- 28 Dec, 2018 40 commits
-
-
Jérôme Glisse authored
Patch series "mmu notifier contextual informations", v2. This patchset adds contextual information, why an invalidation is happening, to mmu notifier callback. This is necessary for user of mmu notifier that wish to maintains their own data structure without having to add new fields to struct vm_area_struct (vma). For instance device can have they own page table that mirror the process address space. When a vma is unmap (munmap() syscall) the device driver can free the device page table for the range. Today we do not have any information on why a mmu notifier call back is happening and thus device driver have to assume that it is always an munmap(). This is inefficient at it means that it needs to re-allocate device page table on next page fault and rebuild the whole device driver data structure for the range. Other use case beside munmap() also exist, for instance it is pointless for device driver to invalidate the device page table when the invalidation is for the soft dirtyness tracking. Or device driver can optimize away mprotect() that change the page table permission access for the range. This patchset enables all this optimizations for device drivers. I do not include any of those in this series but another patchset I am posting will leverage this. The patchset is pretty simple from a code point of view. The first two patches consolidate all mmu notifier arguments into a struct so that it is easier to add/change arguments. The last patch adds the contextual information (munmap, protection, soft dirty, clear, ...). This patch (of 3): To avoid having to change many callback definition everytime we want to add a parameter use a structure to group all parameters for the mmu_notifier invalidate_range_start/end callback. No functional changes with this patch. [akpm@linux-foundation.org: fix drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c kerneldoc] Link: http://lkml.kernel.org/r/20181205053628.3210-2-jglisse@redhat.comSigned-off-by:
Jérôme Glisse <jglisse@redhat.com> Acked-by:
Jan Kara <jack@suse.cz> Acked-by: Jason Gunthorpe <jgg@mellanox.com> [infiniband] Cc: Matthew Wilcox <mawilcox@microsoft.com> Cc: Ross Zwisler <zwisler@kernel.org> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Radim Krcmar <rkrcmar@redhat.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Christian Koenig <christian.koenig@amd.com> Cc: Felix Kuehling <felix.kuehling@amd.com> Cc: Ralph Campbell <rcampbell@nvidia.com> Cc: John Hubbard <jhubbard@nvidia.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Michal Hocko authored
We have received a bug report that an injected MCE about faulty memory prevents memory offline to succeed on 4.4 base kernel. The underlying reason was that the HWPoison page has an elevated reference count and the migration keeps failing. There are two problems with that. First of all it is dubious to migrate the poisoned page because we know that accessing that memory is possible to fail. Secondly it doesn't make any sense to migrate a potentially broken content and preserve the memory corruption over to a new location. Oscar has found out that 4.4 and the current upstream kernels behave slightly differently with his simply testcase === int main(void) { int ret; int i; int fd; char *array = malloc(4096); char *array_locked = malloc(4096); fd = open("/tmp/data", O_RDONLY); read(fd, array, 4095); for (i = 0; i < 4096; i++) array_locked[i] = 'd'; ret = mlock((void *)PAGE_ALIGN((unsigned long)array_locked), sizeof(array_locked)); if (ret) perror("mlock"); sleep (20); ret = madvise((void *)PAGE_ALIGN((unsigned long)array_locked), 4096, MADV_HWPOISON); if (ret) perror("madvise"); for (i = 0; i < 4096; i++) array_locked[i] = 'd'; return 0; } === + offline this memory. In 4.4 kernels he saw the hwpoisoned page to be returned back to the LRU list kernel: [<ffffffff81019ac9>] dump_trace+0x59/0x340 kernel: [<ffffffff81019e9a>] show_stack_log_lvl+0xea/0x170 kernel: [<ffffffff8101ac71>] show_stack+0x21/0x40 kernel: [<ffffffff8132bb90>] dump_stack+0x5c/0x7c kernel: [<ffffffff810815a1>] warn_slowpath_common+0x81/0xb0 kernel: [<ffffffff811a275c>] __pagevec_lru_add_fn+0x14c/0x160 kernel: [<ffffffff811a2eed>] pagevec_lru_move_fn+0xad/0x100 kernel: [<ffffffff811a334c>] __lru_cache_add+0x6c/0xb0 kernel: [<ffffffff81195236>] add_to_page_cache_lru+0x46/0x70 kernel: [<ffffffffa02b4373>] extent_readpages+0xc3/0x1a0 [btrfs] kernel: [<ffffffff811a16d7>] __do_page_cache_readahead+0x177/0x200 kernel: [<ffffffff811a18c8>] ondemand_readahead+0x168/0x2a0 kernel: [<ffffffff8119673f>] generic_file_read_iter+0x41f/0x660 kernel: [<ffffffff8120e50d>] __vfs_read+0xcd/0x140 kernel: [<ffffffff8120e9ea>] vfs_read+0x7a/0x120 kernel: [<ffffffff8121404b>] kernel_read+0x3b/0x50 kernel: [<ffffffff81215c80>] do_execveat_common.isra.29+0x490/0x6f0 kernel: [<ffffffff81215f08>] do_execve+0x28/0x30 kernel: [<ffffffff81095ddb>] call_usermodehelper_exec_async+0xfb/0x130 kernel: [<ffffffff8161c045>] ret_from_fork+0x55/0x80 And that latter confuses the hotremove path because an LRU page is attempted to be migrated and that fails due to an elevated reference count. It is quite possible that the reuse of the HWPoisoned page is some kind of fixed race condition but I am not really sure about that. With the upstream kernel the failure is slightly different. The page doesn't seem to have LRU bit set but isolate_movable_page simply fails and do_migrate_range simply puts all the isolated pages back to LRU and therefore no progress is made and scan_movable_pages finds same set of pages over and over again. Fix both cases by explicitly checking HWPoisoned pages before we even try to get reference on the page, try to unmap it if it is still mapped. As explained by Naoya: : Hwpoison code never unmapped those for no big reason because : Ksm pages never dominate memory, so we simply didn't have strong : motivation to save the pages. Also put WARN_ON(PageLRU) in case there is a race and we can hit LRU HWPoison pages which shouldn't happen but I couldn't convince myself about that. Naoya has noted the following: : Theoretically no such gurantee, because try_to_unmap() doesn't have a : guarantee of success and then memory_failure() returns immediately : when hwpoison_user_mappings fails. : Or the following code (comes after hwpoison_user_mappings block) also impli= : es : that the target page can still have PageLRU flag. : : /* : * Torn down by someone else? : */ : if (PageLRU(p) && !PageSwapCache(p) && p->mapping =3D=3D NULL) { : action_result(pfn, MF_MSG_TRUNCATED_LRU, MF_IGNORED); : res =3D -EBUSY; : goto out; : } : : So I think it's OK to keep "if (WARN_ON(PageLRU(page)))" block in : current version of your patch. Link: http://lkml.kernel.org/r/20181206120135.14079-1-mhocko@kernel.orgSigned-off-by:Michal Hocko <mhocko@suse.com> Reviewed-by:
Oscar Salvador <osalvador@suse.com> Debugged-by:
David Hildenbrand <david@redhat.com> Acked-by:
Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: <stable@vger.kernel.org> Signed-off-by:
-
Oscar Salvador authored
kmemleak_scan() goes through all online nodes and tries to scan all used pages. We can do better and use pfn_to_online_page(), so in case we have CONFIG_MEMORY_HOTPLUG, offlined pages will be skiped automatically. For boxes where CONFIG_MEMORY_HOTPLUG is not present, pfn_to_online_page() will fallback to pfn_valid(). Another little optimization is to check if the page belongs to the node we are currently checking, so in case we have nodes interleaved we will not check the same pfn multiple times. I ran some tests: Add some memory to node1 and node2 making it interleaved: (qemu) object_add memory-backend-ram,id=ram0,size=1G (qemu) device_add pc-dimm,id=dimm0,memdev=ram0,node=1 (qemu) object_add memory-backend-ram,id=ram1,size=1G (qemu) device_add pc-dimm,id=dimm1,memdev=ram1,node=2 (qemu) object_add memory-backend-ram,id=ram2,size=1G (qemu) device_add pc-dimm,id=dimm2,memdev=ram2,node=1 Then, we offline that memory: # for i in {32..39} ; do echo "offline" > /sys/devices/system/node/node1/memory$i/state;done # for i in {48..55} ; do echo "offline" > /sys/devices/system/node/node1/memory$i/state;don # for i in {40..47} ; do echo "offline" > /sys/devices/system/node/node2/memory$i/state;done And we run kmemleak_scan: # echo "scan" > /sys/kernel/debug/kmemleak before the patch: kmemleak: time spend: 41596 us after the patch: kmemleak: time spend: 34899 us [akpm@linux-foundation.org: remove stray newline, per Oscar] Link: http://lkml.kernel.org/r/20181206131918.25099-1-osalvador@suse.deSigned-off-by:Oscar Salvador <osalvador@suse.de> Reviewed-by:
Wei Yang <richard.weiyang@gmail.com> Suggested-by:
Catalin Marinas <catalin.marinas@arm.com> Signed-off-by:
-
Will Deacon authored
Whilst no architectures actually enable support for huge p4d mappings in the vmap area, the code that is implemented should be using break-before-make, as we do for pud and pmd huge entries. Link: http://lkml.kernel.org/r/1544120495-17438-6-git-send-email-will.deacon@arm.comSigned-off-by:
Will Deacon <will.deacon@arm.com> Reviewed-by:
Toshi Kani <toshi.kani@hpe.com> Cc: Chintan Pandya <cpandya@codeaurora.org> Cc: Toshi Kani <toshi.kani@hpe.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Michal Hocko <mhocko@suse.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Sean Christopherson <sean.j.christopherson@intel.com> Signed-off-by:
-
Will Deacon authored
The current ioremap() code uses a phys_addr variable at each level of page table, which is confusingly offset by subtracting the base virtual address being mapped so that adding the current virtual address back on when iterating through the page table entries gives back the corresponding physical address. This is fairly confusing and results in all users of phys_addr having to add the current virtual address back on. Instead, this patch just updates phys_addr when iterating over the page table entries, ensuring that it's always up-to-date and doesn't require explicit offsetting. Link: http://lkml.kernel.org/r/1544120495-17438-5-git-send-email-will.deacon@arm.comSigned-off-by:
Sean Christopherson <sean.j.christopherson@intel.com> Reviewed-by:
-
Will Deacon authored
The core code already has a check for pXd_none(), so remove it from the architecture implementation. Link: http://lkml.kernel.org/r/1544120495-17438-4-git-send-email-will.deacon@arm.comSigned-off-by:
Thomas Gleixner <tglx@linutronix.de> Reviewed-by:
-
Will Deacon authored
The core code already has a check for pXd_none(), so remove it from the architecture implementation. Link: http://lkml.kernel.org/r/1544120495-17438-3-git-send-email-will.deacon@arm.comSigned-off-by:
-
Will Deacon authored
The recently merged API for ensuring break-before-make on page-table entries when installing huge mappings in the vmalloc/ioremap region is fairly counter-intuitive, resulting in the arch freeing functions (e.g. pmd_free_pte_page()) being called even on entries that aren't present. This resulted in a minor bug in the arm64 implementation, giving rise to spurious VM_WARN messages. This patch moves the pXd_present() checks out into the core code, refactoring the callsites at the same time so that we avoid the complex conjunctions when determining whether or not we can put down a huge mapping. Link: http://lkml.kernel.org/r/1544120495-17438-2-git-send-email-will.deacon@arm.comSigned-off-by:
-
Kirill Tkhai authored
page always is not NULL, so we may remove this useless check. Link: http://lkml.kernel.org/r/154419752044.18559.2452963074922917720.stgit@localhost.localdomainSigned-off-by:
Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by:
Cyrill Gorcunov <gorcunov@gmail.com> Reviewed-by:
-
Anthony Yznaga authored
Certain pages that are never mapped to userspace have a type indicated in the page_type field of their struct pages (e.g. PG_buddy). page_type overlaps with _mapcount so set the count to 0 and avoid calling page_mapcount() for these pages. [anthony.yznaga@oracle.com: incorporate feedback from Matthew Wilcox] Link: http://lkml.kernel.org/r/1544481313-27318-1-git-send-email-anthony.yznaga@oracle.com Link: http://lkml.kernel.org/r/1543963526-27917-1-git-send-email-anthony.yznaga@oracle.comSigned-off-by:
Anthony Yznaga <anthony.yznaga@oracle.com> Reviewed-by:
Matthew Wilcox <willy@infradead.org> Reviewed-by:
-
Anthony Yznaga authored
Because kpagecount_read() fakes success if map counts are not being collected, clamp the page count passed to it by walk_pfn() to the pages value returned by the preceding call to kpageflags_read(). Link: http://lkml.kernel.org/r/1543962269-26116-1-git-send-email-anthony.yznaga@oracle.com Fixes: 7f1d23e6 ("tools/vm/page-types.c: include shared map counts") Signed-off-by:
-
Oscar Salvador authored
Since commit 03e85f9d ("mm/page_alloc: Introduce free_area_init_core_hotplug"), some functions changed to only be called during system initialization. Concretly, free_area_init_node() and the functions that hang from it. Also, some variables are no longer used after the system has gone through initialization. So this could be considered as a late clean-up for that patch. This patch changes the functions from __meminit to __init, and the variables from __meminitdata to __initdata. In return, we get some KBs back: Before: Freeing unused kernel image memory: 2472K After: Freeing unused kernel image memory: 2480K Link: http://lkml.kernel.org/r/20181204111507.4808-1-osalvador@suse.deSigned-off-by:
-
Brian Foster authored
write_cache_pages() is used in both background and integrity writeback scenarios by various filesystems. Background writeback is mostly concerned with cleaning a certain number of dirty pages based on various mm heuristics. It may not write the full set of dirty pages or wait for I/O to complete. Integrity writeback is responsible for persisting a set of dirty pages before the writeback job completes. For example, an fsync() call must perform integrity writeback to ensure data is on disk before the call returns. write_cache_pages() unconditionally breaks out of its processing loop in the event of a ->writepage() error. This is fine for background writeback, which had no strict requirements and will eventually come around again. This can cause problems for integrity writeback on filesystems that might need to clean up state associated with failed page writeouts. For example, XFS performs internal delayed allocation accounting before returning a ->writepage() error, where applicable. If the current writeback happens to be associated with an unmount and write_cache_pages() completes the writeback prematurely due to error, the filesystem is unmounted in an inconsistent state if dirty+delalloc pages still exist. To handle this problem, update write_cache_pages() to always process the full set of pages for integrity writeback regardless of ->writepage() errors. Save the first encountered error and return it to the caller once complete. This facilitates XFS (or any other fs that expects integrity writeback to process the entire set of dirty pages) to clean up its internal state completely in the event of persistent mapping errors. Background writeback continues to exit on the first error encountered. [akpm@linux-foundation.org: fix typo in comment] Link: http://lkml.kernel.org/r/20181116134304.32440-1-bfoster@redhat.comSigned-off-by:
Brian Foster <bfoster@redhat.com> Reviewed-by:
-
Wei Yang authored
Function show_mem() is used to print system memory status when user requires or fail to allocate memory. Generally, this is a best effort information so any races with memory hotplug (or very theoretically an early initialization) should be tolerable and the worst that could happen is to print an imprecise node state. Drop the resize lock because this is the only place which might hold the lock from the interrupt context and so all other callers might use a simple spinlock. Even though this doesn't solve any real issue it makes the code easier to follow and tiny more effective. Link: http://lkml.kernel.org/r/20181129235532.9328-1-richard.weiyang@gmail.comSigned-off-by:
-
YueHaibing authored
Fixes gcc '-Wunused-but-set-variable' warning: mm/hmm.c: In function 'hmm_devmem_ref_kill': mm/hmm.c:995:21: warning: variable 'devmem' set but not used [-Wunused-but-set-variable] It not used any more since 35d39f953d4e ("mm, hmm: replace hmm_devmem_pages_create() with devm_memremap_pages()") Link: http://lkml.kernel.org/r/1543629971-128377-1-git-send-email-yuehaibing@huawei.comSigned-off-by:YueHaibing <yuehaibing@huawei.com> Reviewed-by:
-
Wei Yang authored
During online_pages phase, pgdat->nr_zones will be updated in case this zone is empty. Currently the online_pages phase is protected by the global locks (device_device_hotplug_lock and mem_hotplug_lock), which ensures there is no contention during the update of nr_zones. These global locks introduces scalability issues (especially the second one), which slow down code relying on get_online_mems(). This is also a preparation for not having to rely on get_online_mems() but instead some more fine grained locks. The patch moves init_currently_empty_zone under both zone_span_writelock and pgdat_resize_lock because both the pgdat state is changed (nr_zones) and the zone's start_pfn. Also this patch changes the documentation of node_size_lock to include the protection of nr_zones. Link: http://lkml.kernel.org/r/20181203205016.14123-1-richard.weiyang@gmail.comSigned-off-by:
-
Wei Yang authored
Since the information needed in sparse_add_one_section() is node id to allocate proper memory, it is not necessary to pass its pgdat. This patch changes the prototype of sparse_add_one_section() to pass node id directly. This is intended to reduce misleading that sparse_add_one_section() would touch pgdat. Link: http://lkml.kernel.org/r/20181204085657.20472-2-richard.weiyang@gmail.comSigned-off-by:
-
Wei Yang authored
pgdat_resize_lock is used to protect pgdat's memory region information like: node_start_pfn, node_present_pages, etc. While in function sparse_add/remove_one_section(), pgdat_resize_lock is used to protect initialization/release of one mem_section. This looks not proper. These code paths are currently protected by mem_hotplug_lock currently but should there ever be any reason for locking at the sparse layer a dedicated lock should be introduced. Following is the current call trace of sparse_add/remove_one_section() mem_hotplug_begin() arch_add_memory() add_pages() __add_pages() __add_section() sparse_add_one_section() mem_hotplug_done() mem_hotplug_begin() arch_remove_memory() __remove_pages() __remove_section() sparse_remove_one_section() mem_hotplug_done() The comment above the pgdat_resize_lock also mentions "Holding this will also guarantee that any pfn_valid() stays that way.", which is true with the current implementation and false after this patch. But current implementation doesn't meet this comment. There isn't any pfn walkers to take the lock so this looks like a relict from the past. This patch also removes this comment. [richard.weiyang@gmail.com: v4] Link: http://lkml.kernel.org/r/20181204085657.20472-1-richard.weiyang@gmail.com [mhocko@suse.com: changelog suggestion] Link: http://lkml.kernel.org/r/20181128091243.19249-1-richard.weiyang@gmail.comSigned-off-by: -
Yu Zhao authored
Pagetable page doesn't touch page->mapping or have any used field that overlaps with it. No need to clear mapping in dtor. In fact, doing so might mask problems that otherwise would be detected by bad_page(). Link: http://lkml.kernel.org/r/20181128235525.58780-1-yuzhao@google.comSigned-off-by:
Yu Zhao <yuzhao@google.com> Reviewed-by:
-
Minchan Kim authored
If there are lots of write IO with flash device, it could have a wearout problem of storage. To overcome the problem, admin needs to design write limitation to guarantee flash health for entire product life. This patch creates a new knob "writeback_limit" for zram. writeback_limit's default value is 0 so that it doesn't limit any writeback. If admin want to measure writeback count in a certain period, he could know it via /sys/block/zram0/bd_stat's 3rd column. If admin want to limit writeback as per-day 400M, he could do it like below. MB_SHIFT=20 4K_SHIFT=12 echo $((400<<MB_SHIFT>>4K_SHIFT)) > \ /sys/block/zram0/writeback_limit. If admin want to allow further write again, he could do it like below echo 0 > /sys/block/zram0/writeback_limit If admin want to see remaining writeback budget, cat /sys/block/zram0/writeback_limit The writeback_limit count will reset whenever you reset zram (e.g., system reboot, echo 1 > /sys/block/zramX/reset) so keeping how many of writeback happened until you reset the zram to allocate extra writeback budget in next setting is user's job. [minchan@kernel.org: v4] Link: http://lkml.kernel.org/r/20181203024045.153534-8-minchan@kernel.org Link: http://lkml.kernel.org/r/20181127055429.251614-8-minchan@kernel.orgSigned-off-by:
Minchan Kim <minchan@kernel.org> Reviewed-by:
Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Joey Pabalinas <joeypabalinas@gmail.com> Signed-off-by:
-
Minchan Kim authored
bd_stat represents things that happened in the backing device. Currently it supports bd_counts, bd_reads and bd_writes which are helpful to understand wearout of flash and memory saving. [minchan@kernel.org: v4] Link: http://lkml.kernel.org/r/20181203024045.153534-7-minchan@kernel.org Link: http://lkml.kernel.org/r/20181127055429.251614-7-minchan@kernel.orgSigned-off-by:
-
Minchan Kim authored
Add a new feature "zram idle/huge page writeback". In the zram-swap use case, zram usually has many idle/huge swap pages. It's pointless to keep them in memory (ie, zram). To solve this problem, this feature introduces idle/huge page writeback to the backing device so the goal is to save more memory space on embedded systems. Normal sequence to use idle/huge page writeback feature is as follows, while (1) { # mark allocated zram slot to idle echo all > /sys/block/zram0/idle # leave system working for several hours # Unless there is no access for some blocks on zram, # they are still IDLE marked pages. echo "idle" > /sys/block/zram0/writeback or/and echo "huge" > /sys/block/zram0/writeback # write the IDLE or/and huge marked slot into backing device # and free the memory. } Per the discussion at https://lore.kernel.org/lkml/20181122065926.GG3441@jagdpanzerIV/T/#u, This patch removes direct incommpressibe page writeback feature (d2afd25114f4 ("zram: write incompressible pages to backing device")). Below concerns from Sergey: == &< == "IDLE writeback" is superior to "incompressible writeback". "incompressible writeback" is completely unpredictable and uncontrollable; it depens on data patterns and compression algorithms. While "IDLE writeback" is predictable. I even suspect, that, *ideally*, we can remove "incompressible writeback". "IDLE pages" is a super set which also includes "incompressible" pages. So, technically, we still can do "incompressible writeback" from "IDLE writeback" path; but a much more reasonable one, based on a page idling period. I understand that you want to keep "direct incompressible writeback" around. ZRAM is especially popular on devices which do suffer from flash wearout, so I can see "incompressible writeback" path becoming a dead code, long term. == &< == Below concerns from Minchan: == &< == My concern is if we enable CONFIG_ZRAM_WRITEBACK in this implementation, both hugepage/idlepage writeck will turn on. However someuser want to enable only idlepage writeback so we need to introduce turn on/off knob for hugepage or new CONFIG_ZRAM_IDLEPAGE_WRITEBACK for those usecase. I don't want to make it complicated *if possible*. Long term, I imagine we need to make VM aware of new swap hierarchy a little bit different with as-is. For example, first high priority swap can return -EIO or -ENOCOMP, swap try to fallback to next lower priority swap device. With that, hugepage writeback will work tranparently. So we could regard it as regression because incompressible pages doesn't go to backing storage automatically. Instead, user should do it via "echo huge" > /sys/block/zram/writeback" manually. == &< == Link: http://lkml.kernel.org/r/20181127055429.251614-6-minchan@kernel.orgSigned-off-by:Joey Pabalinas <joeypabalinas@gmail.com> Reviewed-by:
-
Minchan Kim authored
To support idle page writeback with upcoming patches, this patch introduces a new ZRAM_IDLE flag. Userspace can mark zram slots as "idle" via "echo all > /sys/block/zramX/idle" which marks every allocated zram slot as ZRAM_IDLE. User could see it by /sys/kernel/debug/zram/zram0/block_state. 300 75.033841 ...i 301 63.806904 s..i 302 63.806919 ..hi Once there is IO for the slot, the mark will be disappeared. 300 75.033841 ... 301 63.806904 s..i 302 63.806919 ..hi Therefore, 300th block is idle zpage. With this feature, user can how many zram has idle pages which are waste of memory. Link: http://lkml.kernel.org/r/20181127055429.251614-5-minchan@kernel.orgSigned-off-by: -
Minchan Kim authored
Rename some variables and restructure some code for better readability in writeback and zs_free_page. Link: http://lkml.kernel.org/r/20181127055429.251614-4-minchan@kernel.orgSigned-off-by:
-
Minchan Kim authored
If blkdev_get fails, we shouldn't do blkdev_put. Otherwise, kernel emits below log. This patch fixes it. WARNING: CPU: 0 PID: 1893 at fs/block_dev.c:1828 blkdev_put+0x105/0x120 Modules linked in: CPU: 0 PID: 1893 Comm: swapoff Not tainted 4.19.0+ #453 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 RIP: 0010:blkdev_put+0x105/0x120 Call Trace: __x64_sys_swapoff+0x46d/0x490 do_syscall_64+0x5a/0x190 entry_SYSCALL_64_after_hwframe+0x49/0xbe irq event stamp: 4466 hardirqs last enabled at (4465): __free_pages_ok+0x1e3/0x490 hardirqs last disabled at (4466): trace_hardirqs_off_thunk+0x1a/0x1c softirqs last enabled at (3420): __do_softirq+0x333/0x446 softirqs last disabled at (3407): irq_exit+0xd1/0xe0 Link: http://lkml.kernel.org/r/20181127055429.251614-3-minchan@kernel.orgSigned-off-by: -
Minchan Kim authored
Patch series "zram idle page writeback", v3. Inherently, swap device has many idle pages which are rare touched since it was allocated. It is never problem if we use storage device as swap. However, it's just waste for zram-swap. This patchset supports zram idle page writeback feature. * Admin can define what is idle page "no access since X time ago" * Admin can define when zram should writeback them * Admin can define when zram should stop writeback to prevent wearout Details are in each patch's description. This patch (of 7): ================================ WARNING: inconsistent lock state 4.19.0+ #390 Not tainted -------------------------------- inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage. zram_verify/2095 [HC0[0]:SC1[1]:HE1:SE0] takes: 00000000b1828693 (&(&zram->bitmap_lock)->rlock){+.?.}, at: put_entry_bdev+0x1e/0x50 {SOFTIRQ-ON-W} state was registered at: _raw_spin_lock+0x2c/0x40 zram_make_request+0x755/0xdc9 generic_make_request+0x373/0x6a0 submit_bio+0x6c/0x140 __swap_writepage+0x3a8/0x480 shrink_page_list+0x1102/0x1a60 shrink_inactive_list+0x21b/0x3f0 shrink_node_memcg.constprop.99+0x4f8/0x7e0 shrink_node+0x7d/0x2f0 do_try_to_free_pages+0xe0/0x300 try_to_free_pages+0x116/0x2b0 __alloc_pages_slowpath+0x3f4/0xf80 __alloc_pages_nodemask+0x2a2/0x2f0 __handle_mm_fault+0x42e/0xb50 handle_mm_fault+0x55/0xb0 __do_page_fault+0x235/0x4b0 page_fault+0x1e/0x30 irq event stamp: 228412 hardirqs last enabled at (228412): [<ffffffff98245846>] __slab_free+0x3e6/0x600 hardirqs last disabled at (228411): [<ffffffff98245625>] __slab_free+0x1c5/0x600 softirqs last enabled at (228396): [<ffffffff98e0031e>] __do_softirq+0x31e/0x427 softirqs last disabled at (228403): [<ffffffff98072051>] irq_exit+0xd1/0xe0 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&(&zram->bitmap_lock)->rlock); <Interrupt> lock(&(&zram->bitmap_lock)->rlock); *** DEADLOCK *** no locks held by zram_verify/2095. stack backtrace: CPU: 5 PID: 2095 Comm: zram_verify Not tainted 4.19.0+ #390 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014 Call Trace: <IRQ> dump_stack+0x67/0x9b print_usage_bug+0x1bd/0x1d3 mark_lock+0x4aa/0x540 __lock_acquire+0x51d/0x1300 lock_acquire+0x90/0x180 _raw_spin_lock+0x2c/0x40 put_entry_bdev+0x1e/0x50 zram_free_page+0xf6/0x110 zram_slot_free_notify+0x42/0xa0 end_swap_bio_read+0x5b/0x170 blk_update_request+0x8f/0x340 scsi_end_request+0x2c/0x1e0 scsi_io_completion+0x98/0x650 blk_done_softirq+0x9e/0xd0 __do_softirq+0xcc/0x427 irq_exit+0xd1/0xe0 do_IRQ+0x93/0x120 common_interrupt+0xf/0xf </IRQ> With writeback feature, zram_slot_free_notify could be called in softirq context by end_swap_bio_read. However, bitmap_lock is not aware of that so lockdep yell out: get_entry_bdev spin_lock(bitmap->lock); irq softirq end_swap_bio_read zram_slot_free_notify zram_slot_lock <-- deadlock prone zram_free_page put_entry_bdev spin_lock(bitmap->lock); <-- deadlock prone With akpm's suggestion (i.e. bitmap operation is already atomic), we could remove bitmap lock. It might fail to find a empty slot if serious contention happens. However, it's not severe problem because huge page writeback has already possiblity to fail if there is severe memory pressure. Worst case is just keeping the incompressible in memory, not storage. The other problem is zram_slot_lock in zram_slot_slot_free_notify. To make it safe is this patch introduces zram_slot_trylock where zram_slot_free_notify uses it. Although it's rare to be contented, this patch adds new debug stat "miss_free" to keep monitoring how often it happens. Link: http://lkml.kernel.org/r/20181127055429.251614-2-minchan@kernel.orgSigned-off-by: -
Qian Cai authored
Kmemleak does not play well with KASAN (tested on both HPE Apollo 70 and Huawei TaiShan 2280 aarch64 servers). After calling start_kernel()->setup_arch()->kasan_init(), kmemleak early log buffer went from something like 280 to 260000 which caused kmemleak disabled and crash dump memory reservation failed. The multitude of kmemleak_alloc() calls is from nested loops while KASAN is setting up full memory mappings, so let early kmemleak allocations skip those memblock_alloc_internal() calls came from kasan_init() given that those early KASAN memory mappings should not reference to other memory. Hence, no kmemleak false positives. kasan_init kasan_map_populate [1] kasan_pgd_populate [2] kasan_pud_populate [3] kasan_pmd_populate [4] kasan_pte_populate [5] kasan_alloc_zeroed_page memblock_alloc_try_nid memblock_alloc_internal kmemleak_alloc [1] for_each_memblock(memory, reg) [2] while (pgdp++, addr = next, addr != end) [3] while (pudp++, addr = next, addr != end && pud_none(READ_ONCE(*pudp))) [4] while (pmdp++, addr = next, addr != end && pmd_none(READ_ONCE(*pmdp))) [5] while (ptep++, addr = next, addr != end && pte_none(READ_ONCE(*ptep))) Link: http://lkml.kernel.org/r/1543442925-17794-1-git-send-email-cai@gmx.usSigned-off-by:Qian Cai <cai@gmx.us> Acked-by:
-
Oscar Salvador authored
This is a preparation for the next patch. Currently, we only call release_mem_region_adjustable() in __remove_pages if the zone is not ZONE_DEVICE, because resources that belong to HMM/devm are being released by themselves with devm_release_mem_region. Since we do not want to touch any zone/page stuff during the removing of the memory (but during the offlining), we do not want to check for the zone here. So we need another way to tell release_mem_region_adjustable() to not realease the resource in case it belongs to HMM/devm. HMM/devm acquires/releases a resource through devm_request_mem_region/devm_release_mem_region. These resources have the flag IORESOURCE_MEM, while resources acquired by hot-add memory path (register_memory_resource()) contain IORESOURCE_SYSTEM_RAM. So, we can check for this flag in release_mem_region_adjustable, and if the resource does not contain such flag, we know that we are dealing with a HMM/devm resource, so we can back off. Link: http://lkml.kernel.org/r/20181127162005.15833-3-osalvador@suse.deSigned-off-by:
Pavel Tatashin <pasha.tatashin@soleen.com> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Jerome Glisse <jglisse@redhat.com> Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Oscar Salvador <osalvador@suse.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Signed-off-by:
-
Oscar Salvador authored
Patch series "Do not touch pages in hot-remove path", v2. This patchset aims for two things: 1) A better definition about offline and hot-remove stage 2) Solving bugs where we can access non-initialized pages during hot-remove operations [2] [3]. This is achieved by moving all page/zone handling to the offline stage, so we do not need to access pages when hot-removing memory. [1] https://patchwork.kernel.org/cover/10691415/ [2] https://patchwork.kernel.org/patch/10547445/ [3] https://www.spinics.net/lists/linux-mm/msg161316.html This patch (of 5): This is a preparation for the following-up patches. The idea of passing the nid is that it will allow us to get rid of the zone parameter afterwards. Link: http://lkml.kernel.org/r/20181127162005.15833-2-osalvador@suse.deSigned-off-by: -
Wei Yang authored
When DEFERRED_STRUCT_PAGE_INIT is configured, only the first section of each node's highest zone is initialized before defer stage. static_init_pgcnt is used to store the number of pages like this: pgdat->static_init_pgcnt = min_t(unsigned long, PAGES_PER_SECTION, pgdat->node_spanned_pages); because we don't want to overflow zone's range. But this is not necessary, since defer_init() is called like this: memmap_init_zone() for pfn in [start_pfn, end_pfn) defer_init(pfn, end_pfn) In case (pgdat->node_spanned_pages < PAGES_PER_SECTION), the loop would stop before calling defer_init(). BTW, comparing PAGES_PER_SECTION with node_spanned_pages is not correct, since nr_initialised is zone based instead of node based. Even node_spanned_pages is bigger than PAGES_PER_SECTION, its highest zone would have pages less than PAGES_PER_SECTION. Link: http://lkml.kernel.org/r/20181122094807.6985-1-richard.weiyang@gmail.comSigned-off-by:Alexander Duyck <alexander.h.duyck@linux.intel.com> Cc: Pavel Tatashin <pasha.tatashin@oracle.com> Cc: Oscar Salvador <osalvador@suse.de> Signed-off-by:
-
Hugh Dickins authored
Waiting on a page migration entry has used wait_on_page_locked() all along since 2006: but you cannot safely wait_on_page_locked() without holding a reference to the page, and that extra reference is enough to make migrate_page_move_mapping() fail with -EAGAIN, when a racing task faults on the entry before migrate_page_move_mapping() gets there. And that failure is retried nine times, amplifying the pain when trying to migrate a popular page. With a single persistent faulter, migration sometimes succeeds; with two or three concurrent faulters, success becomes much less likely (and the more the page was mapped, the worse the overhead of unmapping and remapping it on each try). This is especially a problem for memory offlining, where the outer level retries forever (or until terminated from userspace), because a heavy refault workload can trigger an endless loop of migration failures. wait_on_page_locked() is the wrong tool for the job. David Herrmann (but was he the first?) noticed this issue in 2014: https://marc.info/?l=linux-mm&m=140110465608116&w=2 Tim Chen started a thread in August 2017 which appears relevant: https://marc.info/?l=linux-mm&m=150275941014915&w=2 where Kan Liang went on to implicate __migration_entry_wait(): https://marc.info/?l=linux-mm&m=150300268411980&w=2 and the thread ended up with the v4.14 commits: 2554db91 ("sched/wait: Break up long wake list walk") 11a19c7b ("sched/wait: Introduce wakeup boomark in wake_up_page_bit") Baoquan He reported "Memory hotplug softlock issue" 14 November 2018: https://marc.info/?l=linux-mm&m=154217936431300&w=2 We have all assumed that it is essential to hold a page reference while waiting on a page lock: partly to guarantee that there is still a struct page when MEMORY_HOTREMOVE is configured, but also to protect against reuse of the struct page going to someone who then holds the page locked indefinitely, when the waiter can reasonably expect timely unlocking. But in fact, so long as wait_on_page_bit_common() does the put_page(), and is careful not to rely on struct page contents thereafter, there is no need to hold a reference to the page while waiting on it. That does mean that this case cannot go back through the loop: but that's fine for the page migration case, and even if used more widely, is limited by the "Stop walking if it's locked" optimization in wake_page_function(). Add interface put_and_wait_on_page_locked() to do this, using "behavior" enum in place of "lock" arg to wait_on_page_bit_common() to implement it. No interruptible or killable variant needed yet, but they might follow: I have a vague notion that reporting -EINTR should take precedence over return from wait_on_page_bit_common() without knowing the page state, so arrange it accordingly - but that may be nothing but pedantic. __migration_entry_wait() still has to take a brief reference to the page, prior to calling put_and_wait_on_page_locked(): but now that it is dropped before waiting, the chance of impeding page migration is very much reduced. Should we perhaps disable preemption across this? shrink_page_list()'s __ClearPageLocked(): that was a surprise! This survived a lot of testing before that showed up. PageWaiters may have been set by wait_on_page_bit_common(), and the reference dropped, just before shrink_page_list() succeeds in freezing its last page reference: in such a case, unlock_page() must be used. Follow the suggestion from Michal Hocko, just revert a978d6f5 ("mm: unlockless reclaim") now: that optimization predates PageWaiters, and won't buy much these days; but we can reinstate it for the !PageWaiters case if anyone notices. It does raise the question: should vmscan.c's is_page_cache_freeable() and __remove_mapping() now treat a PageWaiters page as if an extra reference were held? Perhaps, but I don't think it matters much, since shrink_page_list() already had to win its trylock_page(), so waiters are not very common there: I noticed no difference when trying the bigger change, and it's surely not needed while put_and_wait_on_page_locked() is only used for page migration. [willy@infradead.org: add put_and_wait_on_page_locked() kerneldoc] Link: http://lkml.kernel.org/r/alpine.LSU.2.11.1811261121330.1116@eggly.anvilsSigned-off-by:
Hugh Dickins <hughd@google.com> Reported-by:
Baoquan He <bhe@redhat.com> Tested-by:
Andrea Arcangeli <aarcange@redhat.com> Acked-by:
Vlastimil Babka <vbabka@suse.cz> Cc: Matthew Wilcox <willy@infradead.org> Cc: Baoquan He <bhe@redhat.com> Cc: David Hildenbrand <david@redhat.com> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: David Herrmann <dh.herrmann@gmail.com> Cc: Tim Chen <tim.c.chen@linux.intel.com> Cc: Kan Liang <kan.liang@intel.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Davidlohr Bueso <dave@stgolabs.net> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Christoph Lameter <cl@linux.com> Cc: Nick Piggin <npiggin@gmail.com> Signed-off-by:
-
yuzhoujian authored
The current oom report doesn't display victim's memcg context during the global OOM situation. While this information is not strictly needed, it can be really helpful for containerized environments to locate which container has lost a process. Now that we have a single line for the oom context, we can trivially add both the oom memcg (this can be either global_oom or a specific memcg which hits its hard limits) and task_memcg which is the victim's memcg. Below is the single line output in the oom report after this patch. - global oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,global_oom,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> - memcg oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,oom_memcg=<memcg>,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> [penguin-kernel@I-love.SAKURA.ne.jp: use pr_cont() in mem_cgroup_print_oom_context()] Link: http://lkml.kernel.org/r/201812190723.wBJ7NdkN032628@www262.sakura.ne.jp Link: http://lkml.kernel.org/r/1542799799-36184-2-git-send-email-ufo19890607@gmail.comSigned-off-by:
yuzhoujian <yuzhoujian@didichuxing.com> Signed-off-by:
Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by:
-
yuzhoujian authored
OOM report contains several sections. The first one is the allocation context that has triggered the OOM. Then we have cpuset context followed by the stack trace of the OOM path. The tird one is the OOM memory information. Followed by the current memory state of all system tasks. At last, we will show oom eligible tasks and the information about the chosen oom victim. One thing that makes parsing more awkward than necessary is that we do not have a single and easily parsable line about the oom context. This patch is reorganizing the oom report to 1) who invoked oom and what was the allocation request [ 515.902945] tuned invoked oom-killer: gfp_mask=0x6200ca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 2) OOM stack trace [ 515.904273] CPU: 24 PID: 1809 Comm: tuned Not tainted 4.20.0-rc3+ #3 [ 515.905518] Hardware name: Inspur SA5212M4/YZMB-00370-107, BIOS 4.1.10 11/14/2016 [ 515.906821] Call Trace: [ 515.908062] dump_stack+0x5a/0x73 [ 515.909311] dump_header+0x55/0x28c [ 515.914260] oom_kill_process+0x2d8/0x300 [ 515.916708] out_of_memory+0x145/0x4a0 [ 515.917932] __alloc_pages_slowpath+0x7d2/0xa16 [ 515.919157] __alloc_pages_nodemask+0x277/0x290 [ 515.920367] filemap_fault+0x3d0/0x6c0 [ 515.921529] ? filemap_map_pages+0x2b8/0x420 [ 515.922709] ext4_filemap_fault+0x2c/0x40 [ext4] [ 515.923884] __do_fault+0x20/0x80 [ 515.925032] __handle_mm_fault+0xbc0/0xe80 [ 515.926195] handle_mm_fault+0xfa/0x210 [ 515.927357] __do_page_fault+0x233/0x4c0 [ 515.928506] do_page_fault+0x32/0x140 [ 515.929646] ? page_fault+0x8/0x30 [ 515.930770] page_fault+0x1e/0x30 3) OOM memory information [ 515.958093] Mem-Info: [ 515.959647] active_anon:26501758 inactive_anon:1179809 isolated_anon:0 active_file:4402672 inactive_file:483963 isolated_file:1344 unevictable:0 dirty:4886753 writeback:0 unstable:0 slab_reclaimable:148442 slab_unreclaimable:18741 mapped:1347 shmem:1347 pagetables:58669 bounce:0 free:88663 free_pcp:0 free_cma:0 ... 4) current memory state of all system tasks [ 516.079544] [ 744] 0 744 9211 1345 114688 82 0 systemd-journal [ 516.082034] [ 787] 0 787 31764 0 143360 92 0 lvmetad [ 516.084465] [ 792] 0 792 10930 1 110592 208 -1000 systemd-udevd [ 516.086865] [ 1199] 0 1199 13866 0 131072 112 -1000 auditd [ 516.089190] [ 1222] 0 1222 31990 1 110592 157 0 smartd [ 516.091477] [ 1225] 0 1225 4864 85 81920 43 0 irqbalance [ 516.093712] [ 1226] 0 1226 52612 0 258048 426 0 abrtd [ 516.112128] [ 1280] 0 1280 109774 55 299008 400 0 NetworkManager [ 516.113998] [ 1295] 0 1295 28817 37 69632 24 0 ksmtuned [ 516.144596] [ 10718] 0 10718 2622484 1721372 15998976 267219 0 panic [ 516.145792] [ 10719] 0 10719 2622484 1164767 9818112 53576 0 panic [ 516.146977] [ 10720] 0 10720 2622484 1174361 9904128 53709 0 panic [ 516.148163] [ 10721] 0 10721 2622484 1209070 10194944 54824 0 panic [ 516.149329] [ 10722] 0 10722 2622484 1745799 14774272 91138 0 panic 5) oom context (contrains and the chosen victim). oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0-1,task=panic,pid=10737,uid=0 An admin can easily get the full oom context at a single line which makes parsing much easier. Link: http://lkml.kernel.org/r/1542799799-36184-1-git-send-email-ufo19890607@gmail.comSigned-off-by:
-
Alexey Dobriyan authored
and propagate through down the call stack. Link: http://lkml.kernel.org/r/20181124091411.GC10969@avx2Signed-off-by:
Alexey Dobriyan <adobriyan@gmail.com> Reviewed-by:
-
Alexey Dobriyan authored
Those strings are immutable as well. Link: http://lkml.kernel.org/r/20181124090508.GB10877@avx2Signed-off-by:
-
Alexey Dobriyan authored
Those strings are immutable in fact. Link: http://lkml.kernel.org/r/20181124090327.GA10877@avx2Signed-off-by:
-
Mel Gorman authored
An external fragmentation event was previously described as When the page allocator fragments memory, it records the event using the mm_page_alloc_extfrag event. If the fallback_order is smaller than a pageblock order (order-9 on 64-bit x86) then it's considered an event that will cause external fragmentation issues in the future. The kernel reduces the probability of such events by increasing the watermark sizes by calling set_recommended_min_free_kbytes early in the lifetime of the system. This works reasonably well in general but if there are enough sparsely populated pageblocks then the problem can still occur as enough memory is free overall and kswapd stays asleep. This patch introduces a watermark_boost_factor sysctl that allows a zone watermark to be temporarily boosted when an external fragmentation causing events occurs. The boosting will stall allocations that would decrease free memory below the boosted low watermark and kswapd is woken if the calling context allows to reclaim an amount of memory relative to the size of the high watermark and the watermark_boost_factor until the boost is cleared. When kswapd finishes, it wakes kcompactd at the pageblock order to clean some of the pageblocks that may have been affected by the fragmentation event. kswapd avoids any writeback, slab shrinkage and swap from reclaim context during this operation to avoid excessive system disruption in the name of fragmentation avoidance. Care is taken so that kswapd will do normal reclaim work if the system is really low on memory. This was evaluated using the same workloads as "mm, page_alloc: Spread allocations across zones before introducing fragmentation". 1-socket Skylake machine config-global-dhp__workload_thpfioscale XFS (no special madvise) 4 fio threads, 1 THP allocating thread -------------------------------------- 4.20-rc3 extfrag events < order 9: 804694 4.20-rc3+patch: 408912 (49% reduction) 4.20-rc3+patch1-4: 18421 (98% reduction) 4.20.0-rc3 4.20.0-rc3 lowzone-v5r8 boost-v5r8 Amean fault-base-1 653.58 ( 0.00%) 652.71 ( 0.13%) Amean fault-huge-1 0.00 ( 0.00%) 178.93 * -99.00%* 4.20.0-rc3 4.20.0-rc3 lowzone-v5r8 boost-v5r8 Percentage huge-1 0.00 ( 0.00%) 5.12 ( 100.00%) Note that external fragmentation causing events are massively reduced by this path whether in comparison to the previous kernel or the vanilla kernel. The fault latency for huge pages appears to be increased but that is only because THP allocations were successful with the patch applied. 1-socket Skylake machine global-dhp__workload_thpfioscale-madvhugepage-xfs (MADV_HUGEPAGE) ----------------------------------------------------------------- 4.20-rc3 extfrag events < order 9: 291392 4.20-rc3+patch: 191187 (34% reduction) 4.20-rc3+patch1-4: 13464 (95% reduction) thpfioscale Fault Latencies 4.20.0-rc3 4.20.0-rc3 lowzone-v5r8 boost-v5r8 Min fault-base-1 912.00 ( 0.00%) 905.00 ( 0.77%) Min fault-huge-1 127.00 ( 0.00%) 135.00 ( -6.30%) Amean fault-base-1 1467.55 ( 0.00%) 1481.67 ( -0.96%) Amean fault-huge-1 1127.11 ( 0.00%) 1063.88 * 5.61%* 4.20.0-rc3 4.20.0-rc3 lowzone-v5r8 boost-v5r8 Percentage huge-1 77.64 ( 0.00%) 83.46 ( 7.49%) As before, massive reduction in external fragmentation events, some jitter on latencies and an increase in THP allocation success rates. 2-socket Haswell machine config-global-dhp__workload_thpfioscale XFS (no special madvise) 4 fio threads, 5 THP allocating threads ---------------------------------------------------------------- 4.20-rc3 extfrag events < order 9: 215698 4.20-rc3+patch: 200210 (7% reduction) 4.20-rc3+patch1-4: 14263 (93% reduction) 4.20.0-rc3 4.20.0-rc3 lowzone-v5r8 boost-v5r8 Amean fault-base-5 1346.45 ( 0.00%) 1306.87 ( 2.94%) Amean fault-huge-5 3418.60 ( 0.00%) 1348.94 ( 60.54%) 4.20.0-rc3 4.20.0-rc3 lowzone-v5r8 boost-v5r8 Percentage huge-5 0.78 ( 0.00%) 7.91 ( 910.64%) There is a 93% reduction in fragmentation causing events, there is a big reduction in the huge page fault latency and allocation success rate is higher. 2-socket Haswell machine global-dhp__workload_thpfioscale-madvhugepage-xfs (MADV_HUGEPAGE) ----------------------------------------------------------------- 4.20-rc3 extfrag events < order 9: 166352 4.20-rc3+patch: 147463 (11% reduction) 4.20-rc3+patch1-4: 11095 (93% reduction) thpfioscale Fault Latencies 4.20.0-rc3 4.20.0-rc3 lowzone-v5r8 boost-v5r8 Amean fault-base-5 6217.43 ( 0.00%) 7419.67 * -19.34%* Amean fault-huge-5 3163.33 ( 0.00%) 3263.80 ( -3.18%) 4.20.0-rc3 4.20.0-rc3 lowzone-v5r8 boost-v5r8 Percentage huge-5 95.14 ( 0.00%) 87.98 ( -7.53%) There is a large reduction in fragmentation events with some jitter around the latencies and success rates. As before, the high THP allocation success rate does mean the system is under a lot of pressure. However, as the fragmentation events are reduced, it would be expected that the long-term allocation success rate would be higher. Link: http://lkml.kernel.org/r/20181123114528.28802-5-mgorman@techsingularity.netSigned-off-by:Mel Gorman <mgorman@techsingularity.net> Acked-by:
-
Mel Gorman authored
This is a preparation patch that copies the GFP flag __GFP_KSWAPD_RECLAIM into alloc_flags. This is a preparation patch only that avoids having to pass gfp_mask through a long callchain in a future patch. Note that the setting in the fast path happens in alloc_flags_nofragment() and it may be claimed that this has nothing to do with ALLOC_NO_FRAGMENT. That's true in this patch but is not true later so it's done now for easier review to show where the flag needs to be recorded. No functional change. [mgorman@techsingularity.net: ALLOC_KSWAPD flag needs to be applied in the !CONFIG_ZONE_DMA32 case] Link: http://lkml.kernel.org/r/20181126143503.GO23260@techsingularity.net Link: http://lkml.kernel.org/r/20181123114528.28802-4-mgorman@techsingularity.netSigned-off-by:
-
Mel Gorman authored
This is a preparation patch only, no functional change. Link: http://lkml.kernel.org/r/20181123114528.28802-3-mgorman@techsingularity.netSigned-off-by:
-
Mel Gorman authored
Patch series "Fragmentation avoidance improvements", v5. It has been noted before that fragmentation avoidance (aka anti-fragmentation) is not perfect. Given sufficient time or an adverse workload, memory gets fragmented and the long-term success of high-order allocations degrades. This series defines an adverse workload, a definition of external fragmentation events (including serious) ones and a series that reduces the level of those fragmentation events. The details of the workload and the consequences are described in more detail in the changelogs. However, from patch 1, this is a high-level summary of the adverse workload. The exact details are found in the mmtests implementation. The broad details of the workload are as follows; 1. Create an XFS filesystem (not specified in the configuration but done as part of the testing for this patch) 2. Start 4 fio threads that write a number of 64K files inefficiently. Inefficiently means that files are created on first access and not created in advance (fio parameterr create_on_open=1) and fallocate is not used (fallocate=none). With multiple IO issuers this creates a mix of slab and page cache allocations over time. The total size of the files is 150% physical memory so that the slabs and page cache pages get mixed 3. Warm up a number of fio read-only threads accessing the same files created in step 2. This part runs for the same length of time it took to create the files. It'll fault back in old data and further interleave slab and page cache allocations. As it's now low on memory due to step 2, fragmentation occurs as pageblocks get stolen. 4. While step 3 is still running, start a process that tries to allocate 75% of memory as huge pages with a number of threads. The number of threads is based on a (NR_CPUS_SOCKET - NR_FIO_THREADS)/4 to avoid THP threads contending with fio, any other threads or forcing cross-NUMA scheduling. Note that the test has not been used on a machine with less than 8 cores. The benchmark records whether huge pages were allocated and what the fault latency was in microseconds 5. Measure the number of events potentially causing external fragmentation, the fault latency and the huge page allocation success rate. 6. Cleanup Overall the series reduces external fragmentation causing events by over 94% on 1 and 2 socket machines, which in turn impacts high-order allocation success rates over the long term. There are differences in latencies and high-order allocation success rates. Latencies are a mixed bag as they are vulnerable to exact system state and whether allocations succeeded so they are treated as a secondary metric. Patch 1 uses lower zones if they are populated and have free memory instead of fragmenting a higher zone. It's special cased to handle a Normal->DMA32 fallback with the reasons explained in the changelog. Patch 2-4 boosts watermarks temporarily when an external fragmentation event occurs. kswapd wakes to reclaim a small amount of old memory and then wakes kcompactd on completion to recover the system slightly. This introduces some overhead in the slowpath. The level of boosting can be tuned or disabled depending on the tolerance for fragmentation vs allocation latency. Patch 5 stalls some movable allocation requests to let kswapd from patch 4 make some progress. The duration of the stalls is very low but it is possible to tune the system to avoid fragmentation events if larger stalls can be tolerated. The bulk of the improvement in fragmentation avoidance is from patches 1-4 but patch 5 can deal with a rare corner case and provides the option of tuning a system for THP allocation success rates in exchange for some stalls to control fragmentation. This patch (of 5): The page allocator zone lists are iterated based on the watermarks of each zone which does not take anti-fragmentation into account. On x86, node 0 may have multiple zones while other nodes have one zone. A consequence is that tasks running on node 0 may fragment ZONE_NORMAL even though ZONE_DMA32 has plenty of free memory. This patch special cases the allocator fast path such that it'll try an allocation from a lower local zone before fragmenting a higher zone. In this case, stealing of pageblocks or orders larger than a pageblock are still allowed in the fast path as they are uninteresting from a fragmentation point of view. This was evaluated using a benchmark designed to fragment memory before attempting THP allocations. It's implemented in mmtests as the following configurations configs/config-global-dhp__workload_thpfioscale configs/config-global-dhp__workload_thpfioscale-defrag configs/config-global-dhp__workload_thpfioscale-madvhugepage e.g. from mmtests ./run-mmtests.sh --run-monitor --config configs/config-global-dhp__workload_thpfioscale test-run-1 The broad details of the workload are as follows; 1. Create an XFS filesystem (not specified in the configuration but done as part of the testing for this patch). 2. Start 4 fio threads that write a number of 64K files inefficiently. Inefficiently means that files are created on first access and not created in advance (fio parameter create_on_open=1) and fallocate is not used (fallocate=none). With multiple IO issuers this creates a mix of slab and page cache allocations over time. The total size of the files is 150% physical memory so that the slabs and page cache pages get mixed. 3. Warm up a number of fio read-only processes accessing the same files created in step 2. This part runs for the same length of time it took to create the files. It'll refault old data and further interleave slab and page cache allocations. As it's now low on memory due to step 2, fragmentation occurs as pageblocks get stolen. 4. While step 3 is still running, start a process that tries to allocate 75% of memory as huge pages with a number of threads. The number of threads is based on a (NR_CPUS_SOCKET - NR_FIO_THREADS)/4 to avoid THP threads contending with fio, any other threads or forcing cross-NUMA scheduling. Note that the test has not been used on a machine with less than 8 cores. The benchmark records whether huge pages were allocated and what the fault latency was in microseconds. 5. Measure the number of events potentially causing external fragmentation, the fault latency and the huge page allocation success rate. 6. Cleanup the test files. Note that due to the use of IO and page cache that this benchmark is not suitable for running on large machines where the time to fragment memory may be excessive. Also note that while this is one mix that generates fragmentation that it's not the only mix that generates fragmentation. Differences in workload that are more slab-intensive or whether SLUB is used with high-order pages may yield different results. When the page allocator fragments memory, it records the event using the mm_page_alloc_extfrag ftrace event. If the fallback_order is smaller than a pageblock order (order-9 on 64-bit x86) then it's considered to be an "external fragmentation event" that may cause issues in the future. Hence, the primary metric here is the number of external fragmentation events that occur with order < 9. The secondary metric is allocation latency and huge page allocation success rates but note that differences in latencies and what the success rate also can affect the number of external fragmentation event which is why it's a secondary metric. 1-socket Skylake machine config-global-dhp__workload_thpfioscale XFS (no special madvise) 4 fio threads, 1 THP allocating thread -------------------------------------- 4.20-rc3 extfrag events < order 9: 804694 4.20-rc3+patch: 408912 (49% reduction) thpfioscale Fault Latencies 4.20.0-rc3 4.20.0-rc3 vanilla lowzone-v5r8 Amean fault-base-1 662.92 ( 0.00%) 653.58 * 1.41%* Amean fault-huge-1 0.00 ( 0.00%) 0.00 ( 0.00%) 4.20.0-rc3 4.20.0-rc3 vanilla lowzone-v5r8 Percentage huge-1 0.00 ( 0.00%) 0.00 ( 0.00%) Fault latencies are slightly reduced while allocation success rates remain at zero as this configuration does not make any special effort to allocate THP and fio is heavily active at the time and either filling memory or keeping pages resident. However, a 49% reduction of serious fragmentation events reduces the changes of external fragmentation being a problem in the future. Vlastimil asked during review for a breakdown of the allocation types that are falling back. vanilla 3816 MIGRATE_UNMOVABLE 800845 MIGRATE_MOVABLE 33 MIGRATE_UNRECLAIMABLE patch 735 MIGRATE_UNMOVABLE 408135 MIGRATE_MOVABLE 42 MIGRATE_UNRECLAIMABLE The majority of the fallbacks are due to movable allocations and this is consistent for the workload throughout the series so will not be presented again as the primary source of fallbacks are movable allocations. Movable fallbacks are sometimes considered "ok" to fallback because they can be migrated. The problem is that they can fill an unmovable/reclaimable pageblock causing those allocations to fallback later and polluting pageblocks with pages that cannot move. If there is a movable fallback, it is pretty much guaranteed to affect an unmovable/reclaimable pageblock and while it might not be enough to actually cause a unmovable/reclaimable fallback in the future, we cannot know that in advance so the patch takes the only option available to it. Hence, it's important to control them. This point is also consistent throughout the series and will not be repeated. 1-socket Skylake machine global-dhp__workload_thpfioscale-madvhugepage-xfs (MADV_HUGEPAGE) ----------------------------------------------------------------- 4.20-rc3 extfrag events < order 9: 291392 4.20-rc3+patch: 191187 (34% reduction) thpfioscale Fault Latencies 4.20.0-rc3 4.20.0-rc3 vanilla lowzone-v5r8 Amean fault-base-1 1495.14 ( 0.00%) 1467.55 ( 1.85%) Amean fault-huge-1 1098.48 ( 0.00%) 1127.11 ( -2.61%) thpfioscale Percentage Faults Huge 4.20.0-rc3 4.20.0-rc3 vanilla lowzone-v5r8 Percentage huge-1 78.57 ( 0.00%) 77.64 ( -1.18%) Fragmentation events were reduced quite a bit although this is known to be a little variable. The latencies and allocation success rates are similar but they were already quite high. 2-socket Haswell machine config-global-dhp__workload_thpfioscale XFS (no special madvise) 4 fio threads, 5 THP allocating threads ---------------------------------------------------------------- 4.20-rc3 extfrag events < order 9: 215698 4.20-rc3+patch: 200210 (7% reduction) thpfioscale Fault Latencies 4.20.0-rc3 4.20.0-rc3 vanilla lowzone-v5r8 Amean fault-base-5 1350.05 ( 0.00%) 1346.45 ( 0.27%) Amean fault-huge-5 4181.01 ( 0.00%) 3418.60 ( 18.24%) 4.20.0-rc3 4.20.0-rc3 vanilla lowzone-v5r8 Percentage huge-5 1.15 ( 0.00%) 0.78 ( -31.88%) The reduction of external fragmentation events is slight and this is partially due to the removal of __GFP_THISNODE in commit ac5b2c18 ("mm: thp: relax __GFP_THISNODE for MADV_HUGEPAGE mappings") as THP allocations can now spill over to remote nodes instead of fragmenting local memory. 2-socket Haswell machine global-dhp__workload_thpfioscale-madvhugepage-xfs (MADV_HUGEPAGE) ----------------------------------------------------------------- 4.20-rc3 extfrag events < order 9: 166352 4.20-rc3+patch: 147463 (11% reduction) thpfioscale Fault Latencies 4.20.0-rc3 4.20.0-rc3 vanilla lowzone-v5r8 Amean fault-base-5 6138.97 ( 0.00%) 6217.43 ( -1.28%) Amean fault-huge-5 2294.28 ( 0.00%) 3163.33 * -37.88%* thpfioscale Percentage Faults Huge 4.20.0-rc3 4.20.0-rc3 vanilla lowzone-v5r8 Percentage huge-5 96.82 ( 0.00%) 95.14 ( -1.74%) There was a slight reduction in external fragmentation events although the latencies were higher. The allocation success rate is high enough that the system is struggling and there is quite a lot of parallel reclaim and compaction activity. There is also a certain degree of luck on whether processes start on node 0 or not for this patch but the relevance is reduced later in the series. Overall, the patch reduces the number of external fragmentation causing events so the success of THP over long periods of time would be improved for this adverse workload. Link: http://lkml.kernel.org/r/20181123114528.28802-2-mgorman@techsingularity.netSigned-off-by:
-
