- 15 Dec, 2009 40 commits
-
-
KOSAKI Motohiro authored
shrink_all_zone() was introduced by commit d6277db4 (swsusp: rework memory shrinker) for hibernate performance improvement. and sc.swap_cluster_max was introduced by commit a06fe4d307 (Speed freeing memory for suspend). commit a06fe4d307 said Without the patch: Freed 14600 pages in 1749 jiffies = 32.61 MB/s (Anomolous!) Freed 88563 pages in 14719 jiffies = 23.50 MB/s Freed 205734 pages in 32389 jiffies = 24.81 MB/s With the patch: Freed 68252 pages in 496 jiffies = 537.52 MB/s Freed 116464 pages in 569 jiffies = 798.54 MB/s Freed 209699 pages in 705 jiffies = 1161.89 MB/s At that time, their patch was pretty worth. However, Modern Hardware trend and recent VM improvement broke its worth. From several reason, I think we should remove shrink_all_zones() at all. detail: 1) Old days, shrink_zone()'s slowness was mainly caused by stupid io-throttle at no i/o congestion. but current shrink_zone() is sane, not slow. 2) shrink_all_zone() try to shrink all pages at a time. but it doesn't works fine on numa system. example) System has 4GB memory and each node have 2GB. and hibernate need 1GB. optimal) steal 500MB from each node. shrink_all_zones) steal 1GB from node-0. Oh, Cache balancing logic was broken. ;) Unfortunately, Desktop system moved ahead NUMA at nowadays. (Side note, if hibernate require 2GB, shrink_all_zones() never success on above machine) 3) if the node has several I/O flighting pages, shrink_all_zones() makes pretty bad result. schenario) hibernate need 1GB 1) shrink_all_zones() try to reclaim 1GB from Node-0 2) but it only reclaimed 990MB 3) stupidly, shrink_all_zones() try to reclaim 1GB from Node-1 4) it reclaimed 990MB Oh, well. it reclaimed twice much than required. In the other hand, current shrink_zone() has sane baling out logic. then, it doesn't make overkill reclaim. then, we lost shrink_zones()'s risk. 4) SplitLRU VM always keep active/inactive ratio very carefully. inactive list only shrinking break its assumption. it makes unnecessary OOM risk. it obviously suboptimal. Now, shrink_all_memory() is only the wrapper function of do_try_to_free_pages(). it bring good reviewability and debuggability, and solve above problems. side note: Reclaim logic unificication makes two good side effect. - Fix recursive reclaim bug on shrink_all_memory(). it did forgot to use PF_MEMALLOC. it mean the system be able to stuck into deadlock. - Now, shrink_all_memory() got lockdep awareness. it bring good debuggability. Signed-off-by:
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Reviewed-by:
Rik van Riel <riel@redhat.com> Acked-by:
Rafael J. Wysocki <rjw@sisk.pl> Cc: Mel Gorman <mel@csn.ul.ie> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
KOSAKI Motohiro authored
Currently, sc.scap_cluster_max has double meanings. 1) reclaim batch size as isolate_lru_pages()'s argument 2) reclaim baling out thresolds The two meanings pretty unrelated. Thus, Let's separate it. this patch doesn't change any behavior. Signed-off-by:
Minchan Kim <minchan.kim@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Signed-off-by:
-
Alex Chiang authored
Describe NUMA node symlink created for CPUs when CONFIG_NUMA is set. Signed-off-by:
Alex Chiang <achiang@hp.com> Cc: Greg KH <greg@kroah.com> Cc: Randy Dunlap <randy.dunlap@oracle.com> Cc: Gary Hade <garyhade@us.ibm.com> Cc: Badari Pulavarty <pbadari@us.ibm.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Rientjes <rientjes@google.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by:
-
Alex Chiang authored
You can discover which CPUs belong to a NUMA node by examining /sys/devices/system/node/node#/ However, it's not convenient to go in the other direction, when looking at /sys/devices/system/cpu/cpu#/ Yes, you can muck about in sysfs, but adding these symlinks makes life a lot more convenient. Signed-off-by:
David Rientjes <rientjes@google.com> Cc: Gary Hade <garyhade@us.ibm.com> Cc: Badari Pulavarty <pbadari@us.ibm.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Rientjes <rientjes@google.com> Cc: Greg KH <greg@kroah.com> Cc: Randy Dunlap <randy.dunlap@oracle.com> Cc: David Rientjes <rientjes@google.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by:
-
Alex Chiang authored
By returning early if the node is not online, we can unindent the interesting code by two levels. No functional change. Signed-off-by:
-
Alex Chiang authored
By returning early if the node is not online, we can unindent the interesting code by one level. No functional change. Signed-off-by:
-
Alex Chiang authored
Commit c04fc586 (mm: show node to memory section relationship with symlinks in sysfs) created symlinks from nodes to memory sections, e.g. /sys/devices/system/node/node1/memory135 -> ../../memory/memory135 If you're examining the memory section though and are wondering what node it might belong to, you can find it by grovelling around in sysfs, but it's a little cumbersome. Add a reverse symlink for each memory section that points back to the node to which it belongs. Signed-off-by:
-
Hugh Dickins authored
When do_nonlinear_fault() realizes that the page table must have been corrupted for it to have been called, it does print_bad_pte() and returns ... VM_FAULT_OOM, which is hard to understand. It made some sense when I did it for 2.6.15, when do_page_fault() just killed the current process; but nowadays it lets the OOM killer decide who to kill - so page table corruption in one process would be liable to kill another. Change it to return VM_FAULT_SIGBUS instead: that doesn't guarantee that the process will be killed, but is good enough for such a rare abnormality, accompanied as it is by the "BUG: Bad page map" message. And recent HWPOISON work has copied that code into do_swap_page(), when it finds an impossible swap entry: fix that to VM_FAULT_SIGBUS too. Signed-off-by:
Hugh Dickins <hugh.dickins@tiscali.co.uk> Cc: Izik Eidus <ieidus@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Nick Piggin <npiggin@suse.de> Reviewed-by:
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by:
Wu Fengguang <fengguang.wu@intel.com> Reviewed-by:
-
Hugh Dickins authored
CONFIG_DEBUG_SPINLOCK adds 12 or 16 bytes to a 32- or 64-bit spinlock_t, and CONFIG_DEBUG_LOCK_ALLOC adds another 12 or 24 bytes to it: lockdep enables both of those, and CONFIG_LOCK_STAT adds 8 or 16 bytes to that. When 2.6.15 placed the split page table lock inside struct page (usually sized 32 or 56 bytes), only CONFIG_DEBUG_SPINLOCK was a possibility, and we ignored the enlargement (but fitted in CONFIG_GENERIC_LOCKBREAK's 4 by letting the spinlock_t occupy both page->private and page->mapping). Should these debugging options be allowed to double the size of a struct page, when only one minority use of the page (as a page table) needs to fit a spinlock in there? Perhaps not. Take the easy way out: switch off SPLIT_PTLOCK_CPUS when DEBUG_SPINLOCK or DEBUG_LOCK_ALLOC is in force. I've sometimes tried to be cleverer, kmallocing a cacheline for the spinlock when it doesn't fit, but given up each time. Falling back to mm->page_table_lock (as we do when ptlock is not split) lets lockdep check out the strictest path anyway. And now that some arches allow 8192 cpus, use 999999 for infinity. (What has this got to do with KSM swapping? It doesn't care about the size of struct page, but may care about random junk in page->mapping - to be explained separately later.) Signed-off-by:
-
Hugh Dickins authored
KSM swapping will know where page_referenced_one() and try_to_unmap_one() should look. It could hack page->index to get them to do what it wants, but it seems cleaner now to pass the address down to them. Make the same change to page_mkclean_one(), since it follows the same pattern; but there's no real need in its case. Signed-off-by:
-
Hugh Dickins authored
Remove three degrees of obfuscation, left over from when we had CONFIG_UNEVICTABLE_LRU. MLOCK_PAGES is CONFIG_HAVE_MLOCKED_PAGE_BIT is CONFIG_HAVE_MLOCK is CONFIG_MMU. rmap.o (and memory-failure.o) are only built when CONFIG_MMU, so don't need such conditions at all. Somehow, I feel no compulsion to remove the CONFIG_HAVE_MLOCK* lines from 169 defconfigs: leave those to evolve in due course. Signed-off-by:
-
Hugh Dickins authored
There's contorted mlock/munlock handling in try_to_unmap_anon() and try_to_unmap_file(), which we'd prefer not to repeat for KSM swapping. Simplify it by moving it all down into try_to_unmap_one(). One thing is then lost, try_to_munlock()'s distinction between when no vma holds the page mlocked, and when a vma does mlock it, but we could not get mmap_sem to set the page flag. But its only caller takes no interest in that distinction (and is better testing SWAP_MLOCK anyway), so let's keep the code simple and return SWAP_AGAIN for both cases. try_to_unmap_file()'s TTU_MUNLOCK nonlinear handling was particularly amusing: once unravelled, it turns out to have been choosing between two different ways of doing the same nothing. Ah, no, one way was actually returning SWAP_FAIL when it meant to return SWAP_SUCCESS. [kosaki.motohiro@jp.fujitsu.com: comment adding to mlocking in try_to_unmap_one] [akpm@linux-foundation.org: remove test of MLOCK_PAGES] Signed-off-by:
-
Hugh Dickins authored
At present we define PageAnon(page) by the low PAGE_MAPPING_ANON bit set in page->mapping, with the higher bits a pointer to the anon_vma; and have defined PageKsm(page) as that with NULL anon_vma. But KSM swapping will need to store a pointer there: so in preparation for that, now define PAGE_MAPPING_FLAGS as the low two bits, including PAGE_MAPPING_KSM (always set along with PAGE_MAPPING_ANON, until some other use for the bit emerges). Declare page_rmapping(page) to return the pointer part of page->mapping, and page_anon_vma(page) to return the anon_vma pointer when that's what it is. Use these in a few appropriate places: notably, unuse_vma() has been testing page->mapping, but is better to be testing page_anon_vma() (cases may be added in which flag bits are set without any pointer). Signed-off-by:
-
KOSAKI Motohiro authored
If reclaim fails to make sufficient progress, the priority is raised. Once the priority is higher, kswapd starts waiting on congestion. However, if the zone is below the min watermark then kswapd needs to continue working without delay as there is a danger of an increased rate of GFP_ATOMIC allocation failure. This patch changes the conditions under which kswapd waits on congestion by only going to sleep if the min watermarks are being met. [mel@csn.ul.ie: add stats to track how relevant the logic is] [mel@csn.ul.ie: make kswapd only check its own zones and rename the relevant counters] Signed-off-by:
Mel Gorman <mel@csn.ul.ie> Reviewed-by:
-
Mel Gorman authored
After kswapd balances all zones in a pgdat, it goes to sleep. In the event of no IO congestion, kswapd can go to sleep very shortly after the high watermark was reached. If there are a constant stream of allocations from parallel processes, it can mean that kswapd went to sleep too quickly and the high watermark is not being maintained for sufficient length time. This patch makes kswapd go to sleep as a two-stage process. It first tries to sleep for HZ/10. If it is woken up by another process or the high watermark is no longer met, it's considered a premature sleep and kswapd continues work. Otherwise it goes fully to sleep. This adds more counters to distinguish between fast and slow breaches of watermarks. A "fast" premature sleep is one where the low watermark was hit in a very short time after kswapd going to sleep. A "slow" premature sleep indicates that the high watermark was breached after a very short interval. Signed-off-by:
-
Huang Shijie authored
When the code jumps to the `out', `referenced' is still zero. So there is no need to check it. Signed-off-by:
Huang Shijie <shijie8@gmail.com> Acked-by:
-
Huang Shijie authored
Just simplify the code when `mlocked' is true. Signed-off-by:
-
Huang Shijie authored
Fix the comment for try_to_unmap_anon() with the new arguments. Signed-off-by:
-
Vincent Li authored
Commit 543ade1f ("Streamline generic_file_* interfaces and filemap cleanups") removed generic_file_write() in filemap. Change the comment in vmscan pageout() to __generic_file_aio_write(). Signed-off-by:
Vincent Li <macli@brc.ubc.ca> Cc: Christoph Hellwig <hch@lst.de> Signed-off-by:
-
Lee Schermerhorn authored
Seems that page_io.c doesn't really need to know that page_private(page) is the swp_entry 'val'. Rework map_swap_page() to do what its name says and map a page to a page offset in the swap space. The only other caller of map_swap_page() is internal to mm/swapfile.c and it does want to map a swap entry to the 'sector'. So rename map_swap_page() to map_swap_entry(), make it 'static' and and implement map_swap_page() as a wrapper around that. Signed-off-by:
Lee Schermerhorn <lee.schermerhorn@hp.com> Cc: Hugh Dickins <hugh.dickins@tiscali.co.uk> Signed-off-by:
-
Hugh Dickins authored
Reorder (and comment) the fields of swap_info_struct, to make better use of its cachelines: it's good for swap_duplicate() in particular if unsigned int max and swap_map are near the start. Signed-off-by:
-
Hugh Dickins authored
While we're fiddling with the swap_map values, let's assign a particular value to shmem/tmpfs swap pages: their swap counts are never incremented, and it helps swapoff's try_to_unuse() a little if it can immediately distinguish those pages from process pages. Since we've no use for SWAP_MAP_BAD | COUNT_CONTINUED, we might as well use that 0xbf value for SWAP_MAP_SHMEM. Signed-off-by:
-
Hugh Dickins authored
Swap is duplicated (reference count incremented by one) whenever the same swap page is inserted into another mm (when forking finds a swap entry in place of a pte, or when reclaim unmaps a pte to insert the swap entry). swap_info_struct's vmalloc'ed swap_map is the array of these reference counts: but what happens when the unsigned short (or unsigned char since the preceding patch) is full? (and its high bit is kept for a cache flag) We then lose track of it, never freeing, leaving it in use until swapoff: at which point we _hope_ that a single pass will have found all instances, assume there are no more, and will lose user data if we're wrong. Swapping of KSM pages has not yet been enabled; but it is implemented, and makes it very easy for a user to overflow the maximum swap count: possible with ordinary process pages, but unlikely, even when pid_max has been raised from PID_MAX_DEFAULT. This patch implements swap count continuations: when the count overflows, a continuation page is allocated and linked to the original vmalloc'ed map page, and this used to hold the continuation counts for that entry and its neighbours. These continuation pages are seldom referenced: the common paths all work on the original swap_map, only referring to a continuation page when the low "digit" of a count is incremented or decremented through SWAP_MAP_MAX. Signed-off-by:
-
Hugh Dickins authored
Halve the vmalloc'ed swap_map array from unsigned shorts to unsigned chars: it's still very unusual to reach a swap count of 126, and the next patch allows it to be extended indefinitely. Signed-off-by:
-
Hugh Dickins authored
Though swap_count() is useful, I'm finding that swap_has_cache() and encode_swapmap() obscure what happens in the swap_map entry, just at those points where I need to understand it. Remove them, and pass more usable "usage" values to scan_swap_map(), swap_entry_free() and __swap_duplicate(), instead of the SWAP_MAP and SWAP_CACHE enum. Signed-off-by:
-
Hugh Dickins authored
Move CONFIG_HIBERNATION's swapdev_block() into the main CONFIG_HIBERNATION block, remove extraneous whitespace and return, fix typo in a comment. Signed-off-by:
-
Hugh Dickins authored
Make better use of the space by folding first swap_extent into its swap_info_struct, instead of just the list_head: swap partitions need only that one, and for others it's used as a circular list anyway. [jirislaby@gmail.com: fix crash on double swapon] Signed-off-by:
Jiri Slaby <jirislaby@gmail.com> Signed-off-by:
-
Hugh Dickins authored
The swap_info_struct is only 76 or 104 bytes, but it does seem wrong to reserve an array of about 30 of them in bss, when most people will want only one. Change swap_info[] to an array of pointers. That does need a "type" field in the structure: pack it as a char with next type and short prio (aha, char is unsigned by default on PowerPC). Use the (admittedly peculiar) name "type" throughout for this index. /proc/swaps does not take swap_lock: I wouldn't want it to, but do take care with barriers when adding a new item to the array (never removed). Signed-off-by:
-
Hugh Dickins authored
The swap_info_struct is mostly private to mm/swapfile.c, with only one other in-tree user: get_swap_bio(). Adjust its interface to map_swap_page(), so that we can then remove get_swap_info_struct(). But there is a popular user out-of-tree, TuxOnIce: so leave the declaration of swap_info_struct in linux/swap.h. Signed-off-by:
-
Jan Beulich authored
- avoid wasting more precious resources (DMA or DMA32 pools), when being called through vmalloc_32{,_user}() - explicitly allow using high memory here even if the outer allocation request doesn't allow it Signed-off-by:Jan Beulich <jbeulich@novell.com> Acked-by:
-
David Rientjes authored
Objects passed to NODEMASK_ALLOC() are relatively small in size and are backed by slab caches that are not of large order, traditionally never greater than PAGE_ALLOC_COSTLY_ORDER. Thus, using GFP_KERNEL for these allocations on large machines when CONFIG_NODES_SHIFT > 8 will cause the page allocator to loop endlessly in the allocation attempt, each time invoking both direct reclaim or the oom killer. This is of particular interest when using NODEMASK_ALLOC() from a mempolicy context (either directly in mm/mempolicy.c or the mempolicy constrained hugetlb allocations) since the oom killer always kills current when allocations are constrained by mempolicies. So for all present use cases in the kernel, current would end up being oom killed when direct reclaim fails. That would allow the NODEMASK_ALLOC() to succeed but current would have sacrificed itself upon returning. This patch adds gfp flags to NODEMASK_ALLOC() to pass to kmalloc() on CONFIG_NODES_SHIFT > 8; this parameter is a nop on other configurations. All current use cases either directly from hugetlb code or indirectly via NODEMASK_SCRATCH() union __GFP_NORETRY to avoid direct reclaim and the oom killer when the slab allocator needs to allocate additional pages. The side-effect of this change is that all current use cases of either NODEMASK_ALLOC() or NODEMASK_SCRATCH() need appropriate -ENOMEM handling when the allocation fails (never for CONFIG_NODES_SHIFT <= 8). All current use cases were audited and do have appropriate error handling at this time. Signed-off-by:
-
Lee Schermerhorn authored
Offload the registration and unregistration of per node hstate sysfs attributes to a worker thread rather than attempt the allocation/attachment or detachment/freeing of the attributes in the context of the memory hotplug handler. I don't know that this is absolutely required, but the registration can sleep in allocations and other mem hot plug handlers do it this way. If it turns out this is NOT required, we can drop this patch. N.B., Only tested build, boot, libhugetlbfs regression. i.e., no memory hotplug testing. Signed-off-by:Andi Kleen <andi@firstfloor.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Lee Schermerhorn <lee.schermerhorn@hp.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Randy Dunlap <randy.dunlap@oracle.com> Cc: Nishanth Aravamudan <nacc@us.ibm.com> Cc: David Rientjes <rientjes@google.com> Cc: Adam Litke <agl@us.ibm.com> Cc: Andy Whitcroft <apw@canonical.com> Cc: Eric Whitney <eric.whitney@hp.com> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by:
-
Lee Schermerhorn authored
Register per node hstate attributes only for nodes with memory. As suggested by David Rientjes. With Memory Hotplug, memory can be added to a memoryless node and a node with memory can become memoryless. Therefore, add a memory on/off-line notifier callback to [un]register a node's attributes on transition to/from memoryless state. N.B., Only tested build, boot, libhugetlbfs regression. i.e., no memory hotplug testing. Signed-off-by: -
David Rientjes authored
When memory is hot-removed, its node must be cleared in N_HIGH_MEMORY if there are no present pages left. In such a situation, kswapd must also be stopped since it has nothing left to do. Signed-off-by:
-
Lee Schermerhorn authored
Register per node hstate sysfs attributes only for nodes with memory. Global replacement of 'all online nodes" with "all nodes with memory" in mm/hugetlb.c. Suggested by David Rientjes. A subsequent patch will handle adding/removing of per node hstate sysfs attributes when nodes transition to/from memoryless state via memory hotplug. NOTE: this patch has not been tested with memoryless nodes. Signed-off-by:
-
Lee Schermerhorn authored
Update the kernel huge tlb documentation to describe the numa memory policy based huge page management. Additionaly, the patch includes a fair amount of rework to improve consistency, eliminate duplication and set the context for documenting the memory policy interaction. Signed-off-by:
-
Lee Schermerhorn authored
Add the per huge page size control/query attributes to the per node sysdevs: /sys/devices/system/node/node<ID>/hugepages/hugepages-<size>/ nr_hugepages - r/w free_huge_pages - r/o surplus_huge_pages - r/o The patch attempts to re-use/share as much of the existing global hstate attribute initialization and handling, and the "nodes_allowed" constraint processing as possible. Calling set_max_huge_pages() with no node indicates a change to global hstate parameters. In this case, any non-default task mempolicy will be used to generate the nodes_allowed mask. A valid node id indicates an update to that node's hstate parameters, and the count argument specifies the target count for the specified node. From this info, we compute the target global count for the hstate and construct a nodes_allowed node mask contain only the specified node. Setting the node specific nr_hugepages via the per node attribute effectively ignores any task mempolicy or cpuset constraints. With this patch: (me):ls /sys/devices/system/node/node0/hugepages/hugepages-2048kB ./ ../ free_hugepages nr_hugepages surplus_hugepages Starting from: Node 0 HugePages_Total: 0 Node 0 HugePages_Free: 0 Node 0 HugePages_Surp: 0 Node 1 HugePages_Total: 0 Node 1 HugePages_Free: 0 Node 1 HugePages_Surp: 0 Node 2 HugePages_Total: 0 Node 2 HugePages_Free: 0 Node 2 HugePages_Surp: 0 Node 3 HugePages_Total: 0 Node 3 HugePages_Free: 0 Node 3 HugePages_Surp: 0 vm.nr_hugepages = 0 Allocate 16 persistent huge pages on node 2: (me):echo 16 >/sys/devices/system/node/node2/hugepages/hugepages-2048kB/nr_hugepages [Note that this is equivalent to: numactl -m 2 hugeadmin --pool-pages-min 2M:+16 ] Yields: Node 0 HugePages_Total: 0 Node 0 HugePages_Free: 0 Node 0 HugePages_Surp: 0 Node 1 HugePages_Total: 0 Node 1 HugePages_Free: 0 Node 1 HugePages_Surp: 0 Node 2 HugePages_Total: 16 Node 2 HugePages_Free: 16 Node 2 HugePages_Surp: 0 Node 3 HugePages_Total: 0 Node 3 HugePages_Free: 0 Node 3 HugePages_Surp: 0 vm.nr_hugepages = 16 Global controls work as expected--reduce pool to 8 persistent huge pages: (me):echo 8 >/sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages Node 0 HugePages_Total: 0 Node 0 HugePages_Free: 0 Node 0 HugePages_Surp: 0 Node 1 HugePages_Total: 0 Node 1 HugePages_Free: 0 Node 1 HugePages_Surp: 0 Node 2 HugePages_Total: 8 Node 2 HugePages_Free: 8 Node 2 HugePages_Surp: 0 Node 3 HugePages_Total: 0 Node 3 HugePages_Free: 0 Node 3 HugePages_Surp: 0 Signed-off-by:
-
Lee Schermerhorn authored
Move definition of NUMA_NO_NODE from ia64 and x86_64 arch specific headers to generic header 'linux/numa.h' for use in generic code. NUMA_NO_NODE replaces bare '-1' where it's used in this series to indicate "no node id specified". Ultimately, it can be used to replace the -1 elsewhere where it is used similarly. Signed-off-by:
-
Lee Schermerhorn authored
This patch derives a "nodes_allowed" node mask from the numa mempolicy of the task modifying the number of persistent huge pages to control the allocation, freeing and adjusting of surplus huge pages when the pool page count is modified via the new sysctl or sysfs attribute "nr_hugepages_mempolicy". The nodes_allowed mask is derived as follows: * For "default" [NULL] task mempolicy, a NULL nodemask_t pointer is produced. This will cause the hugetlb subsystem to use node_online_map as the "nodes_allowed". This preserves the behavior before this patch. * For "preferred" mempolicy, including explicit local allocation, a nodemask with the single preferred node will be produced. "local" policy will NOT track any internode migrations of the task adjusting nr_hugepages. * For "bind" and "interleave" policy, the mempolicy's nodemask will be used. * Other than to inform the construction of the nodes_allowed node mask, the actual mempolicy mode is ignored. That is, all modes behave like interleave over the resulting nodes_allowed mask with no "fallback". See the updated documentation [next patch] for more information about the implications of this patch. Examples: Starting with: Node 0 HugePages_Total: 0 Node 1 HugePages_Total: 0 Node 2 HugePages_Total: 0 Node 3 HugePages_Total: 0 Default behavior [with or without this patch] balances persistent hugepage allocation across nodes [with sufficient contiguous memory]: sysctl vm.nr_hugepages[_mempolicy]=32 yields: Node 0 HugePages_Total: 8 Node 1 HugePages_Total: 8 Node 2 HugePages_Total: 8 Node 3 HugePages_Total: 8 Of course, we only have nr_hugepages_mempolicy with the patch, but with default mempolicy, nr_hugepages_mempolicy behaves the same as nr_hugepages. Applying mempolicy--e.g., with numactl [using '-m' a.k.a. '--membind' because it allows multiple nodes to be specified and it's easy to type]--we can allocate huge pages on individual nodes or sets of nodes. So, starting from the condition above, with 8 huge pages per node, add 8 more to node 2 using: numactl -m 2 sysctl vm.nr_hugepages_mempolicy=40 This yields: Node 0 HugePages_Total: 8 Node 1 HugePages_Total: 8 Node 2 HugePages_Total: 16 Node 3 HugePages_Total: 8 The incremental 8 huge pages were restricted to node 2 by the specified mempolicy. Similarly, we can use mempolicy to free persistent huge pages from specified nodes: numactl -m 0,1 sysctl vm.nr_hugepages_mempolicy=32 yields: Node 0 HugePages_Total: 4 Node 1 HugePages_Total: 4 Node 2 HugePages_Total: 16 Node 3 HugePages_Total: 8 The 8 huge pages freed were balanced over nodes 0 and 1. [rientjes@google.com: accomodate reworked NODEMASK_ALLOC] Signed-off-by:
-
Lee Schermerhorn authored
Factor init_nodemask_of_node() out of the nodemask_of_node() macro. This will be used to populate the huge pages "nodes_allowed" nodemask for a single node when basing nodes_allowed on a preferred/local mempolicy or when a persistent huge page pool page count is modified via a per node sysfs attribute. Signed-off-by:
-
