
Matthew Wilcox authored
The last in-tree caller of block_write_full_page_endio() was removed in January 2013. It's time to remove the EXPORT_SYMBOL, which leaves block_write_full_page() as the only caller of block_write_full_page_endio(), so inline block_write_full_page_endio() into block_write_full_page(). Signed-off-by:
-
NeilBrown authored
When a loopback NFS mount is active and the backing device for the NFS mount becomes congested, that can impose throttling delays on the nfsd threads. These delays significantly reduce throughput and so the NFS mount remains congested. This results in a livelock and the reduced throughput persists. This livelock has been found in testing with the 'wait_iff_congested' call, and could possibly be caused by the 'congestion_wait' call. This livelock is similar to the deadlock which justified the introduction of PF_LESS_THROTTLE, and the same flag can be used to remove this livelock. To minimise the impact of the change, we still throttle nfsd when the filesystem it is writing to is congested, but not when some separate filesystem (e.g. the NFS filesystem) is congested. Signed-off-by:
NeilBrown <neilb@suse.de> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by:
-
Mel Gorman authored
Migration of misplaced transhuge pages uses page_add_new_anon_rmap() when putting the page back as it avoided an atomic operations and added the new page to the correct LRU. A side-effect is that the page gets marked activated as part of the migration meaning that transhuge and base pages are treated differently from an aging perspective than base page migration. This patch uses page_add_anon_rmap() and putback_lru_page() on completion of a transhuge migration similar to base page migration. It would require fewer atomic operations to use lru_cache_add without taking an additional reference to the page. The downside would be that it's still different to base page migration and unevictable pages may be added to the wrong LRU for cleaning up later. Testing of the usual workloads did not show any adverse impact to the change. Signed-off-by:
Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Sasha Levin <sasha.levin@oracle.com> Acked-by:
Hugh Dickins <hughd@google.com> Signed-off-by:
-
Vladimir Davydov authored
At present, we have the following mutexes protecting data related to per memcg kmem caches: - slab_mutex. This one is held during the whole kmem cache creation and destruction paths. We also take it when updating per root cache memcg_caches arrays (see memcg_update_all_caches). As a result, taking it guarantees there will be no changes to any kmem cache (including per memcg). Why do we need something else then? The point is it is private to slab implementation and has some internal dependencies with other mutexes (get_online_cpus). So we just don't want to rely upon it and prefer to introduce additional mutexes instead. - activate_kmem_mutex. Initially it was added to synchronize initializing kmem limit (memcg_activate_kmem). However, since we can grow per root cache memcg_caches arrays only on kmem limit initialization (see memcg_update_all_caches), we also employ it to protect against memcg_caches arrays relocation (e.g. see __kmem_cache_destroy_memcg_children). - We have a convention not to take slab_mutex in memcontrol.c, but we want to walk over per memcg memcg_slab_caches lists there (e.g. for destroying all memcg caches on offline). So we have per memcg slab_caches_mutex's protecting those lists. The mutexes are taken in the following order: activate_kmem_mutex -> slab_mutex -> memcg::slab_caches_mutex Such a syncrhonization scheme has a number of flaws, for instance: - We can't call kmem_cache_{destroy,shrink} while walking over a memcg::memcg_slab_caches list due to locking order. As a result, in mem_cgroup_destroy_all_caches we schedule the memcg_cache_params::destroy work shrinking and destroying the cache. - We don't have a mutex to synchronize per memcg caches destruction between memcg offline (mem_cgroup_destroy_all_caches) and root cache destruction (__kmem_cache_destroy_memcg_children). Currently we just don't bother about it. This patch simplifies it by substituting per memcg slab_caches_mutex's with the global memcg_slab_mutex. It will be held whenever a new per memcg cache is created or destroyed, so it protects per root cache memcg_caches arrays and per memcg memcg_slab_caches lists. The locking order is following: activate_kmem_mutex -> memcg_slab_mutex -> slab_mutex This allows us to call kmem_cache_{create,shrink,destroy} under the memcg_slab_mutex. As a result, we don't need memcg_cache_params::destroy work any more - we can simply destroy caches while iterating over a per memcg slab caches list. Also using the global mutex simplifies synchronization between concurrent per memcg caches creation/destruction, e.g. mem_cgroup_destroy_all_caches vs __kmem_cache_destroy_memcg_children. The downside of this is that we substitute per-memcg slab_caches_mutex's with a hummer-like global mutex, but since we already take either the slab_mutex or the cgroup_mutex along with a memcg::slab_caches_mutex, it shouldn't hurt concurrency a lot. Signed-off-by:Vladimir Davydov <vdavydov@parallels.com> Acked-by:
Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Pekka Enberg <penberg@kernel.org> Signed-off-by:
-
Vladimir Davydov authored
Currently we have two pairs of kmemcg-related functions that are called on slab alloc/free. The first is memcg_{bind,release}_pages that count the total number of pages allocated on a kmem cache. The second is memcg_{un}charge_slab that {un}charge slab pages to kmemcg resource counter. Let's just merge them to keep the code clean. Signed-off-by: -
Vladimir Davydov authored
This patchset is a part of preparations for kmemcg re-parenting. It targets at simplifying kmemcg work-flows and synchronization. First, it removes async per memcg cache destruction (see patches 1, 2). Now caches are only destroyed on memcg offline. That means the caches that are not empty on memcg offline will be leaked. However, they are already leaked, because memcg_cache_params::nr_pages normally never drops to 0 so the destruction work is never scheduled except kmem_cache_shrink is called explicitly. In the future I'm planning reaping such dead caches on vmpressure or periodically. Second, it substitutes per memcg slab_caches_mutex's with the global memcg_slab_mutex, which should be taken during the whole per memcg cache creation/destruction path before the slab_mutex (see patch 3). This greatly simplifies synchronization among various per memcg cache creation/destruction paths. I'm still not quite sure about the end picture, in particular I don't know whether we should reap dead memcgs' kmem caches periodically or try to merge them with their parents (see https://lkml.org/lkml/2014/4/20/38 for more details), but whichever way we choose, this set looks like a reasonable change to me, because it greatly simplifies kmemcg work-flows and eases further development. This patch (of 3): After a memcg is offlined, we mark its kmem caches that cannot be deleted right now due to pending objects as dead by setting the memcg_cache_params::dead flag, so that memcg_release_pages will schedule cache destruction (memcg_cache_params::destroy) as soon as the last slab of the cache is freed (memcg_cache_params::nr_pages drops to zero). I guess the idea was to destroy the caches as soon as possible, i.e. immediately after freeing the last object. However, it just doesn't work that way, because kmem caches always preserve some pages for the sake of performance, so that nr_pages never gets to zero unless the cache is shrunk explicitly using kmem_cache_shrink. Of course, we could account the total number of objects on the cache or check if all the slabs allocated for the cache are empty on kmem_cache_free and schedule destruction if so, but that would be too costly. Thus we have a piece of code that works only when we explicitly call kmem_cache_shrink, but complicates the whole picture a lot. Moreover, it's racy in fact. For instance, kmem_cache_shrink may free the last slab and thus schedule cache destruction before it finishes checking that the cache is empty, which can lead to use-after-free. So I propose to remove this async cache destruction from memcg_release_pages, and check if the cache is empty explicitly after calling kmem_cache_shrink instead. This will simplify things a lot w/o introducing any functional changes. And regarding dead memcg caches (i.e. those that are left hanging around after memcg offline for they have objects), I suppose we should reap them either periodically or on vmpressure as Glauber suggested initially. I'm going to implement this later. Signed-off-by:
-
Michal Hocko authored
Eric has reported that he can see task(s) stuck in memcg OOM handler regularly. The only way out is to echo 0 > $GROUP/memory.oom_control His usecase is: - Setup a hierarchy with memory and the freezer (disable kernel oom and have a process watch for oom). - In that memory cgroup add a process with one thread per cpu. - In one thread slowly allocate once per second I think it is 16M of ram and mlock and dirty it (just to force the pages into ram and stay there). - When oom is achieved loop: * attempt to freeze all of the tasks. * if frozen send every task SIGKILL, unfreeze, remove the directory in cgroupfs. Eric has then pinpointed the issue to be memcg specific. All tasks are sitting on the memcg_oom_waitq when memcg oom is disabled. Those that have received fatal signal will bypass the charge and should continue on their way out. The tricky part is that the exit path might trigger a page fault (e.g. exit_robust_list), thus the memcg charge, while its memcg is still under OOM because nobody has released any charges yet. Unlike with the in-kernel OOM handler the exiting task doesn't get TIF_MEMDIE set so it doesn't shortcut further charges of the killed task and falls to the memcg OOM again without any way out of it as there are no fatal signals pending anymore. This patch fixes the issue by checking PF_EXITING early in mem_cgroup_try_charge and bypass the charge same as if it had fatal signal pending or TIF_MEMDIE set. Normally exiting tasks (aka not killed) will bypass the charge now but this should be OK as the task is leaving and will release memory and increasing the memory pressure just to release it in a moment seems dubious wasting of cycles. Besides that charges after exit_signals should be rare. I am bringing this patch again (rebased on the current mmotm tree). I hope we can move forward finally. If there is still an opposition then I would really appreciate a concurrent approach so that we can discuss alternatives. http://comments.gmane.org/gmane.linux.kernel.stable/77650 is a reference to the followup discussion when the patch has been dropped from the mmotm last time. Reported-by:Eric W. Biederman <ebiederm@xmission.com> Signed-off-by:
Michal Hocko <mhocko@suse.cz> Acked-by:
David Rientjes <rientjes@google.com> Acked-by:
-
Mel Gorman authored
throttle_direct_reclaim() is meant to trigger during swap-over-network during which the min watermark is treated as a pfmemalloc reserve. It throttes on the first node in the zonelist but this is flawed. The user-visible impact is that a process running on CPU whose local memory node has no ZONE_NORMAL will stall for prolonged periods of time, possibly indefintely. This is due to throttle_direct_reclaim thinking the pfmemalloc reserves are depleted when in fact they don't exist on that node. On a NUMA machine running a 32-bit kernel (I know) allocation requests from CPUs on node 1 would detect no pfmemalloc reserves and the process gets throttled. This patch adjusts throttling of direct reclaim to throttle based on the first node in the zonelist that has a usable ZONE_NORMAL or lower zone. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by:
-
Oleg Nesterov authored
CONFIG_MM_OWNER makes no sense. It is not user-selectable, it is only selected by CONFIG_MEMCG automatically. So we can kill this option in init/Kconfig and do s/CONFIG_MM_OWNER/CONFIG_MEMCG/ globally. Signed-off-by:
Oleg Nesterov <oleg@redhat.com> Acked-by:
-
Huang Shijie authored
Remove the first mapping check for vma_link. Move the mutex_lock into the braces when vma->vm_file is true. Signed-off-by:
Huang Shijie <b32955@freescale.com> Signed-off-by:
-
Jianyu Zhan authored
In mm/swap.c, __lru_cache_add() is exported, but actually there are no users outside this file. This patch unexports __lru_cache_add(), and makes it static. It also exports lru_cache_add_file(), as it is use by cifs and fuse, which can loaded as modules. Signed-off-by:
Jianyu Zhan <nasa4836@gmail.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shaohua Li <shli@kernel.org> Cc: Bob Liu <bob.liu@oracle.com> Cc: Seth Jennings <sjenning@linux.vnet.ibm.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Rafael Aquini <aquini@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Acked-by:
Rik van Riel <riel@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Khalid Aziz <khalid.aziz@oracle.com> Cc: Christoph Hellwig <hch@lst.de> Reviewed-by:
Zhang Yanfei <zhangyanfei@cn.fujitsu.com> Signed-off-by:
-
Jonathan Gonzalez V authored
Performing vma lookups without taking the mm->mmap_sem is asking for trouble. While doing the search, the vma in question can be modified or even removed before returning to the caller. Take the lock (exclusively) in order to avoid races while iterating through the vmacache and/or rbtree. Signed-off-by:
Jonathan Gonzalez V <zeus@gnu.org> Signed-off-by:
Davidlohr Bueso <davidlohr@hp.com> Cc: Inki Dae <inki.dae@samsung.com> Cc: Joonyoung Shim <jy0922.shim@samsung.com> Cc: David Airlie <airlied@linux.ie> Signed-off-by:
-
Davidlohr Bueso authored
Performing vma lookups without taking the mm->mmap_sem is asking for trouble. While doing the search, the vma in question can be modified or even removed before returning to the caller. Take the lock (shared) in order to avoid races while iterating through the vmacache and/or rbtree. [akpm@linux-foundation.org: CSE current->active_mm, per Vineet] Signed-off-by:
Vineet Gupta <vgupta@synopsys.com> Signed-off-by:
-
Vladimir Davydov authored
Kmemcg is currently under development and lacks some important features. In particular, it does not have support of kmem reclaim on memory pressure inside cgroup, which practically makes it unusable in real life. Let's warn about it in both Kconfig and Documentation to prevent complaints arising. Signed-off-by:
-
Dave Hansen authored
Nobody outputs memory addresses in decimal. PFNs are essentially addresses, and they're gibberish in decimal. Output them in hex. Also, add the nid and zone name to give a little more context to the message. Signed-off-by:
Dave Hansen <dave.hansen@linux.intel.com> Acked-by:
-
Vlastimil Babka authored
isolate_freepages() is currently somewhat hard to follow thanks to many looks like it is related to the 'low_pfn' variable, but in fact it is not. This patch renames the 'high_pfn' variable to a hopefully less confusing name, and slightly changes its handling without a functional change. A comment made obsolete by recent changes is also updated. [akpm@linux-foundation.org: comment fixes, per Minchan] [iamjoonsoo.kim@lge.com: cleanups] Signed-off-by:
Vlastimil Babka <vbabka@suse.cz> Cc: Minchan Kim <minchan@kernel.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Christoph Lameter <cl@linux.com> Cc: Rik van Riel <riel@redhat.com> Cc: Dongjun Shin <d.j.shin@samsung.com> Cc: Sunghwan Yun <sunghwan.yun@samsung.com> Signed-off-by:
Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by:
-
Heesub Shin authored
Remove code lines currently not in use or never called. Signed-off-by:
Heesub Shin <heesub.shin@samsung.com> Acked-by:
-
Vlastimil Babka authored
For the MIGRATE_RESERVE pages, it is useful when they do not get misplaced on free_list of other migratetype, otherwise they might get allocated prematurely and e.g. fragment the MIGRATE_RESEVE pageblocks. While this cannot be avoided completely when allocating new MIGRATE_RESERVE pageblocks in min_free_kbytes sysctl handler, we should prevent the misplacement where possible. Currently, it is possible for the misplacement to happen when a MIGRATE_RESERVE page is allocated on pcplist through rmqueue_bulk() as a fallback for other desired migratetype, and then later freed back through free_pcppages_bulk() without being actually used. This happens because free_pcppages_bulk() uses get_freepage_migratetype() to choose the free_list, and rmqueue_bulk() calls set_freepage_migratetype() with the *desired* migratetype and not the page's original MIGRATE_RESERVE migratetype. This patch fixes the problem by moving the call to set_freepage_migratetype() from rmqueue_bulk() down to __rmqueue_smallest() and __rmqueue_fallback() where the actual page's migratetype (e.g. from which free_list the page is taken from) is used. Note that this migratetype might be different from the pageblock's migratetype due to freepage stealing decisions. This is OK, as page stealing never uses MIGRATE_RESERVE as a fallback, and also takes care to leave all MIGRATE_CMA pages on the correct freelist. Therefore, as an additional benefit, the call to get_pageblock_migratetype() from rmqueue_bulk() when CMA is enabled, can be removed completely. This relies on the fact that MIGRATE_CMA pageblocks are created only during system init, and the above. The related is_migrate_isolate() check is also unnecessary, as memory isolation has other ways to move pages between freelists, and drain pcp lists containing pages that should be isolated. The buffered_rmqueue() can also benefit from calling get_freepage_migratetype() instead of get_pageblock_migratetype(). Signed-off-by:
Yong-Taek Lee <ytk.lee@samsung.com> Reported-by:
Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Suggested-by:
Minchan Kim <minchan@kernel.org> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Cc: Hugh Dickins <hughd@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: "Wang, Yalin" <Yalin.Wang@sonymobile.com> Signed-off-by:
-
Andrew Morton authored
Signed-off-by:
-
Vladimir Davydov authored
When we create a sl[au]b cache, we allocate kmem_cache_node structures for each online NUMA node. To handle nodes taken online/offline, we register memory hotplug notifier and allocate/free kmem_cache_node corresponding to the node that changes its state for each kmem cache. To synchronize between the two paths we hold the slab_mutex during both the cache creationg/destruction path and while tuning per-node parts of kmem caches in memory hotplug handler, but that's not quite right, because it does not guarantee that a newly created cache will have all kmem_cache_nodes initialized in case it races with memory hotplug. For instance, in case of slub: CPU0 CPU1 ---- ---- kmem_cache_create: online_pages: __kmem_cache_create: slab_memory_callback: slab_mem_going_online_callback: lock slab_mutex for each slab_caches list entry allocate kmem_cache node unlock slab_mutex lock slab_mutex init_kmem_cache_nodes: for_each_node_state(node, N_NORMAL_MEMORY) allocate kmem_cache node add kmem_cache to slab_caches list unlock slab_mutex online_pages (continued): node_states_set_node As a result we'll get a kmem cache with not all kmem_cache_nodes allocated. To avoid issues like that we should hold get/put_online_mems() during the whole kmem cache creation/destruction/shrink paths, just like we deal with cpu hotplug. This patch does the trick. Note, that after it's applied, there is no need in taking the slab_mutex for kmem_cache_shrink any more, so it is removed from there. Signed-off-by: -
Vladimir Davydov authored
kmem_cache_{create,destroy,shrink} need to get a stable value of cpu/node online mask, because they init/destroy/access per-cpu/node kmem_cache parts, which can be allocated or destroyed on cpu/mem hotplug. To protect against cpu hotplug, these functions use {get,put}_online_cpus. However, they do nothing to synchronize with memory hotplug - taking the slab_mutex does not eliminate the possibility of race as described in patch 2. What we need there is something like get_online_cpus, but for memory. We already have lock_memory_hotplug, which serves for the purpose, but it's a bit of a hammer right now, because it's backed by a mutex. As a result, it imposes some limitations to locking order, which are not desirable, and can't be used just like get_online_cpus. That's why in patch 1 I substitute it with get/put_online_mems, which work exactly like get/put_online_cpus except they block not cpu, but memory hotplug. [ v1 can be found at https://lkml.org/lkml/2014/4/6/68. I NAK'ed it by myself, because it used an rw semaphore for get/put_online_mems, making them dead lock prune. ] This patch (of 2): {un}lock_memory_hotplug, which is used to synchronize against memory hotplug, is currently backed by a mutex, which makes it a bit of a hammer - threads that only want to get a stable value of online nodes mask won't be able to proceed concurrently. Also, it imposes some strong locking ordering rules on it, which narrows down the set of its usage scenarios. This patch introduces get/put_online_mems, which are the same as get/put_online_cpus, but for memory hotplug, i.e. executing a code inside a get/put_online_mems section will guarantee a stable value of online nodes, present pages, etc. lock_memory_hotplug()/unlock_memory_hotplug() are removed altogether. Signed-off-by: -
Vladimir Davydov authored
It is only used in slab and should not be used anywhere else so there is no need in exporting it. Signed-off-by:
-
Mel Gorman authored
pgdat->reclaim_nodes tracks if a remote node is allowed to be reclaimed by zone_reclaim due to its distance. As it is expected that zone_reclaim_mode will be rarely enabled it is unreasonable for all machines to take a penalty. Fortunately, the zone_reclaim_mode() path is already slow and it is the path that takes the hit. Signed-off-by:
Christoph Lameter <cl@linux.com> Signed-off-by:
-
Mel Gorman authored
When it was introduced, zone_reclaim_mode made sense as NUMA distances punished and workloads were generally partitioned to fit into a NUMA node. NUMA machines are now common but few of the workloads are NUMA-aware and it's routine to see major performance degradation due to zone_reclaim_mode being enabled but relatively few can identify the problem. Those that require zone_reclaim_mode are likely to be able to detect when it needs to be enabled and tune appropriately so lets have a sensible default for the bulk of users. This patch (of 2): zone_reclaim_mode causes processes to prefer reclaiming memory from local node instead of spilling over to other nodes. This made sense initially when NUMA machines were almost exclusively HPC and the workload was partitioned into nodes. The NUMA penalties were sufficiently high to justify reclaiming the memory. On current machines and workloads it is often the case that zone_reclaim_mode destroys performance but not all users know how to detect this. Favour the common case and disable it by default. Users that are sophisticated enough to know they need zone_reclaim_mode will detect it. Signed-off-by:
-
Luiz Capitulino authored
HugeTLB is limited to allocating hugepages whose size are less than MAX_ORDER order. This is so because HugeTLB allocates hugepages via the buddy allocator. Gigantic pages (that is, pages whose size is greater than MAX_ORDER order) have to be allocated at boottime. However, boottime allocation has at least two serious problems. First, it doesn't support NUMA and second, gigantic pages allocated at boottime can't be freed. This commit solves both issues by adding support for allocating gigantic pages during runtime. It works just like regular sized hugepages, meaning that the interface in sysfs is the same, it supports NUMA, and gigantic pages can be freed. For example, on x86_64 gigantic pages are 1GB big. To allocate two 1G gigantic pages on node 1, one can do: # echo 2 > \ /sys/devices/system/node/node1/hugepages/hugepages-1048576kB/nr_hugepages And to free them all: # echo 0 > \ /sys/devices/system/node/node1/hugepages/hugepages-1048576kB/nr_hugepages The one problem with gigantic page allocation at runtime is that it can't be serviced by the buddy allocator. To overcome that problem, this commit scans all zones from a node looking for a large enough contiguous region. When one is found, it's allocated by using CMA, that is, we call alloc_contig_range() to do the actual allocation. For example, on x86_64 we scan all zones looking for a 1GB contiguous region. When one is found, it's allocated by alloc_contig_range(). One expected issue with that approach is that such gigantic contiguous regions tend to vanish as runtime goes by. The best way to avoid this for now is to make gigantic page allocations very early during system boot, say from a init script. Other possible optimization include using compaction, which is supported by CMA but is not explicitly used by this commit. It's also important to note the following: 1. Gigantic pages allocated at boottime by the hugepages= command-line option can be freed at runtime just fine 2. This commit adds support for gigantic pages only to x86_64. The reason is that I don't have access to nor experience with other archs. The code is arch indepedent though, so it should be simple to add support to different archs 3. I didn't add support for hugepage overcommit, that is allocating a gigantic page on demand when /proc/sys/vm/nr_overcommit_hugepages > 0. The reason is that I don't think it's reasonable to do the hard and long work required for allocating a gigantic page at fault time. But it should be simple to add this if wanted [akpm@linux-foundation.org: coding-style fixes] Signed-off-by:Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by:
Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Reviewed-by:
Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Rik van Riel <riel@redhat.com> Cc: Yinghai Lu <yinghai@kernel.org> Signed-off-by:
-
Luiz Capitulino authored
Next commit will add new code which will want to call for_each_node_mask_to_alloc() macro. Move it, its buddy for_each_node_mask_to_free() and their dependencies up in the file so the new code can use them. This is just code movement, no logic change. Signed-off-by:
Andrea Arcangeli <aarcange@redhat.com> Reviewed-by:
Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Reviewed-by:
-
Luiz Capitulino authored
Hugepages pages never get the PG_reserved bit set, so don't clear it. However, note that if the bit gets mistakenly set free_pages_check() will catch it. Signed-off-by:
-
Luiz Capitulino authored
Signed-off-by:
-
Luiz Capitulino authored
The HugeTLB subsystem uses the buddy allocator to allocate hugepages during runtime. This means that hugepages allocation during runtime is limited to MAX_ORDER order. For archs supporting gigantic pages (that is, page sizes greater than MAX_ORDER), this in turn means that those pages can't be allocated at runtime. HugeTLB supports gigantic page allocation during boottime, via the boot allocator. To this end the kernel provides the command-line options hugepagesz= and hugepages=, which can be used to instruct the kernel to allocate N gigantic pages during boot. For example, x86_64 supports 2M and 1G hugepages, but only 2M hugepages can be allocated and freed at runtime. If one wants to allocate 1G gigantic pages, this has to be done at boot via the hugepagesz= and hugepages= command-line options. Now, gigantic page allocation at boottime has two serious problems: 1. Boottime allocation is not NUMA aware. On a NUMA machine the kernel evenly distributes boottime allocated hugepages among nodes. For example, suppose you have a four-node NUMA machine and want to allocate four 1G gigantic pages at boottime. The kernel will allocate one gigantic page per node. On the other hand, we do have users who want to be able to specify which NUMA node gigantic pages should allocated from. So that they can place virtual machines on a specific NUMA node. 2. Gigantic pages allocated at boottime can't be freed At this point it's important to observe that regular hugepages allocated at runtime don't have those problems. This is so because HugeTLB interface for runtime allocation in sysfs supports NUMA and runtime allocated pages can be freed just fine via the buddy allocator. This series adds support for allocating gigantic pages at runtime. It does so by allocating gigantic pages via CMA instead of the buddy allocator. Releasing gigantic pages is also supported via CMA. As this series builds on top of the existing HugeTLB interface, it makes gigantic page allocation and releasing just like regular sized hugepages. This also means that NUMA support just works. For example, to allocate two 1G gigantic pages on node 1, one can do: # echo 2 > \ /sys/devices/system/node/node1/hugepages/hugepages-1048576kB/nr_hugepages And, to release all gigantic pages on the same node: # echo 0 > \ /sys/devices/system/node/node1/hugepages/hugepages-1048576kB/nr_hugepages Please, refer to patch 5/5 for full technical details. Finally, please note that this series is a follow up for a previous series that tried to extend the command-line options set to be NUMA aware: http://marc.info/?l=linux-mm&m=139593335312191&w=2 During the discussion of that series it was agreed that having runtime allocation support for gigantic pages was a better solution. This patch (of 5): This function is going to be used by non-init code in a future commit. Signed-off-by: -
Duan Jiong authored
Fix a coccinelle error regarding usage of IS_ERR and PTR_ERR instead of PTR_ERR_OR_ZERO. Signed-off-by:
Duan Jiong <duanj.fnst@cn.fujitsu.com> Acked-by:
-
Vladimir Davydov authored
Signed-off-by:
-
Li Zhong authored
Seems we all agree that information about SECTION, e.g. section size, sections per memory block should be kept as kernel internals, and not exposed to userspace. This patch updates Documentation/memory-hotplug.txt to refer to memory blocks instead of memory sections where appropriate and added a paragraph to explain that memory blocks are made of memory sections. The documentation update is mostly provided by Nathan. Also, as end_phys_index in code is actually not the end section id, but the end memory block id, which should always be the same as phys_index. So it is removed here. Signed-off-by:
Li Zhong <zhong@linux.vnet.ibm.com> Reviewed-by:
-
Dave Hansen authored
I recently added a patch to let folks pass a "reason" string dump_page() which gets dumped out along with the page's data. This essentially saves the bug-reader a trip in to the source to figure out why we BUG_ON()'d. The new VM_BUG_ON_PAGE() passes in NULL for "reason". It seems like we might as well pass the BUG_ON() condition if we have it. This will bloat kernels a bit with ~160 new strings, but this is all under a debugging option anyway. page:ffffea0008560280 count:1 mapcount:0 mapping:(null) index:0x0 page flags: 0xbfffc0000000001(locked) page dumped because: VM_BUG_ON_PAGE(PageLocked(page)) ------------[ cut here ]------------ kernel BUG at /home/davehans/linux.git/mm/filemap.c:464! invalid opcode: 0000 [#1] SMP CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.14.0+ #251 Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 ... [akpm@linux-foundation.org: include stringify.h] Signed-off-by:
-
Johannes Weiner authored
Per-memcg swappiness and oom killing can currently not be tweaked on a memcg that is part of a hierarchy, but not the root of that hierarchy. Users have complained that they can't configure this when they turned on hierarchy mode. In fact, with hierarchy mode becoming the default, this restriction disables the tunables entirely. But there is no good reason for this restriction. The settings for swappiness and OOM killing are taken from whatever memcg whose limit triggered reclaim and OOM invocation, regardless of its position in the hierarchy tree. Allow setting swappiness on any group. The knob on the root memcg already reads the global VM swappiness, make it writable as well. Allow disabling the OOM killer on any non-root memcg. Signed-off-by:
-
Sebastian Ott authored
Memory obtained via mempool_alloc is not always zeroed even when called with __GFP_ZERO. Add a note and VM_BUG_ON statement to make that clear. [akpm@linux-foundation.org: use VM_WARN_ON_ONCE] Signed-off-by:
Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by:
-
Andrew Morton authored
WARN_ON() and WARN_ON_ONCE(), dependent on CONFIG_DEBUG_VM Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Signed-off-by:
-
