- 30 Apr, 2021 40 commits
-
-
Nicholas Piggin authored
vmalloc_to_page returns NULL for addresses mapped by larger pages[*]. Whether or not a vmap is huge depends on the architecture details, alignments, boot options, etc., which the caller can not be expected to know. Therefore HUGE_VMAP is a regression for vmalloc_to_page. This change teaches vmalloc_to_page about larger pages, and returns the struct page that corresponds to the offset within the large page. This makes the API agnostic to mapping implementation details. [*] As explained by commit 029c54b0 ("mm/vmalloc.c: huge-vmap: fail gracefully on unexpected huge vmap mappings") [npiggin@gmail.com: sparc32: add stub pud_page define for walking huge vmalloc page tables] Link: https://lkml.kernel.org/r/20210324232825.1157363-1-npiggin@gmail.com Link: https://lkml.kernel.org/r/20210317062402.533919-3-npiggin@gmail.comSigned-off-by:
Nicholas Piggin <npiggin@gmail.com> Reviewed-by:
Miaohe Lin <linmiaohe@huawei.com> Reviewed-by:
Christoph Hellwig <hch@lst.de> Cc: Borislav Petkov <bp@alien8.de> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Ding Tianhong <dingtianhong@huawei.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Russell King <linux@armlinux.org.uk> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Uladzislau Rezki (Sony) <urezki@gmail.com> Cc: Will Deacon <will@kernel.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: David S. Miller <davem@davemloft.net> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Nicholas Piggin authored
Patch series "huge vmalloc mappings", v13. The kernel virtual mapping layer grew support for mapping memory with > PAGE_SIZE ptes with commit 0ddab1d2 ("lib/ioremap.c: add huge I/O map capability interfaces"), and implemented support for using those huge page mappings with ioremap. According to the submission, the use-case is mapping very large non-volatile memory devices, which could be GB or TB: https://lore.kernel.org/lkml/1425404664-19675-1-git-send-email-toshi.kani@hp.com/ The benefit is said to be in the overhead of maintaining the mapping, perhaps both in memory overhead and setup / teardown time. Memory overhead for the mapping with a 4kB page and 8 byte page table is 2GB per TB of mapping, down to 4MB / TB with 2MB pages. The same huge page vmap infrastructure can be quite easily adapted and used for mapping vmalloc memory pages without more complexity for arch or core vmap code. However unlike ioremap, vmalloc page table overhead is not a real problem, so the advantage to justify this is performance. Several of the most structures in the kernel (e.g., vfs and network hash tables) are allocated with vmalloc on NUMA machines, in order to distribute access bandwidth over the machine. Mapping these with larger pages can improve TLB usage significantly, for example this reduces TLB misses by nearly 30x on a `git diff` workload on a 2-node POWER9 (59,800 -> 2,100) and reduces CPU cycles by 0.54%, due to vfs hashes being allocated with 2MB pages. [ Other numbers? - The difference is even larger in a guest due to more costly TLB misses. - Eric Dumazet was keen on the network hash performance possibilities. - Other archs? Ding was doing x86 testing. ] The kernel module allocator also uses vmalloc to map module images even on non-NUMA, which can result in high iTLB pressure on highly modular distro type of kernels. This series does not implement huge mappings for modules yet, but it's a step along the way. Rick Edgecombe was looking at that IIRC. The per-cpu allocator similarly might be able to take advantage of this. Also on the todo list. The disadvantages of this I can see are: * Memory fragmentation can waste some physical memory because it will attempt to allocate larger pages to fit the required size, rounding up (once the requested size is >= 2MB). - I don't see it being a big problem in practice unless some user crops up that allocates thousands of 2.5MB ranges. We can tewak heuristics a bit there if needed to reduce peak waste. * Less granular mappings can make the NUMA distribution less balanced. - Similar to the above. - Could also allocate all major system hashes with one allocation up-front and spread them all across the one block, which should help overall NUMA distribution and reduce fragmentation waste. * Callers might expect something about the underlying allocated pages. - Tried to keep the apperance of base PAGE_SIZE pages throughout the APIs and exposed data structures. - Added a VM_NO_HUGE_VMAP flag to hammer troublesome cases with. - Finally, added a nohugevmalloc boot option to turn it off (independent of nohugeiomap). This patch (of 14): ARM uses its own PMD folding scheme which is missing pud_page which should just pass through to pmd_page. Move this from the 3-level page table to common header. Link: https://lkml.kernel.org/r/20210317062402.533919-2-npiggin@gmail.comSigned-off-by:
-
Serapheim Dimitropoulos authored
vread() has been linearly searching vmap_area_list for looking up vmalloc areas to read from. These same areas are also tracked by a rb_tree (vmap_area_root) which offers logarithmic lookup. This patch modifies vread() to use the rb_tree structure instead of the list and the speedup for heavy /proc/kcore readers can be pretty significant. Below are the wall clock measurements of a Python application that leverages the drgn debugging library to read and interpret data read from /proc/kcore. Before the patch: ----- $ time sudo sdb -e 'dbuf | head 3000 | wc' (unsigned long)3000 real 0m22.446s user 0m2.321s sys 0m20.690s ----- With the patch: ----- $ time sudo sdb -e 'dbuf | head 3000 | wc' (unsigned long)3000 real 0m2.104s user 0m2.043s sys 0m0.921s ----- Link: https://lkml.kernel.org/r/20210209190253.108763-1-serapheim@delphix.comSigned-off-by:
Serapheim Dimitropoulos <serapheim@delphix.com> Reviewed-by:
Uladzislau Rezki (Sony) <urezki@gmail.com> Signed-off-by:
-
Christoph Hellwig authored
remap_vmalloc_range_partial is only used to implement remap_vmalloc_range and by procfs. Unexport it. Link: https://lkml.kernel.org/r/20210301082235.932968-3-hch@lst.deSigned-off-by:
-
Christoph Hellwig authored
Patch series "remap_vmalloc_range cleanups". This series removes an open coded instance of remap_vmalloc_range and removes the unused remap_vmalloc_range_partial export. This patch (of 2): Use remap_vmalloc_range instead of open coding it using remap_vmalloc_range_partial. Link: https://lkml.kernel.org/r/20210301082235.932968-1-hch@lst.de Link: https://lkml.kernel.org/r/20210301082235.932968-2-hch@lst.deSigned-off-by:
-
Wang Wensheng authored
sparse_buffer_init() and sparse_buffer_fini() should appear in pair, or a WARN issue would be through the next time sparse_buffer_init() runs. Add the missing sparse_buffer_fini() in error branch. Link: https://lkml.kernel.org/r/20210325113155.118574-1-wangwensheng4@huawei.com Fixes: 85c77f79 ("mm/sparse: add new sparse_init_nid() and sparse_init()") Signed-off-by:
Wang Wensheng <wangwensheng4@huawei.com> Reviewed-by:
David Hildenbrand <david@redhat.com> Reviewed-by:
Oscar Salvador <osalvador@suse.de> Cc: Pavel Tatashin <pasha.tatashin@oracle.com> Signed-off-by:
-
Zhiyuan Dai authored
strlcpy is marked as deprecated in Documentation/process/deprecated.rst, and there is no functional difference when the caller expects truncation (when not checking the return value). strscpy is relatively better as it also avoids scanning the whole source string. Link: https://lkml.kernel.org/r/1613962050-14188-1-git-send-email-daizhiyuan@phytium.com.cnSigned-off-by:
Zhiyuan Dai <daizhiyuan@phytium.com.cn> Signed-off-by:
-
Brian Geffon authored
This test extends the current mremap tests to validate that the MREMAP_DONTUNMAP operation can be performed on shmem mappings. Link: https://lkml.kernel.org/r/20210323182520.2712101-3-bgeffon@google.comSigned-off-by:
Brian Geffon <bgeffon@google.com> Cc: Axel Rasmussen <axelrasmussen@google.com> Cc: Lokesh Gidra <lokeshgidra@google.com> Cc: Mike Rapoport <rppt@linux.vnet.ibm.com> Cc: Peter Xu <peterx@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: "Michael S . Tsirkin" <mst@redhat.com> Cc: Brian Geffon <bgeffon@google.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Sonny Rao <sonnyrao@google.com> Cc: Minchan Kim <minchan@kernel.org> Cc: "Kirill A . Shutemov" <kirill@shutemov.name> Cc: Dmitry Safonov <dima@arista.com> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Alejandro Colomar <alx.manpages@gmail.com> Signed-off-by:
-
Brian Geffon authored
This reverts commit cd544fd1. As discussed in [1] this commit was a no-op because the mapping type was checked in vma_to_resize before move_vma is ever called. This meant that vm_ops->mremap() would never be called on such mappings. Furthermore, we've since expanded support of MREMAP_DONTUNMAP to non-anonymous mappings, and these special mappings are still protected by the existing check of !VM_DONTEXPAND and !VM_PFNMAP which will result in a -EINVAL. 1. https://lkml.org/lkml/2020/12/28/2340 Link: https://lkml.kernel.org/r/20210323182520.2712101-2-bgeffon@google.comSigned-off-by:
Hugh Dickins <hughd@google.com> Reviewed-by:
Dmitry Safonov <0x7f454c46@gmail.com> Cc: Alejandro Colomar <alx.manpages@gmail.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Axel Rasmussen <axelrasmussen@google.com> Cc: "Kirill A . Shutemov" <kirill@shutemov.name> Cc: Lokesh Gidra <lokeshgidra@google.com> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: "Michael S . Tsirkin" <mst@redhat.com> Cc: Mike Rapoport <rppt@linux.vnet.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Peter Xu <peterx@redhat.com> Cc: Sonny Rao <sonnyrao@google.com> Cc: Vlastimil Babka <vbabka@suse.cz> Signed-off-by:
-
Brian Geffon authored
Patch series "mm: Extend MREMAP_DONTUNMAP to non-anonymous mappings", v5. This patch (of 3): Currently MREMAP_DONTUNMAP only accepts private anonymous mappings. This restriction was placed initially for simplicity and not because there exists a technical reason to do so. This change will widen the support to include any mappings which are not VM_DONTEXPAND or VM_PFNMAP. The primary use case is to support MREMAP_DONTUNMAP on mappings which may have been created from a memfd. This change will result in mremap(MREMAP_DONTUNMAP) returning -EINVAL if VM_DONTEXPAND or VM_PFNMAP mappings are specified. Lokesh Gidra who works on the Android JVM, provided an explanation of how such a feature will improve Android JVM garbage collection: "Android is developing a new garbage collector (GC), based on userfaultfd. The garbage collector will use userfaultfd (uffd) on the java heap during compaction. On accessing any uncompacted page, the application threads will find it missing, at which point the thread will create the compacted page and then use UFFDIO_COPY ioctl to get it mapped and then resume execution. Before starting this compaction, in a stop-the-world pause the heap will be mremap(MREMAP_DONTUNMAP) so that the java heap is ready to receive UFFD_EVENT_PAGEFAULT events after resuming execution. To speedup mremap operations, pagetable movement was optimized by moving PUD entries instead of PTE entries [1]. It was necessary as mremap of even modest sized memory ranges also took several milliseconds, and stopping the application for that long isn't acceptable in response-time sensitive cases. With UFFDIO_CONTINUE feature [2], it will be even more efficient to implement this GC, particularly the 'non-moveable' portions of the heap. It will also help in reducing the need to copy (UFFDIO_COPY) the pages. However, for this to work, the java heap has to be on a 'shared' vma. Currently MREMAP_DONTUNMAP only supports private anonymous mappings, this patch will enable using UFFDIO_CONTINUE for the new userfaultfd-based heap compaction." [1] https://lore.kernel.org/linux-mm/20201215030730.NC3CU98e4%25akpm@linux-foundation.org/ [2] https://lore.kernel.org/linux-mm/20210302000133.272579-1-axelrasmussen@google.com/ Link: https://lkml.kernel.org/r/20210323182520.2712101-1-bgeffon@google.comSigned-off-by:
Lokesh Gidra <lokeshgidra@google.com> Reviewed-by:
-
Huang Ying authored
With NUMA balancing, in hint page fault handler, the faulting page will be migrated to the accessing node if necessary. During the migration, TLB will be shot down on all CPUs that the process has run on recently. Because in the hint page fault handler, the PTE will be made accessible before the migration is tried. The overhead of TLB shooting down can be high, so it's better to be avoided if possible. In fact, if we delay mapping the page until migration, that can be avoided. This is what this patch doing. For the multiple threads applications, it's possible that a page is accessed by multiple threads almost at the same time. In the original implementation, because the first thread will install the accessible PTE before migrating the page, the other threads may access the page directly before the page is made inaccessible again during migration. While with the patch, the second thread will go through the page fault handler too. And because of the PageLRU() checking in the following code path, migrate_misplaced_page() numamigrate_isolate_page() isolate_lru_page() the migrate_misplaced_page() will return 0, and the PTE will be made accessible in the second thread. This will introduce a little more overhead. But we think the possibility for a page to be accessed by the multiple threads at the same time is low, and the overhead difference isn't too large. If this becomes a problem in some workloads, we need to consider how to reduce the overhead. To test the patch, we run a test case as follows on a 2-socket Intel server (1 NUMA node per socket) with 128GB DRAM (64GB per socket). 1. Run a memory eater on NUMA node 1 to use 40GB memory before running pmbench. 2. Run pmbench (normal accessing pattern) with 8 processes, and 8 threads per process, so there are 64 threads in total. The working-set size of each process is 8960MB, so the total working-set size is 8 * 8960MB = 70GB. The CPU of all pmbench processes is bound to node 1. The pmbench processes will access some DRAM on node 0. 3. After the pmbench processes run for 10 seconds, kill the memory eater. Now, some pages will be migrated from node 0 to node 1 via NUMA balancing. Test results show that, with the patch, the pmbench throughput (page accesses/s) increases 5.5%. The number of the TLB shootdowns interrupts reduces 98% (from ~4.7e7 to ~9.7e5) with about 9.2e6 pages (35.8GB) migrated. From the perf profile, it can be found that the CPU cycles spent by try_to_unmap() and its callees reduces from 6.02% to 0.47%. That is, the CPU cycles spent by TLB shooting down decreases greatly. Link: https://lkml.kernel.org/r/20210408132236.1175607-1-ying.huang@intel.comSigned-off-by:"Huang, Ying" <ying.huang@intel.com> Reviewed-by:
Mel Gorman <mgorman@suse.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Peter Xu <peterx@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: "Matthew Wilcox" <willy@infradead.org> Cc: Will Deacon <will@kernel.org> Cc: Michel Lespinasse <walken@google.com> Cc: Arjun Roy <arjunroy@google.com> Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> Signed-off-by:
-
Christoph Hellwig authored
remap_io_sg claims that the pgprot is pre-verified using an io_mapping, but actually does not get passed an io_mapping and just uses the pgprot in the VMA. Remove the apply_to_page_range abuse and just loop over remap_pfn_range for each segment. Note: this could use io_mapping_map_user by passing an iomap to remap_io_sg if the maintainers can verify that the pgprot in the iomap in the only caller is indeed the desired one here. Link: https://lkml.kernel.org/r/20210326055505.1424432-5-hch@lst.deSigned-off-by:
-
Christoph Hellwig authored
Replace the home-grown remap_io_mapping that abuses apply_to_page_range with the proper io_mapping_map_user interface. Link: https://lkml.kernel.org/r/20210326055505.1424432-4-hch@lst.deSigned-off-by:
-
Christoph Hellwig authored
Add a helper that calls remap_pfn_range for an struct io_mapping, relying on the pgprot pre-validation done when creating the mapping instead of doing it at runtime. Link: https://lkml.kernel.org/r/20210326055505.1424432-3-hch@lst.deSigned-off-by:
-
Christoph Hellwig authored
Patch series "add remap_pfn_range_notrack instead of reinventing it in i915", v2. i915 has some reason to want to avoid the track_pfn_remap overhead in remap_pfn_range. Add a function to the core VM to do just that rather than reinventing the functionality poorly in the driver. Note that the remap_io_sg path does get exercises when using Xorg on my Thinkpad X1, so this should be considered lightly tested, I've not managed to hit the remap_io_mapping path at all. This patch (of 4): Add a version of remap_pfn_range that does not call track_pfn_range. This will be used to fix horrible abuses of VM internals in the i915 driver. Link: https://lkml.kernel.org/r/20210326055505.1424432-1-hch@lst.de Link: https://lkml.kernel.org/r/20210326055505.1424432-2-hch@lst.deSigned-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Peter Zijlstra <peterz@infradead.org> Signed-off-by:
-
Ovidiu Panait authored
Adjust the rss_stat tracepoint to print the name of the resident page type that got updated (e.g. MM_ANONPAGES/MM_FILEPAGES), rather than the numeric index corresponding to it (the __entry->member value): Before this patch: ------------------ rss_stat: mm_id=1216113068 curr=0 member=1 size=28672B rss_stat: mm_id=1216113068 curr=0 member=1 size=0B rss_stat: mm_id=534402304 curr=1 member=0 size=188416B rss_stat: mm_id=534402304 curr=1 member=1 size=40960B After this patch: ----------------- rss_stat: mm_id=1726253524 curr=1 type=MM_ANONPAGES size=40960B rss_stat: mm_id=1726253524 curr=1 type=MM_FILEPAGES size=663552B rss_stat: mm_id=1726253524 curr=1 type=MM_ANONPAGES size=65536B rss_stat: mm_id=1726253524 curr=1 type=MM_FILEPAGES size=647168B Use TRACE_DEFINE_ENUM()/__print_symbolic() logic to map the enum values to the strings they represent, so that userspace tools can also parse the raw data correctly. Link: https://lkml.kernel.org/r/20210310162305.4862-1-ovidiu.panait@windriver.comSigned-off-by:
Ovidiu Panait <ovidiu.panait@windriver.com> Suggested-by:
Steven Rostedt (VMware) <rostedt@goodmis.org> Reviewed-by:
-
Oscar Salvador authored
We can optimize in the case we are adding consecutive sections, so no memset(PAGE_UNUSED) is needed. In that case, let us keep track where the unused range of the previous memory range begins, so we can compare it with start of the range to be added. If they are equal, we know sections are added consecutively. For that purpose, let us introduce 'unused_pmd_start', which always holds the beginning of the unused memory range. In the case a section does not contiguously follow the previous one, we know we can memset [unused_pmd_start, PMD_BOUNDARY) with PAGE_UNUSE. This patch is based on a similar patch by David Hildenbrand: https://lore.kernel.org/linux-mm/20200722094558.9828-10-david@redhat.com/ Link: https://lkml.kernel.org/r/20210309214050.4674-5-osalvador@suse.deSigned-off-by:
Dave Hansen <dave.hansen@linux.intel.com> Cc: Andy Lutomirski <luto@kernel.org> Cc: Borislav Petkov <bp@alien8.de> Cc: "H . Peter Anvin" <hpa@zytor.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Zi Yan <ziy@nvidia.com> Signed-off-by:
-
Oscar Salvador authored
When sizeof(struct page) is not a power of 2, sections do not span a PMD anymore and so when populating them some parts of the PMD will remain unused. Because of this, PMDs will be left behind when depopulating sections since remove_pmd_table() thinks that those unused parts are still in use. Fix this by marking the unused parts with PAGE_UNUSED, so memchr_inv() will do the right thing and will let us free the PMD when the last user of it is gone. This patch is based on a similar patch by David Hildenbrand: https://lore.kernel.org/linux-mm/20200722094558.9828-9-david@redhat.com/ [osalvador@suse.de: go back to the ifdef version] Link: https://lkml.kernel.org/r/YGy++mSft7K4u+88@localhost.localdomain Link: https://lkml.kernel.org/r/20210309214050.4674-4-osalvador@suse.deSigned-off-by:
-
Oscar Salvador authored
There is no code to allocate 1GB pages when mapping the vmemmap range as this might waste some memory and requires more complexity which is not really worth. Drop the dead code both for the aligned and unaligned cases and leave only the direct map handling. Link: https://lkml.kernel.org/r/20210309214050.4674-3-osalvador@suse.deSigned-off-by:
-
Oscar Salvador authored
Patch series "Cleanup and fixups for vmemmap handling", v6. This series contains cleanups to remove dead code that handles unaligned cases for 4K and 1GB pages (patch#1 and patch#2) when removing the vemmmap range, and a fix (patch#3) to handle the case when two vmemmap ranges intersect the same PMD. This patch (of 4): remove_pte_table() is prepared to handle the case where either the start or the end of the range is not PAGE aligned. This cannot actually happen: __populate_section_memmap enforces the range to be PMD aligned, so as long as the size of the struct page remains multiple of 8, the vmemmap range will be aligned to PAGE_SIZE. Drop the dead code and place a VM_BUG_ON in vmemmap_{populate,free} to catch nasty cases. Note that the VM_BUG_ON is placed in there because vmemmap_{populate,free= } is the gate of all removing and freeing page tables logic. Link: https://lkml.kernel.org/r/20210309214050.4674-1-osalvador@suse.de Link: https://lkml.kernel.org/r/20210309214050.4674-2-osalvador@suse.deSigned-off-by: -
Zhiyuan Dai authored
Add a comment explaining the value of the ISSTATIC parameter, Inform the reader that this is not a coding style issue. Link: https://lkml.kernel.org/r/1613964695-17614-1-git-send-email-daizhiyuan@phytium.com.cnSigned-off-by:
-
Wang Qing authored
Smatch gives the warning: do_numa_page() warn: assigning (-11) to unsigned variable 'migrated' Link: https://lkml.kernel.org/r/1614603421-2681-1-git-send-email-wangqing@vivo.comSigned-off-by:
Wang Qing <wangqing@vivo.com> Reviewed-by:
-
Johannes Weiner authored
When the unsigned page_counter underflows, even just by a few pages, a cgroup will not be able to run anything afterwards and trigger the OOM killer in a loop. Underflows shouldn't happen, but when they do in practice, we may just be off by a small amount that doesn't interfere with the normal operation - consequences don't need to be that dire. Reset the page_counter to 0 upon underflow. We'll issue a warning that the accounting will be off and then try to keep limping along. [ We used to do this with the original res_counter, where it was a more straight-forward correction inside the spinlock section. I didn't carry it forward into the lockless page counters for simplicity, but it turns out this is quite useful in practice. ] Link: https://lkml.kernel.org/r/20210408143155.2679744-1-hannes@cmpxchg.orgSigned-off-by:
Johannes Weiner <hannes@cmpxchg.org> Acked-by:
Michal Hocko <mhocko@suse.com> Acked-by:
Chris Down <chris@chrisdown.name> Reviewed-by:
Shakeel Butt <shakeelb@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by:
-
Wan Jiabing authored
struct mem_cgroup is declared twice. One has been declared at forward struct declaration. Remove the duplicate. Link: https://lkml.kernel.org/r/20210330020246.2265371-1-wanjiabing@vivo.comSigned-off-by:
Wan Jiabing <wanjiabing@vivo.com> Reviewed-by:
Muchun Song <songmuchun@bytedance.com> Acked-by:
-
Muchun Song authored
The page only can be marked as kmem when CONFIG_MEMCG_KMEM is enabled. So move PageMemcgKmem() to the scope of the CONFIG_MEMCG_KMEM. As a bonus, on !CONFIG_MEMCG_KMEM build some code can be compiled out. Link: https://lkml.kernel.org/r/20210319163821.20704-8-songmuchun@bytedance.comSigned-off-by:
Roman Gushchin <guro@fb.com> Reviewed-by:
-
Muchun Song authored
There is only one user of __memcg_kmem_charge(), so manually inline __memcg_kmem_charge() to obj_cgroup_charge_pages(). Similarly manually inline __memcg_kmem_uncharge() into obj_cgroup_uncharge_pages() and call obj_cgroup_uncharge_pages() in obj_cgroup_release(). This is just code cleanup without any functionality changes. Link: https://lkml.kernel.org/r/20210319163821.20704-7-songmuchun@bytedance.comSigned-off-by:
-
Muchun Song authored
Since Roman's series "The new cgroup slab memory controller" applied. All slab objects are charged via the new APIs of obj_cgroup. The new APIs introduce a struct obj_cgroup to charge slab objects. It prevents long-living objects from pinning the original memory cgroup in the memory. But there are still some corner objects (e.g. allocations larger than order-1 page on SLUB) which are not charged via the new APIs. Those objects (include the pages which are allocated from buddy allocator directly) are charged as kmem pages which still hold a reference to the memory cgroup. We want to reuse the obj_cgroup APIs to charge the kmem pages. If we do that, we should store an object cgroup pointer to page->memcg_data for the kmem pages. Finally, page->memcg_data will have 3 different meanings. 1) For the slab pages, page->memcg_data points to an object cgroups vector. 2) For the kmem pages (exclude the slab pages), page->memcg_data points to an object cgroup. 3) For the user pages (e.g. the LRU pages), page->memcg_data points to a memory cgroup. We do not change the behavior of page_memcg() and page_memcg_rcu(). They are also suitable for LRU pages and kmem pages. Why? Because memory allocations pinning memcgs for a long time - it exists at a larger scale and is causing recurring problems in the real world: page cache doesn't get reclaimed for a long time, or is used by the second, third, fourth, ... instance of the same job that was restarted into a new cgroup every time. Unreclaimable dying cgroups pile up, waste memory, and make page reclaim very inefficient. We can convert LRU pages and most other raw memcg pins to the objcg direction to fix this problem, and then the page->memcg will always point to an object cgroup pointer. At that time, LRU pages and kmem pages will be treated the same. The implementation of page_memcg() will remove the kmem page check. This patch aims to charge the kmem pages by using the new APIs of obj_cgroup. Finally, the page->memcg_data of the kmem page points to an object cgroup. We can use the __page_objcg() to get the object cgroup associated with a kmem page. Or we can use page_memcg() to get the memory cgroup associated with a kmem page, but caller must ensure that the returned memcg won't be released (e.g. acquire the rcu_read_lock or css_set_lock). Link: https://lkml.kernel.org/r/20210401030141.37061-1-songmuchun@bytedance.com Link: https://lkml.kernel.org/r/20210319163821.20704-6-songmuchun@bytedance.comSigned-off-by: -
Muchun Song authored
Just like assignment to ug->memcg, we only need to update ug->dummy_page if memcg changed. So move it to there. This is a very small optimization. Link: https://lkml.kernel.org/r/20210319163821.20704-5-songmuchun@bytedance.comSigned-off-by:
-
Muchun Song authored
page_memcg() is not suitable for use by page_expected_state() and page_bad_reason(). Because it can BUG_ON() for the slab pages when CONFIG_DEBUG_VM is enabled. As neither lru, nor kmem, nor slab page should have anything left in there by the time the page is freed, what we care about is whether the value of page->memcg_data is 0. So just directly access page->memcg_data here. Link: https://lkml.kernel.org/r/20210319163821.20704-4-songmuchun@bytedance.comSigned-off-by:
-
Muchun Song authored
We know that the unit of slab object charging is bytes, the unit of kmem page charging is PAGE_SIZE. If we want to reuse obj_cgroup APIs to charge the kmem pages, we should pass PAGE_SIZE (as third parameter) to obj_cgroup_charge(). Because the size is already PAGE_SIZE, we can skip touch the objcg stock. And obj_cgroup_{un}charge_pages() are introduced to charge in units of page level. In the latter patch, we also can reuse those two helpers to charge or uncharge a number of kernel pages to a object cgroup. This is just a code movement without any functional changes. Link: https://lkml.kernel.org/r/20210319163821.20704-3-songmuchun@bytedance.comSigned-off-by: -
Muchun Song authored
Patch series "Use obj_cgroup APIs to charge kmem pages", v5. Since Roman's series "The new cgroup slab memory controller" applied. All slab objects are charged with the new APIs of obj_cgroup. The new APIs introduce a struct obj_cgroup to charge slab objects. It prevents long-living objects from pinning the original memory cgroup in the memory. But there are still some corner objects (e.g. allocations larger than order-1 page on SLUB) which are not charged with the new APIs. Those objects (include the pages which are allocated from buddy allocator directly) are charged as kmem pages which still hold a reference to the memory cgroup. E.g. We know that the kernel stack is charged as kmem pages because the size of the kernel stack can be greater than 2 pages (e.g. 16KB on x86_64 or arm64). If we create a thread (suppose the thread stack is charged to memory cgroup A) and then move it from memory cgroup A to memory cgroup B. Because the kernel stack of the thread hold a reference to the memory cgroup A. The thread can pin the memory cgroup A in the memory even if we remove the cgroup A. If we want to see this scenario by using the following script. We can see that the system has added 500 dying cgroups (This is not a real world issue, just a script to show that the large kmallocs are charged as kmem pages which can pin the memory cgroup in the memory). #!/bin/bash cat /proc/cgroups | grep memory cd /sys/fs/cgroup/memory echo 1 > memory.move_charge_at_immigrate for i in range{1..500} do mkdir kmem_test echo $$ > kmem_test/cgroup.procs sleep 3600 & echo $$ > cgroup.procs echo `cat kmem_test/cgroup.procs` > cgroup.procs rmdir kmem_test done cat /proc/cgroups | grep memory This patchset aims to make those kmem pages to drop the reference to memory cgroup by using the APIs of obj_cgroup. Finally, we can see that the number of the dying cgroups will not increase if we run the above test script. This patch (of 7): The rcu_read_lock/unlock only can guarantee that the memcg will not be freed, but it cannot guarantee the success of css_get (which is in the refill_stock when cached memcg changed) to memcg. rcu_read_lock() memcg = obj_cgroup_memcg(old) __memcg_kmem_uncharge(memcg) refill_stock(memcg) if (stock->cached != memcg) // css_get can change the ref counter from 0 back to 1. css_get(&memcg->css) rcu_read_unlock() This fix is very like the commit: eefbfa7f ("mm: memcg/slab: fix use after free in obj_cgroup_charge") Fix this by holding a reference to the memcg which is passed to the __memcg_kmem_uncharge() before calling __memcg_kmem_uncharge(). Link: https://lkml.kernel.org/r/20210319163821.20704-1-songmuchun@bytedance.com Link: https://lkml.kernel.org/r/20210319163821.20704-2-songmuchun@bytedance.com Fixes: 3de7d4f2 ("mm: memcg/slab: optimize objcg stock draining") Signed-off-by: -
Shakeel Butt authored
Currently the kernel adds the page, allocated for swapin, to the swapcache before charging the page. This is fine but now we want a per-memcg swapcache stat which is essential for folks who wants to transparently migrate from cgroup v1's memsw to cgroup v2's memory and swap counters. In addition charging a page before exposing it to other parts of the kernel is a step in the right direction. To correctly maintain the per-memcg swapcache stat, this patch has adopted to charge the page before adding it to swapcache. One challenge in this option is the failure case of add_to_swap_cache() on which we need to undo the mem_cgroup_charge(). Specifically undoing mem_cgroup_uncharge_swap() is not simple. To resolve the issue, this patch decouples the charging for swapin pages from mem_cgroup_charge(). Two new functions are introduced, mem_cgroup_swapin_charge_page() for just charging the swapin page and mem_cgroup_swapin_uncharge_swap() for uncharging the swap slot once the page has been successfully added to the swapcache. [shakeelb@google.com: set page->private before calling swap_readpage] Link: https://lkml.kernel.org/r/20210318015959.2986837-1-shakeelb@google.com Link: https://lkml.kernel.org/r/20210305212639.775498-1-shakeelb@google.comSigned-off-by:
Heiko Carstens <hca@linux.ibm.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by:
-
Johannes Weiner authored
With memcg having switched to rstat, memory.stat output is precise. Update the cgroup selftest to reflect the expectations and error tolerances of the new implementation. Also add newly tracked types of memory to the memory.stat side of the equation, since they're included in memory.current and could throw false positives. Link: https://lkml.kernel.org/r/20210209163304.77088-9-hannes@cmpxchg.orgSigned-off-by:
Michal Koutný <mkoutny@suse.com> Acked-by:
-
Johannes Weiner authored
There are two functions to flush the per-cpu data of an lruvec into the rest of the cgroup tree: when the cgroup is being freed, and when a CPU disappears during hotplug. The difference is whether all CPUs or just one is being collected, but the rest of the flushing code is the same. Merge them into one function and share the common code. Link: https://lkml.kernel.org/r/20210209163304.77088-8-hannes@cmpxchg.orgSigned-off-by:
-
Johannes Weiner authored
Replace the memory controller's custom hierarchical stats code with the generic rstat infrastructure provided by the cgroup core. The current implementation does batched upward propagation from the write side (i.e. as stats change). The per-cpu batches introduce an error, which is multiplied by the number of subgroups in a tree. In systems with many CPUs and sizable cgroup trees, the error can be large enough to confuse users (e.g. 32 batch pages * 32 CPUs * 32 subgroups results in an error of up to 128M per stat item). This can entirely swallow allocation bursts inside a workload that the user is expecting to see reflected in the statistics. In the past, we've done read-side aggregation, where a memory.stat read would have to walk the entire subtree and add up per-cpu counts. This became problematic with lazily-freed cgroups: we could have large subtrees where most cgroups were entirely idle. Hence the switch to change-driven upward propagation. Unfortunately, it needed to trade accuracy for speed due to the write side being so hot. Rstat combines the best of both worlds: from the write side, it cheaply maintains a queue of cgroups that have pending changes, so that the read side can do selective tree aggregation. This way the reported stats will always be precise and recent as can be, while the aggregation can skip over potentially large numbers of idle cgroups. The way rstat works is that it implements a tree for tracking cgroups with pending local changes, as well as a flush function that walks the tree upwards. The controller then drives this by 1) telling rstat when a local cgroup stat changes (e.g. mod_memcg_state) and 2) when a flush is required to get uptodate hierarchy stats for a given subtree (e.g. when memory.stat is read). The controller also provides a flush callback that is called during the rstat flush walk for each cgroup and aggregates its local per-cpu counters and propagates them upwards. This adds a second vmstats to struct mem_cgroup (MEMCG_NR_STAT + NR_VM_EVENT_ITEMS) to track pending subtree deltas during upward aggregation. It removes 3 words from the per-cpu data. It eliminates memcg_exact_page_state(), since memcg_page_state() is now exact. [akpm@linux-foundation.org: merge fix] [hannes@cmpxchg.org: fix a sleep in atomic section problem] Link: https://lkml.kernel.org/r/20210315234100.64307-1-hannes@cmpxchg.org Link: https://lkml.kernel.org/r/20210209163304.77088-7-hannes@cmpxchg.orgSigned-off-by:
Balbir Singh <bsingharora@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by:
-
Johannes Weiner authored
Current users of the rstat code can source root-level statistics from the native counters of their respective subsystem, allowing them to forego aggregation at the root level. This optimization is currently implemented inside the generic rstat code, which doesn't track the root cgroup and doesn't invoke the subsystem flush callbacks on it. However, the memory controller cannot do this optimization, because cgroup1 breaks out memory specifically for the local level, including at the root level. In preparation for the memory controller switching to rstat, move the optimization from rstat core to the controllers. Afterwards, rstat will always track the root cgroup for changes and invoke the subsystem callbacks on it; and it's up to the subsystem to special-case and skip aggregation of the root cgroup if it can source this information through other, cheaper means. This is the case for the io controller and the cgroup base stats. In their respective flush callbacks, check whether the parent is the root cgroup, and if so, skip the unnecessary upward propagation. The extra cost of tracking the root cgroup is negligible: on stat changes, we actually remove a branch that checks for the root. The queueing for a flush touches only per-cpu data, and only the first stat change since a flush requires a (per-cpu) lock. Link: https://lkml.kernel.org/r/20210209163304.77088-6-hannes@cmpxchg.orgSigned-off-by:
Tejun Heo <tj@kernel.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Michal Koutný <mkoutny@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by:
-
Johannes Weiner authored
Rstat currently only supports the default hierarchy in cgroup2. In order to replace memcg's private stats infrastructure - used in both cgroup1 and cgroup2 - with rstat, the latter needs to support cgroup1. The initialization and destruction callbacks for regular cgroups are already in place. Remove the cgroup_on_dfl() guards to handle cgroup1. The initialization of the root cgroup is currently hardcoded to only handle cgrp_dfl_root.cgrp. Move those callbacks to cgroup_setup_root() and cgroup_destroy_root() to handle the default root as well as the various cgroup1 roots we may set up during mounting. The linking of css to cgroups happens in code shared between cgroup1 and cgroup2 as well. Simply remove the cgroup_on_dfl() guard. Linkage of the root css to the root cgroup is a bit trickier: per default, the root css of a subsystem controller belongs to the default hierarchy (i.e. the cgroup2 root). When a controller is mounted in its cgroup1 version, the root css is stolen and moved to the cgroup1 root; on unmount, the css moves back to the default hierarchy. Annotate rebind_subsystems() to move the root css linkage along between roots. Link: https://lkml.kernel.org/r/20210209163304.77088-5-hannes@cmpxchg.orgSigned-off-by:
-
Johannes Weiner authored
There are no users outside of the memory controller itself. The rest of the kernel cares either about node or lruvec stats. Link: https://lkml.kernel.org/r/20210209163304.77088-4-hannes@cmpxchg.orgSigned-off-by:
-
Johannes Weiner authored
No need to encapsulate a simple struct member access. Link: https://lkml.kernel.org/r/20210209163304.77088-3-hannes@cmpxchg.orgSigned-off-by:
-
Johannes Weiner authored
Patch series "mm: memcontrol: switch to rstat", v3. This series converts memcg stats tracking to the streamlined rstat infrastructure provided by the cgroup core code. rstat is already used by the CPU controller and the IO controller. This change is motivated by recent accuracy problems in memcg's custom stats code, as well as the benefits of sharing common infra with other controllers. The current memcg implementation does batched tree aggregation on the write side: local stat changes are cached in per-cpu counters, which are then propagated upward in batches when a threshold (32 pages) is exceeded. This is cheap, but the error introduced by the lazy upward propagation adds up: 32 pages times CPUs times cgroups in the subtree. We've had complaints from service owners that the stats do not reliably track and react to allocation behavior as expected, sometimes swallowing the results of entire test applications. The original memcg stat implementation used to do tree aggregation exclusively on the read side: local stats would only ever be tracked in per-cpu counters, and a memory.stat read would iterate the entire subtree and sum those counters up. This didn't keep up with the times: - Cgroup trees are much bigger now. We switched to lazily-freed cgroups, where deleted groups would hang around until their remaining page cache has been reclaimed. This can result in large subtrees that are expensive to walk, while most of the groups are idle and their statistics don't change much anymore. - Automated monitoring increased. With the proliferation of userspace oom killing, proactive reclaim, and higher-resolution logging of workload trends in general, top-level stat files are polled at least once a second in many deployments. - The lifetime of cgroups got shorter. Where most cgroup setups in the past would have a few large policy-oriented cgroups for everything running on the system, newer cgroup deployments tend to create one group per application - which gets deleted again as the processes exit. An aggregation scheme that doesn't retain child data inside the parents loses event history of the subtree. Rstat addresses all three of those concerns through intelligent, persistent read-side aggregation. As statistics change at the local level, rstat tracks - on a per-cpu basis - only those parts of a subtree that have changes pending and require aggregation. The actual aggregation occurs on the colder read side - which can now skip over (potentially large) numbers of recently idle cgroups. === The test_kmem cgroup selftest is currently failing due to excessive cumulative vmstat drift from 100 subgroups: ok 1 test_kmem_basic memory.current = 8810496 slab + anon + file + kernel_stack = 17074568 slab = 6101384 anon = 946176 file = 0 kernel_stack = 10027008 not ok 2 test_kmem_memcg_deletion ok 3 test_kmem_proc_kpagecgroup ok 4 test_kmem_kernel_stacks ok 5 test_kmem_dead_cgroups ok 6 test_percpu_basic As you can see, memory.stat items far exceed memory.current. The kernel stack alone is bigger than all of charged memory. That's because the memory of the test has been uncharged from memory.current, but the negative vmstat deltas are still sitting in the percpu caches. The test at this time isn't even counting percpu, pagetables etc. yet, which would further contribute to the error. The last patch in the series updates the test to include them - as well as reduces the vmstat tolerances in general to only expect page_counter batching. With all patches applied, the (now more stringent) test succeeds: ok 1 test_kmem_basic ok 2 test_kmem_memcg_deletion ok 3 test_kmem_proc_kpagecgroup ok 4 test_kmem_kernel_stacks ok 5 test_kmem_dead_cgroups ok 6 test_percpu_basic === A kernel build test confirms that overhead is comparable. Two kernels are built simultaneously in a nested tree with several idle siblings: root - kernelbuild - one - two - three - four - build-a (defconfig, make -j16) `- build-b (defconfig, make -j16) `- idle-1 `- ... `- idle-9 During the builds, kernelbuild/memory.stat is read once a second. A perf diff shows that the changes in cycle distribution is minimal. Top 10 kernel symbols: 0.09% +0.08% [kernel.kallsyms] [k] __mod_memcg_lruvec_state 0.00% +0.06% [kernel.kallsyms] [k] cgroup_rstat_updated 0.08% -0.05% [kernel.kallsyms] [k] __mod_memcg_state.part.0 0.16% -0.04% [kernel.kallsyms] [k] release_pages 0.00% +0.03% [kernel.kallsyms] [k] __count_memcg_events 0.01% +0.03% [kernel.kallsyms] [k] mem_cgroup_charge_statistics.constprop.0 0.10% -0.02% [kernel.kallsyms] [k] get_mem_cgroup_from_mm 0.05% -0.02% [kernel.kallsyms] [k] mem_cgroup_update_lru_size 0.57% +0.01% [kernel.kallsyms] [k] asm_exc_page_fault === The on-demand aggregated stats are now fully accurate: $ grep -e nr_inactive_file /proc/vmstat | awk '{print($1,$2*4096)}'; \ grep -e inactive_file /sys/fs/cgroup/memory.stat vanilla: patched: nr_inactive_file 1574105088 nr_inactive_file 1027801088 inactive_file 1577410560 inactive_file 1027801088 === This patch (of 8): The memcg hotunplug callback erroneously flushes counts on the local CPU, not the counts of the CPU going away; those counts will be lost. Flush the CPU that is actually going away. Also simplify the code a bit by using mod_memcg_state() and count_memcg_events() instead of open-coding the upward flush - this is comparable to how vmstat.c handles hotunplug flushing. Link: https://lkml.kernel.org/r/20210209163304.77088-1-hannes@cmpxchg.org Link: https://lkml.kernel.org/r/20210209163304.77088-2-hannes@cmpxchg.org Fixes: a983b5eb ("mm: memcontrol: fix excessive complexity in memory.stat reporting") Signed-off-by:
-
