- 16 Jan, 2016 40 commits
-
-
Dan Williams authored
For the purpose of communicating the optional presence of a 'struct page' for the pfn returned from ->direct_access(), introduce a type that encapsulates a page-frame-number plus flags. These flags contain the historical "page_link" encoding for a scatterlist entry, but can also denote "device memory". Where "device memory" is a set of pfns that are not part of the kernel's linear mapping by default, but are accessed via the same memory controller as ram. The motivation for this new type is large capacity persistent memory that needs struct page entries in the 'memmap' to support 3rd party DMA (i.e. O_DIRECT I/O with a persistent memory source/target). However, we also need it in support of maintaining a list of mapped inodes which need to be unmapped at driver teardown or freeze_bdev() time. Signed-off-by:
Dan Williams <dan.j.williams@intel.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Dave Hansen <dave@sr71.net> Cc: Ross Zwisler <ross.zwisler@linux.intel.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Dan Williams authored
To date, we have implemented two I/O usage models for persistent memory, PMEM (a persistent "ram disk") and DAX (mmap persistent memory into userspace). This series adds a third, DAX-GUP, that allows DAX mappings to be the target of direct-i/o. It allows userspace to coordinate DMA/RDMA from/to persistent memory. The implementation leverages the ZONE_DEVICE mm-zone that went into 4.3-rc1 (also discussed at kernel summit) to flag pages that are owned and dynamically mapped by a device driver. The pmem driver, after mapping a persistent memory range into the system memmap via devm_memremap_pages(), arranges for DAX to distinguish pfn-only versus page-backed pmem-pfns via flags in the new pfn_t type. The DAX code, upon seeing a PFN_DEV+PFN_MAP flagged pfn, flags the resulting pte(s) inserted into the process page tables with a new _PAGE_DEVMAP flag. Later, when get_user_pages() is walking ptes it keys off _PAGE_DEVMAP to pin the device hosting the page range active. Finally, get_page() and put_page() are modified to take references against the device driver established page mapping. Finally, this need for "struct page" for persistent memory requires memory capacity to store the memmap array. Given the memmap array for a large pool of persistent may exhaust available DRAM introduce a mechanism to allocate the memmap from persistent memory. The new "struct vmem_altmap *" parameter to devm_memremap_pages() enables arch_add_memory() to use reserved pmem capacity rather than the page allocator. This patch (of 18): The core has developed a need for a "pfn_t" type [1]. Move the existing pfn_t in KVM to kvm_pfn_t [2]. [1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.html [2]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002218.htmlSigned-off-by:
Christoffer Dall <christoffer.dall@linaro.org> Cc: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by:
-
Dan Williams authored
The core has developed a need for a "pfn_t" type [1]. Convert the usage of pfn_t by usermode-linux to an unsigned long, and update pfn_to_phys() to drop its expectation of a typed pfn. [1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.htmlSigned-off-by:
-
Toshi Kani authored
An infinite loop of PMD faults was observed when attempted to mlock() a private read-only PMD mmap'd range of a DAX file. __dax_pmd_fault() simply returns with VM_FAULT_FALLBACK when falling back to PTE on COW. However, __handle_mm_fault() returns without falling back to handle_pte_fault() because a PMD map is present in this case. Change __dax_pmd_fault() to split the PMD map, if present, before returning with VM_FAULT_FALLBACK. Signed-off-by:
Toshi Kani <toshi.kani@hpe.com> Signed-off-by:
-
Ross Zwisler authored
Prior to this change DAX PMD mappings that were made read-only were never able to be made writable again. This is because the code in insert_pfn_pmd() that calls pmd_mkdirty() and pmd_mkwrite() would skip these calls if the PMD already existed in the page table. Instead, if we are doing a write always mark the PMD entry as dirty and writeable. Without this code we can get into a condition where we mark the PMD as read-only, and then on a subsequent write fault we get into an infinite loop of PMD faults where we try unsuccessfully to make the PMD writeable. Signed-off-by:
Ross Zwisler <ross.zwisler@linux.intel.com> Signed-off-by:
Jeff Moyer <jmoyer@redhat.com> Reported-by:
-
Dan Williams authored
The DAX implementation needs to protect new calls to ->direct_access() and usage of its return value against the driver for the underlying block device being disabled. Use blk_queue_enter()/blk_queue_exit() to hold off blk_cleanup_queue() from proceeding, or otherwise fail new mapping requests if the request_queue is being torn down. This also introduces blk_dax_ctl to simplify the interface from fs/dax.c through dax_map_atomic() to bdev_direct_access(). [willy@linux.intel.com: fix read() of a hole] Signed-off-by:
-
Dan Williams authored
If a ->direct_access() implementation ever returns a map count less than PAGE_SIZE, catch the error in bdev_direct_access(). This simplifies error checking in upper layers. Signed-off-by:
-
Dan Williams authored
dax_clear_blocks is currently performing a cond_resched() after every PAGE_SIZE memset. We need not check so frequently, for example md-raid only calls cond_resched() at stripe granularity. Also, in preparation for introducing a dax_map_atomic() operation that temporarily pins a dax mapping move the call to cond_resched() to the outer loop. The worst case latency between calls to cond_resched() after this change is 500us the average latency is 133us. This is up from a 10us max and 4us average. Signed-off-by:
Jan Kara <jack@suse.com> Reviewed-by:
-
Dan Williams authored
To date, we have implemented two I/O usage models for persistent memory, PMEM (a persistent "ram disk") and DAX (mmap persistent memory into userspace). This series adds a third, DAX-GUP, that allows DAX mappings to be the target of direct-i/o. It allows userspace to coordinate DMA/RDMA from/to persistent memory. The implementation leverages the ZONE_DEVICE mm-zone that went into 4.3-rc1 (also discussed at kernel summit) to flag pages that are owned and dynamically mapped by a device driver. The pmem driver, after mapping a persistent memory range into the system memmap via devm_memremap_pages(), arranges for DAX to distinguish pfn-only versus page-backed pmem-pfns via flags in the new pfn_t type. The DAX code, upon seeing a PFN_DEV+PFN_MAP flagged pfn, flags the resulting pte(s) inserted into the process page tables with a new _PAGE_DEVMAP flag. Later, when get_user_pages() is walking ptes it keys off _PAGE_DEVMAP to pin the device hosting the page range active. Finally, get_page() and put_page() are modified to take references against the device driver established page mapping. Finally, this need for "struct page" for persistent memory requires memory capacity to store the memmap array. Given the memmap array for a large pool of persistent may exhaust available DRAM introduce a mechanism to allocate the memmap from persistent memory. The new "struct vmem_altmap *" parameter to devm_memremap_pages() enables arch_add_memory() to use reserved pmem capacity rather than the page allocator. This patch (of 25): Both __dax_pmd_fault, and clear_pmem() were taking special steps to clear memory a page at a time to take advantage of non-temporal clear_page() implementations. However, x86_64 does not use non-temporal instructions for clear_page(), and arch_clear_pmem() was always incurring the cost of __arch_wb_cache_pmem(). Clean up the assumption that doing clear_pmem() a page at a time is more performant. Signed-off-by:
Dave Hansen <dave.hansen@linux.intel.com> Reviewed-by:
-
Kirill A. Shutemov authored
Sasha Levin has reported KASAN out-of-bounds bug[1]. It points to "if (!is_swap_pte(pte[i]))" in unfreeze_page_vma() as a problematic access. The cause is that split_huge_page() doesn't handle THP correctly if it's not allingned to PMD boundary. It can happen after mremap(). Test-case (not always triggers the bug): #define _GNU_SOURCE #include <stdio.h> #include <stdlib.h> #include <sys/mman.h> #define MB (1024UL*1024) #define SIZE (2*MB) #define BASE ((void *)0x400000000000) int main() { char *p; p = mmap(BASE, SIZE, PROT_READ | PROT_WRITE, MAP_FIXED | MAP_PRIVATE | MAP_ANONYMOUS | MAP_POPULATE, -1, 0); if (p == MAP_FAILED) perror("mmap"), exit(1); p = mremap(BASE, SIZE, SIZE, MREMAP_FIXED | MREMAP_MAYMOVE, BASE + SIZE + 8192); if (p == MAP_FAILED) perror("mremap"), exit(1); system("echo 1 > /sys/kernel/debug/split_huge_pages"); return 0; } The patch fixes freeze and unfreeze paths to handle page table boundary crossing. It also makes mapcount vs count check in split_huge_page_to_list() stricter: - after freeze we don't expect any subpage mapped as we remove them from rmap when setting up migration entries; - count must be 1, meaning only caller has reference to the page; [1] https://gist.github.com/sashalevin/c67fbea55e7c0576972aSigned-off-by:Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Reported-by:
Sasha Levin <sasha.levin@oracle.com> Signed-off-by:
-
Minchan Kim authored
We don't need to split THP page when MADV_FREE syscall is called if [start, len] is aligned with THP size. The split could be done when VM decide to free it in reclaim path if memory pressure is heavy. With that, we could avoid unnecessary THP split. For the feature, this patch changes pte dirtness marking logic of THP. Now, it marks every ptes of pages dirty unconditionally in splitting, which makes MADV_FREE void. So, instead, this patch propagates pmd dirtiness to all pages via PG_dirty and restores pte dirtiness from PG_dirty. With this, if pmd is clean(ie, MADV_FREEed) when split happens(e,g, shrink_page_list), all of pages are clean too so we could discard them. Signed-off-by:
Minchan Kim <minchan@kernel.org> Cc: Kirill A. Shutemov <kirill@shutemov.name> Cc: Hugh Dickins <hughd@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: "James E.J. Bottomley" <jejb@parisc-linux.org> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Shaohua Li <shli@kernel.org> Cc: <yalin.wang2010@gmail.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Chen Gang <gang.chen.5i5j@gmail.com> Cc: Chris Zankel <chris@zankel.net> Cc: Daniel Micay <danielmicay@gmail.com> Cc: Darrick J. Wong <darrick.wong@oracle.com> Cc: David S. Miller <davem@davemloft.net> Cc: Helge Deller <deller@gmx.de> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Jason Evans <je@fb.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Matt Turner <mattst88@gmail.com> Cc: Max Filippov <jcmvbkbc@gmail.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Mika Penttil <mika.penttila@nextfour.com> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Richard Henderson <rth@twiddle.net> Cc: Rik van Riel <riel@redhat.com> Cc: Roland Dreier <roland@kernel.org> Cc: Russell King <rmk@arm.linux.org.uk> Cc: Shaohua Li <shli@kernel.org> Cc: Will Deacon <will.deacon@arm.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by:
-
Minchan Kim authored
MADV_FREE needs pmd_dirty and pmd_mkclean for detecting recent overwrite of the contents since MADV_FREE syscall is called for THP page. This patch adds pmd_mkclean for THP page MADV_FREE support. Signed-off-by:
-
Minchan Kim authored
MADV_FREE needs pmd_dirty and pmd_mkclean for detecting recent overwrite of the contents since MADV_FREE syscall is called for THP page. This patch adds pmd_mkclean for THP page MADV_FREE support. Signed-off-by:
-
Minchan Kim authored
MADV_FREE needs pmd_dirty and pmd_mkclean for detecting recent overwrite of the contents since MADV_FREE syscall is called for THP page. This patch adds pmd_dirty and pmd_mkclean for THP page MADV_FREE support. Signed-off-by:
-
Minchan Kim authored
MADV_FREE needs pmd_dirty and pmd_mkclean for detecting recent overwrite of the contents since MADV_FREE syscall is called for THP page. This patch adds pmd_dirty and pmd_mkclean for THP page MADV_FREE support. Signed-off-by:
-
Minchan Kim authored
MADV_FREE needs pmd_dirty and pmd_mkclean for detecting recent overwrite of the contents since MADV_FREE syscall is called for THP page. This patch adds pmd_dirty and pmd_mkclean for THP page MADV_FREE support. Signed-off-by:
-
Minchan Kim authored
The MADV_FREE patchset changes page reclaim to simply free a clean anonymous page with no dirty ptes, instead of swapping it out; but KSM uses clean write-protected ptes to reference the stable ksm page. So be sure to mark that page dirty, so it's never mistakenly discarded. [hughd@google.com: adjusted comments] Signed-off-by:
Hugh Dickins <hughd@google.com> Cc: "James E.J. Bottomley" <jejb@parisc-linux.org> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Shaohua Li <shli@kernel.org> Cc: <yalin.wang2010@gmail.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Chen Gang <gang.chen.5i5j@gmail.com> Cc: Chris Zankel <chris@zankel.net> Cc: Daniel Micay <danielmicay@gmail.com> Cc: Darrick J. Wong <darrick.wong@oracle.com> Cc: David S. Miller <davem@davemloft.net> Cc: Helge Deller <deller@gmx.de> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Jason Evans <je@fb.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Kirill A. Shutemov <kirill@shutemov.name> Cc: Matt Turner <mattst88@gmail.com> Cc: Max Filippov <jcmvbkbc@gmail.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Mika Penttil <mika.penttila@nextfour.com> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Richard Henderson <rth@twiddle.net> Cc: Rik van Riel <riel@redhat.com> Cc: Roland Dreier <roland@kernel.org> Cc: Russell King <rmk@arm.linux.org.uk> Cc: Shaohua Li <shli@kernel.org> Cc: Will Deacon <will.deacon@arm.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by:
-
Minchan Kim authored
MADV_FREE is a hint that it's okay to discard pages if there is memory pressure and we use reclaimers(ie, kswapd and direct reclaim) to free them so there is no value keeping them in the active anonymous LRU so this patch moves them to inactive LRU list's head. This means that MADV_FREE-ed pages which were living on the inactive list are reclaimed first because they are more likely to be cold rather than recently active pages. An arguable issue for the approach would be whether we should put the page to the head or tail of the inactive list. I chose head because the kernel cannot make sure it's really cold or warm for every MADV_FREE usecase but at least we know it's not *hot*, so landing of inactive head would be a comprimise for various usecases. This fixes suboptimal behavior of MADV_FREE when pages living on the active list will sit there for a long time even under memory pressure while the inactive list is reclaimed heavily. This basically breaks the whole purpose of using MADV_FREE to help the system to free memory which is might not be used. Signed-off-by:
Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Shaohua Li <shli@kernel.org> Cc: "James E.J. Bottomley" <jejb@parisc-linux.org> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: <yalin.wang2010@gmail.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Chen Gang <gang.chen.5i5j@gmail.com> Cc: Chris Zankel <chris@zankel.net> Cc: Daniel Micay <danielmicay@gmail.com> Cc: Darrick J. Wong <darrick.wong@oracle.com> Cc: David S. Miller <davem@davemloft.net> Cc: Helge Deller <deller@gmx.de> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Jason Evans <je@fb.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Kirill A. Shutemov <kirill@shutemov.name> Cc: Matt Turner <mattst88@gmail.com> Cc: Max Filippov <jcmvbkbc@gmail.com> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Mika Penttil <mika.penttila@nextfour.com> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Richard Henderson <rth@twiddle.net> Cc: Roland Dreier <roland@kernel.org> Cc: Russell King <rmk@arm.linux.org.uk> Cc: Will Deacon <will.deacon@arm.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by:
-
Minchan Kim authored
When I test below piece of code with 12 processes(ie, 512M * 12 = 6G consume) on my (3G ram + 12 cpu + 8G swap, the madvise_free is siginficat slower (ie, 2x times) than madvise_dontneed. loop = 5; mmap(512M); while (loop--) { memset(512M); madvise(MADV_FREE or MADV_DONTNEED); } The reason is lots of swapin. 1) dontneed: 1,612 swapin 2) madvfree: 879,585 swapin If we find hinted pages were already swapped out when syscall is called, it's pointless to keep the swapped-out pages in pte. Instead, let's free the cold page because swapin is more expensive than (alloc page + zeroing). With this patch, it reduced swapin from 879,585 to 1,878 so elapsed time 1) dontneed: 6.10user 233.50system 0:50.44elapsed 2) madvfree: 6.03user 401.17system 1:30.67elapsed 2) madvfree + below patch: 6.70user 339.14system 1:04.45elapsed Signed-off-by:Michal Hocko <mhocko@suse.com> Acked-by:
-
Chen Gang authored
For uapi, need try to let all macros have same value, and MADV_FREE is added into main branch recently, so need redefine MADV_FREE for it. At present, '8' can be shared with all architectures, so redefine it to '8'. [sudipm.mukherjee@gmail.com: correct uniform value of MADV_FREE] Signed-off-by:
Chen Gang <gang.chen.5i5j@gmail.com> Signed-off-by:
Sudip Mukherjee <sudip@vectorindia.org> Signed-off-by:
-
Minchan Kim authored
Most architectures use asm-generic, but alpha, mips, parisc, xtensa need their own definitions. This patch defines MADV_FREE for them so it should fix build break for their architectures. Maybe, I should split and feed pieces to arch maintainers but included here for mmotm convenience. [gang.chen.5i5j@gmail.com: let MADV_FREE have same value for all architectures] Signed-off-by:
Max Filippov <jcmvbkbc@gmail.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Richard Henderson <rth@twiddle.net> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: "James E.J. Bottomley" <jejb@parisc-linux.org> Cc: Helge Deller <deller@gmx.de> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Chris Zankel <chris@zankel.net> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Shaohua Li <shli@kernel.org> Cc: <yalin.wang2010@gmail.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Chen Gang <gang.chen.5i5j@gmail.com> Cc: Daniel Micay <danielmicay@gmail.com> Cc: Darrick J. Wong <darrick.wong@oracle.com> Cc: David S. Miller <davem@davemloft.net> Cc: Hugh Dickins <hughd@google.com> Cc: Jason Evans <je@fb.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Kirill A. Shutemov <kirill@shutemov.name> Cc: Matt Turner <mattst88@gmail.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Michal Hocko <mhocko@suse.com> Cc: Mika Penttil <mika.penttila@nextfour.com> Cc: Rik van Riel <riel@redhat.com> Cc: Roland Dreier <roland@kernel.org> Cc: Russell King <rmk@arm.linux.org.uk> Cc: Shaohua Li <shli@kernel.org> Cc: Will Deacon <will.deacon@arm.com> Signed-off-by:
-
Minchan Kim authored
Linux doesn't have an ability to free pages lazy while other OS already have been supported that named by madvise(MADV_FREE). The gain is clear that kernel can discard freed pages rather than swapping out or OOM if memory pressure happens. Without memory pressure, freed pages would be reused by userspace without another additional overhead(ex, page fault + allocation + zeroing). Jason Evans said: : Facebook has been using MAP_UNINITIALIZED : (https://lkml.org/lkml/2012/1/18/308) in some of its applications for : several years, but there are operational costs to maintaining this : out-of-tree in our kernel and in jemalloc, and we are anxious to retire it : in favor of MADV_FREE. When we first enabled MAP_UNINITIALIZED it : increased throughput for much of our workload by ~5%, and although the : benefit has decreased using newer hardware and kernels, there is still : enough benefit that we cannot reasonably retire it without a replacement. : : Aside from Facebook operations, there are numerous broadly used : applications that would benefit from MADV_FREE. The ones that immediately : come to mind are redis, varnish, and MariaDB. I don't have much insight : into Android internals and development process, but I would hope to see : MADV_FREE support eventually end up there as well to benefit applications : linked with the integrated jemalloc. : : jemalloc will use MADV_FREE once it becomes available in the Linux kernel. : In fact, jemalloc already uses MADV_FREE or equivalent everywhere it's : available: *BSD, OS X, Windows, and Solaris -- every platform except Linux : (and AIX, but I'm not sure it even compiles on AIX). The lack of : MADV_FREE on Linux forced me down a long series of increasingly : sophisticated heuristics for madvise() volume reduction, and even so this : remains a common performance issue for people using jemalloc on Linux. : Please integrate MADV_FREE; many people will benefit substantially. How it works: When madvise syscall is called, VM clears dirty bit of ptes of the range. If memory pressure happens, VM checks dirty bit of page table and if it found still "clean", it means it's a "lazyfree pages" so VM could discard the page instead of swapping out. Once there was store operation for the page before VM peek a page to reclaim, dirty bit is set so VM can swap out the page instead of discarding. One thing we should notice is that basically, MADV_FREE relies on dirty bit in page table entry to decide whether VM allows to discard the page or not. IOW, if page table entry includes marked dirty bit, VM shouldn't discard the page. However, as a example, if swap-in by read fault happens, page table entry doesn't have dirty bit so MADV_FREE could discard the page wrongly. For avoiding the problem, MADV_FREE did more checks with PageDirty and PageSwapCache. It worked out because swapped-in page lives on swap cache and since it is evicted from the swap cache, the page has PG_dirty flag. So both page flags check effectively prevent wrong discarding by MADV_FREE. However, a problem in above logic is that swapped-in page has PG_dirty still after they are removed from swap cache so VM cannot consider the page as freeable any more even if madvise_free is called in future. Look at below example for detail. ptr = malloc(); memset(ptr); .. .. .. heavy memory pressure so all of pages are swapped out .. .. var = *ptr; -> a page swapped-in and could be removed from swapcache. Then, page table doesn't mark dirty bit and page descriptor includes PG_dirty .. .. madvise_free(ptr); -> It doesn't clear PG_dirty of the page. .. .. .. .. heavy memory pressure again. .. In this time, VM cannot discard the page because the page .. has *PG_dirty* To solve the problem, this patch clears PG_dirty if only the page is owned exclusively by current process when madvise is called because PG_dirty represents ptes's dirtiness in several processes so we could clear it only if we own it exclusively. Firstly, heavy users would be general allocators(ex, jemalloc, tcmalloc and hope glibc supports it) and jemalloc/tcmalloc already have supported the feature for other OS(ex, FreeBSD) barrios@blaptop:~/benchmark/ebizzy$ lscpu Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 12 On-line CPU(s) list: 0-11 Thread(s) per core: 1 Core(s) per socket: 1 Socket(s): 12 NUMA node(s): 1 Vendor ID: GenuineIntel CPU family: 6 Model: 2 Stepping: 3 CPU MHz: 3200.185 BogoMIPS: 6400.53 Virtualization: VT-x Hypervisor vendor: KVM Virtualization type: full L1d cache: 32K L1i cache: 32K L2 cache: 4096K NUMA node0 CPU(s): 0-11 ebizzy benchmark(./ebizzy -S 10 -n 512) Higher avg is better. vanilla-jemalloc MADV_free-jemalloc 1 thread records: 10 records: 10 avg: 2961.90 avg: 12069.70 std: 71.96(2.43%) std: 186.68(1.55%) max: 3070.00 max: 12385.00 min: 2796.00 min: 11746.00 2 thread records: 10 records: 10 avg: 5020.00 avg: 17827.00 std: 264.87(5.28%) std: 358.52(2.01%) max: 5244.00 max: 18760.00 min: 4251.00 min: 17382.00 4 thread records: 10 records: 10 avg: 8988.80 avg: 27930.80 std: 1175.33(13.08%) std: 3317.33(11.88%) max: 9508.00 max: 30879.00 min: 5477.00 min: 21024.00 8 thread records: 10 records: 10 avg: 13036.50 avg: 33739.40 std: 170.67(1.31%) std: 5146.22(15.25%) max: 13371.00 max: 40572.00 min: 12785.00 min: 24088.00 16 thread records: 10 records: 10 avg: 11092.40 avg: 31424.20 std: 710.60(6.41%) std: 3763.89(11.98%) max: 12446.00 max: 36635.00 min: 9949.00 min: 25669.00 32 thread records: 10 records: 10 avg: 11067.00 avg: 34495.80 std: 971.06(8.77%) std: 2721.36(7.89%) max: 12010.00 max: 38598.00 min: 9002.00 min: 30636.00 In summary, MADV_FREE is about much faster than MADV_DONTNEED. This patch (of 12): Add core MADV_FREE implementation. [akpm@linux-foundation.org: small cleanups] Signed-off-by:
-
Jerome Marchand authored
The use of idr_remove() is forbidden in the callback functions of idr_for_each(). It is therefore unsafe to call idr_remove in zram_remove(). This patch moves the call to idr_remove() from zram_remove() to hot_remove_store(). In the detroy_devices() path, idrs are removed by idr_destroy(). This solves an use-after-free detected by KASan. [akpm@linux-foundation.org: fix coding stype, per Sergey] Signed-off-by:
Jerome Marchand <jmarchan@redhat.com> Acked-by:
Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Minchan Kim <minchan@kernel.org> Cc: <stable@vger.kernel.org> [4.2+] Signed-off-by:
-
Vladimir Davydov authored
page_referenced_one() and page_idle_clear_pte_refs_one() duplicate the code for looking up pte of a (possibly transhuge) page. Move this code to a new helper function, page_check_address_transhuge(), and make the above mentioned functions use it. This is just a cleanup, no functional changes are intended. Signed-off-by:
Vladimir Davydov <vdavydov@virtuozzo.com> Reviewed-by:
-
Kirill A. Shutemov authored
During freeze_page(), we remove the page from rmap. It munlocks the page if it was mlocked. clear_page_mlock() uses thelru cache, which temporary pins the page. Let's drain the lru cache before checking page's count vs. mapcount. The change makes mlocked page split on first attempt, if it was not pinned by somebody else. Signed-off-by:
Vlastimil Babka <vbabka@suse.cz> Signed-off-by:
-
Kirill A. Shutemov authored
Writing 1 into 'split_huge_pages' will try to find and split all huge pages in the system. This is useful for debuging. [akpm@linux-foundation.org: fix printk text, per Vlastimil] Signed-off-by:
-
Kirill A. Shutemov authored
Both page_referenced() and page_idle_clear_pte_refs_one() assume that THP can only be mapped with PMD, so there's no reason to look on PTEs for PageTransHuge() pages. That's no true anymore: THP can be mapped with PTEs too. The patch removes PageTransHuge() test from the functions and opencode page table check. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by:
-
Kirill A. Shutemov authored
Before THP refcounting rework, THP was not allowed to cross VMA boundary. So, if we have THP and we split it, PG_mlocked can be safely transferred to small pages. With new THP refcounting and naive approach to mlocking we can end up with this scenario: 1. we have a mlocked THP, which belong to one VM_LOCKED VMA. 2. the process does munlock() on the *part* of the THP: - the VMA is split into two, one of them VM_LOCKED; - huge PMD split into PTE table; - THP is still mlocked; 3. split_huge_page(): - it transfers PG_mlocked to *all* small pages regrardless if it blong to any VM_LOCKED VMA. We probably could munlock() all small pages on split_huge_page(), but I think we have accounting issue already on step two. Instead of forbidding mlocked pages altogether, we just avoid mlocking PTE-mapped THPs and munlock THPs on split_huge_pmd(). This means PTE-mapped THPs will be on normal lru lists and will be split under memory pressure by vmscan. After the split vmscan will detect unevictable small pages and mlock them. With this approach we shouldn't hit situation like described above. Signed-off-by: -
Kirill A. Shutemov authored
The patch updates Documentation/vm/transhuge.txt to reflect changes in THP design. Signed-off-by:
-
Kirill A. Shutemov authored
All parts of THP with new refcounting are now in place. We can now allow to enable THP. Signed-off-by:
Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by:
-
Kirill A. Shutemov authored
Currently we don't split huge page on partial unmap. It's not an ideal situation. It can lead to memory overhead. Furtunately, we can detect partial unmap on page_remove_rmap(). But we cannot call split_huge_page() from there due to locking context. It's also counterproductive to do directly from munmap() codepath: in many cases we will hit this from exit(2) and splitting the huge page just to free it up in small pages is not what we really want. The patch introduce deferred_split_huge_page() which put the huge page into queue for splitting. The splitting itself will happen when we get memory pressure via shrinker interface. The page will be dropped from list on freeing through compound page destructor. Signed-off-by:
-
Kirill A. Shutemov authored
We are not able to migrate THPs. It means it's not enough to split only PMD on migration -- we need to split compound page under it too. Signed-off-by:
-
Kirill A. Shutemov authored
This patch adds implementation of split_huge_page() for new refcountings. Unlike previous implementation, new split_huge_page() can fail if somebody holds GUP pin on the page. It also means that pin on page would prevent it from bening split under you. It makes situation in many places much cleaner. The basic scheme of split_huge_page(): - Check that sum of mapcounts of all subpage is equal to page_count() plus one (caller pin). Foll off with -EBUSY. This way we can avoid useless PMD-splits. - Freeze the page counters by splitting all PMD and setup migration PTEs. - Re-check sum of mapcounts against page_count(). Page's counts are stable now. -EBUSY if page is pinned. - Split compound page. - Unfreeze the page by removing migration entries. Signed-off-by: -
Naoya Horiguchi authored
Some mm-related BUG_ON()s could trigger from hwpoison code due to recent changes in thp refcounting rule. This patch fixes them up. In the new refcounting, we no longer use tail->_mapcount to keep tail's refcount, and thereby we can simplify get/put_hwpoison_page(). And another change is that tail's refcount is not transferred to the raw page during thp split (more precisely, in new rule we don't take refcount on tail page any more.) So when we need thp split, we have to transfer the refcount properly to the 4kB soft-offlined page before migration. thp split code goes into core code only when precheck (total_mapcount(head) == page_count(head) - 1) passes to avoid useless split, where we assume that one refcount is held by the caller of thp split and the others are taken via mapping. To meet this assumption, this patch moves thp split part in soft_offline_page() after get_any_page(). [akpm@linux-foundation.org: remove unneeded #define, per Kirill] Signed-off-by:
Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Acked-by:
-
Naoya Horiguchi authored
I saw the following BUG_ON triggered in a testcase where a process calls madvise(MADV_SOFT_OFFLINE) on thps, along with a background process that calls migratepages command repeatedly (doing ping-pong among different NUMA nodes) for the first process: Soft offlining page 0x60000 at 0x700000600000 __get_any_page: 0x60000 free buddy page page:ffffea0001800000 count:0 mapcount:-127 mapping: (null) index:0x1 flags: 0x1fffc0000000000() page dumped because: VM_BUG_ON_PAGE(atomic_read(&page->_count) == 0) ------------[ cut here ]------------ kernel BUG at /src/linux-dev/include/linux/mm.h:342! invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC Modules linked in: cfg80211 rfkill crc32c_intel serio_raw virtio_balloon i2c_piix4 virtio_blk virtio_net ata_generic pata_acpi CPU: 3 PID: 3035 Comm: test_alloc_gene Tainted: G O 4.4.0-rc8-v4.4-rc8-160107-1501-00000-rc8+ #74 Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 task: ffff88007c63d5c0 ti: ffff88007c210000 task.ti: ffff88007c210000 RIP: 0010:[<ffffffff8118998c>] [<ffffffff8118998c>] put_page+0x5c/0x60 RSP: 0018:ffff88007c213e00 EFLAGS: 00010246 Call Trace: put_hwpoison_page+0x4e/0x80 soft_offline_page+0x501/0x520 SyS_madvise+0x6bc/0x6f0 entry_SYSCALL_64_fastpath+0x12/0x6a Code: 8b fc ff ff 5b 5d c3 48 89 df e8 b0 fa ff ff 48 89 df 31 f6 e8 c6 7d ff ff 5b 5d c3 48 c7 c6 08 54 a2 81 48 89 df e8 a4 c5 01 00 <0f> 0b 66 90 66 66 66 66 90 55 48 89 e5 41 55 41 54 53 48 8b 47 RIP [<ffffffff8118998c>] put_page+0x5c/0x60 RSP <ffff88007c213e00> The root cause resides in get_any_page() which retries to get a refcount of the page to be soft-offlined. This function calls put_hwpoison_page(), expecting that the target page is putback to LRU list. But it can be also freed to buddy. So the second check need to care about such case. Fixes: af8fae7c ("mm/memory-failure.c: clean up soft_offline_page()") Signed-off-by: -
Kirill A. Shutemov authored
We're going to use migration entries instead of compound_lock() to stabilize page refcounts. Setup and remove migration entries require page to be locked. Some of split_huge_page() callers already have the page locked. Let's require everybody to lock the page before calling split_huge_page(). Signed-off-by:
-
Kirill A. Shutemov authored
We are going to use migration PTE entries to stabilize page counts. If the page is mapped with PMDs we need to split the PMD and setup migration entries. It's reasonable to combine these operations to avoid double-scanning over the page table. Signed-off-by:
-
Kirill A. Shutemov authored
Original split_huge_page() combined two operations: splitting PMDs into tables of PTEs and splitting underlying compound page. This patch implements split_huge_pmd() which split given PMD without splitting other PMDs this page mapped with or underlying compound page. Without tail page refcounting, implementation of split_huge_pmd() is pretty straight-forward. Signed-off-by:
-
Kirill A. Shutemov authored
We're going to have THP mapped with PTEs. It will confuse numabalancing. Let's skip them for now. Signed-off-by:
-
Kirill A. Shutemov authored
Let's define page_mapped() to be true for compound pages if any sub-pages of the compound page is mapped (with PMD or PTE). On other hand page_mapcount() return mapcount for this particular small page. This will make cases like page_get_anon_vma() behave correctly once we allow huge pages to be mapped with PTE. Most users outside core-mm should use page_mapcount() instead of page_mapped(). Signed-off-by:
-
