- 26 Apr, 2024 40 commits
-
-
Ryan Roberts authored
Rework madvise_cold_or_pageout_pte_range() to avoid splitting any large folio that is fully and contiguously mapped in the pageout/cold vm range. This change means that large folios will be maintained all the way to swap storage. This both improves performance during swap-out, by eliding the cost of splitting the folio, and sets us up nicely for maintaining the large folio when it is swapped back in (to be covered in a separate series). Folios that are not fully mapped in the target range are still split, but note that behavior is changed so that if the split fails for any reason (folio locked, shared, etc) we now leave it as is and move to the next pte in the range and continue work on the proceeding folios. Previously any failure of this sort would cause the entire operation to give up and no folios mapped at higher addresses were paged out or made cold. Given large folios are becoming more common, this old behavior would have likely lead to wasted opportunities. While we are at it, change the code that clears young from the ptes to use ptep_test_and_clear_young(), via the new mkold_ptes() batch helper function. This is more efficent than get_and_clear/modify/set, especially for contpte mappings on arm64, where the old approach would require unfolding/refolding and the new approach can be done in place. Link: https://lkml.kernel.org/r/20240408183946.2991168-8-ryan.roberts@arm.comSigned-off-by:
Ryan Roberts <ryan.roberts@arm.com> Reviewed-by:
Barry Song <v-songbaohua@oppo.com> Acked-by:
David Hildenbrand <david@redhat.com> Cc: Barry Song <21cnbao@gmail.com> Cc: Chris Li <chrisl@kernel.org> Cc: Gao Xiang <xiang@kernel.org> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Kefeng Wang <wangkefeng.wang@huawei.com> Cc: Lance Yang <ioworker0@gmail.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Yang Shi <shy828301@gmail.com> Cc: Yu Zhao <yuzhao@google.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org>
-
Ryan Roberts authored
Now that swap supports storing all mTHP sizes, avoid splitting large folios before swap-out. This benefits performance of the swap-out path by eliding split_folio_to_list(), which is expensive, and also sets us up for swapping in large folios in a future series. If the folio is partially mapped, we continue to split it since we want to avoid the extra IO overhead and storage of writing out pages uneccessarily. THP_SWPOUT and THP_SWPOUT_FALLBACK counters should continue to count events only for PMD-mappable folios to avoid user confusion. THP_SWPOUT already has the appropriate guard. Add a guard for THP_SWPOUT_FALLBACK. It may be appropriate to add per-size counters in future. Link: https://lkml.kernel.org/r/20240408183946.2991168-7-ryan.roberts@arm.comSigned-off-by:
-
Ryan Roberts authored
Multi-size THP enables performance improvements by allocating large, pte-mapped folios for anonymous memory. However I've observed that on an arm64 system running a parallel workload (e.g. kernel compilation) across many cores, under high memory pressure, the speed regresses. This is due to bottlenecking on the increased number of TLBIs added due to all the extra folio splitting when the large folios are swapped out. Therefore, solve this regression by adding support for swapping out mTHP without needing to split the folio, just like is already done for PMD-sized THP. This change only applies when CONFIG_THP_SWAP is enabled, and when the swap backing store is a non-rotating block device. These are the same constraints as for the existing PMD-sized THP swap-out support. Note that no attempt is made to swap-in (m)THP here - this is still done page-by-page, like for PMD-sized THP. But swapping-out mTHP is a prerequisite for swapping-in mTHP. The main change here is to improve the swap entry allocator so that it can allocate any power-of-2 number of contiguous entries between [1, (1 << PMD_ORDER)]. This is done by allocating a cluster for each distinct order and allocating sequentially from it until the cluster is full. This ensures that we don't need to search the map and we get no fragmentation due to alignment padding for different orders in the cluster. If there is no current cluster for a given order, we attempt to allocate a free cluster from the list. If there are no free clusters, we fail the allocation and the caller can fall back to splitting the folio and allocates individual entries (as per existing PMD-sized THP fallback). The per-order current clusters are maintained per-cpu using the existing infrastructure. This is done to avoid interleving pages from different tasks, which would prevent IO being batched. This is already done for the order-0 allocations so we follow the same pattern. As is done for order-0 per-cpu clusters, the scanner now can steal order-0 entries from any per-cpu-per-order reserved cluster. This ensures that when the swap file is getting full, space doesn't get tied up in the per-cpu reserves. This change only modifies swap to be able to accept any order mTHP. It doesn't change the callers to elide doing the actual split. That will be done in separate changes. Link: https://lkml.kernel.org/r/20240408183946.2991168-6-ryan.roberts@arm.comSigned-off-by:
"Huang, Ying" <ying.huang@intel.com> Cc: Barry Song <21cnbao@gmail.com> Cc: Barry Song <v-songbaohua@oppo.com> Cc: Chris Li <chrisl@kernel.org> Cc: David Hildenbrand <david@redhat.com> Cc: Gao Xiang <xiang@kernel.org> Cc: Kefeng Wang <wangkefeng.wang@huawei.com> Cc: Lance Yang <ioworker0@gmail.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Yang Shi <shy828301@gmail.com> Cc: Yu Zhao <yuzhao@google.com> Signed-off-by:
-
Ryan Roberts authored
We are about to allow swap storage of any mTHP size. To prepare for that, let's change get_swap_pages() to take a folio order parameter instead of nr_pages. This makes the interface self-documenting; a power-of-2 number of pages must be provided. We will also need the order internally so this simplifies accessing it. Link: https://lkml.kernel.org/r/20240408183946.2991168-5-ryan.roberts@arm.comSigned-off-by:
-
Ryan Roberts authored
struct percpu_cluster stores the index of cpu's current cluster and the offset of the next entry that will be allocated for the cpu. These two pieces of information are redundant because the cluster index is just (offset / SWAPFILE_CLUSTER). The only reason for explicitly keeping the cluster index is because the structure used for it also has a flag to indicate "no cluster". However this data structure also contains a spin lock, which is never used in this context, as a side effect the code copies the spinlock_t structure, which is questionable coding practice in my view. So let's clean this up and store only the next offset, and use a sentinal value (SWAP_NEXT_INVALID) to indicate "no cluster". SWAP_NEXT_INVALID is chosen to be 0, because 0 will never be seen legitimately; The first page in the swap file is the swap header, which is always marked bad to prevent it from being allocated as an entry. This also prevents the cluster to which it belongs being marked free, so it will never appear on the free list. This change saves 16 bytes per cpu. And given we are shortly going to extend this mechanism to be per-cpu-AND-per-order, we will end up saving 16 * 9 = 144 bytes per cpu, which adds up if you have 256 cpus in the system. Link: https://lkml.kernel.org/r/20240408183946.2991168-4-ryan.roberts@arm.comSigned-off-by:
-
Ryan Roberts authored
Now that we no longer have a convenient flag in the cluster to determine if a folio is large, free_swap_and_cache() will take a reference and lock a large folio much more often, which could lead to contention and (e.g.) failure to split large folios, etc. Let's solve that problem by batch freeing swap and cache with a new function, free_swap_and_cache_nr(), to free a contiguous range of swap entries together. This allows us to first drop a reference to each swap slot before we try to release the cache folio. This means we only try to release the folio once, only taking the reference and lock once - much better than the previous 512 times for the 2M THP case. Contiguous swap entries are gathered in zap_pte_range() and madvise_free_pte_range() in a similar way to how present ptes are already gathered in zap_pte_range(). While we are at it, let's simplify by converting the return type of both functions to void. The return value was used only by zap_pte_range() to print a bad pte, and was ignored by everyone else, so the extra reporting wasn't exactly guaranteed. We will still get the warning with most of the information from get_swap_device(). With the batch version, we wouldn't know which pte was bad anyway so could print the wrong one. [ryan.roberts@arm.com: fix a build warning on parisc] Link: https://lkml.kernel.org/r/20240409111840.3173122-1-ryan.roberts@arm.com Link: https://lkml.kernel.org/r/20240408183946.2991168-3-ryan.roberts@arm.comSigned-off-by:
-
Ryan Roberts authored
Patch series "Swap-out mTHP without splitting", v7. This series adds support for swapping out multi-size THP (mTHP) without needing to first split the large folio via split_huge_page_to_list_to_order(). It closely follows the approach already used to swap-out PMD-sized THP. There are a couple of reasons for swapping out mTHP without splitting: - Performance: It is expensive to split a large folio and under extreme memory pressure some workloads regressed performance when using 64K mTHP vs 4K small folios because of this extra cost in the swap-out path. This series not only eliminates the regression but makes it faster to swap out 64K mTHP vs 4K small folios. - Memory fragmentation avoidance: If we can avoid splitting a large folio memory is less likely to become fragmented, making it easier to re-allocate a large folio in future. - Performance: Enables a separate series [7] to swap-in whole mTHPs, which means we won't lose the TLB-efficiency benefits of mTHP once the memory has been through a swap cycle. I've done what I thought was the smallest change possible, and as a result, this approach is only employed when the swap is backed by a non-rotating block device (just as PMD-sized THP is supported today). Discussion against the RFC concluded that this is sufficient. Performance Testing =================== I've run some swap performance tests on Ampere Altra VM (arm64) with 8 CPUs. The VM is set up with a 35G block ram device as the swap device and the test is run from inside a memcg limited to 40G memory. I've then run `usemem` from vm-scalability with 70 processes, each allocating and writing 1G of memory. I've repeated everything 6 times and taken the mean performance improvement relative to 4K page baseline: | alloc size | baseline | + this series | | | mm-unstable (~v6.9-rc1) | | |:-----------|------------------------:|------------------------:| | 4K Page | 0.0% | 1.3% | | 64K THP | -13.6% | 46.3% | | 2M THP | 91.4% | 89.6% | So with this change, the 64K swap performance goes from a 14% regression to a 46% improvement. While 2M shows a small regression I'm confident that this is just noise. [1] https://lore.kernel.org/linux-mm/20231010142111.3997780-1-ryan.roberts@arm.com/ [2] https://lore.kernel.org/linux-mm/20231017161302.2518826-1-ryan.roberts@arm.com/ [3] https://lore.kernel.org/linux-mm/20231025144546.577640-1-ryan.roberts@arm.com/ [4] https://lore.kernel.org/linux-mm/20240311150058.1122862-1-ryan.roberts@arm.com/ [5] https://lore.kernel.org/linux-mm/20240327144537.4165578-1-ryan.roberts@arm.com/ [6] https://lore.kernel.org/linux-mm/20240403114032.1162100-1-ryan.roberts@arm.com/ [7] https://lore.kernel.org/linux-mm/20240304081348.197341-1-21cnbao@gmail.com/ [8] https://lore.kernel.org/linux-mm/CAGsJ_4yMOow27WDvN2q=E4HAtDd2PJ=OQ5Pj9DG+6FLWwNuXUw@mail.gmail.com/ [9] https://lore.kernel.org/linux-mm/579d5127-c763-4001-9625-4563a9316ac3@redhat.com/ This patch (of 7): As preparation for supporting small-sized THP in the swap-out path, without first needing to split to order-0, Remove the CLUSTER_FLAG_HUGE, which, when present, always implies PMD-sized THP, which is the same as the cluster size. The only use of the flag was to determine whether a swap entry refers to a single page or a PMD-sized THP in swap_page_trans_huge_swapped(). Instead of relying on the flag, we now pass in order, which originates from the folio's order. This allows the logic to work for folios of any order. The one snag is that one of the swap_page_trans_huge_swapped() call sites does not have the folio. But it was only being called there to shortcut a call __try_to_reclaim_swap() in some cases. __try_to_reclaim_swap() gets the folio and (via some other functions) calls swap_page_trans_huge_swapped(). So I've removed the problematic call site and believe the new logic should be functionally equivalent. That said, removing the fast path means that we will take a reference and trylock a large folio much more often, which we would like to avoid. The next patch will solve this. Removing CLUSTER_FLAG_HUGE also means we can remove split_swap_cluster() which used to be called during folio splitting, since split_swap_cluster()'s only job was to remove the flag. Link: https://lkml.kernel.org/r/20240408183946.2991168-1-ryan.roberts@arm.com Link: https://lkml.kernel.org/r/20240408183946.2991168-2-ryan.roberts@arm.comSigned-off-by:Chris Li <chrisl@kernel.org> Acked-by:
-
Baolin Wang authored
Commit 44042b44 ("mm/page_alloc: allow high-order pages to be stored on the per-cpu lists") extends the PCP allocator to store THP pages, and it determines whether to cache THP pages in PCP by comparing with pageblock_order. But the pageblock_order is not always equal to THP order. It might also be MAX_PAGE_ORDER, which could prevent PCP from caching THP pages. Therefore, using HPAGE_PMD_ORDER instead to determine the need for caching THP for PCP will fix this issue Link: https://lkml.kernel.org/r/a25c9e14cd03907d5978b60546a69e6aa3fc2a7d.1712151833.git.baolin.wang@linux.alibaba.com Fixes: 44042b44 ("mm/page_alloc: allow high-order pages to be stored on the per-cpu lists") Signed-off-by:
Baolin Wang <baolin.wang@linux.alibaba.com> Acked-by:
Vlastimil Babka <vbabka@suse.cz> Cc: Mel Gorman <mgorman@techsingularity.net> Reviewed-by:
Barry Song <baohua@kernel.org> Signed-off-by:
-
Matthew Wilcox (Oracle) authored
Replace two calls to compound_head() with one. Link: https://lkml.kernel.org/r/20240403171456.1445117-5-willy@infradead.orgSigned-off-by:
Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Christian Brauner <brauner@kernel.org> Signed-off-by:
-
Matthew Wilcox (Oracle) authored
Both callers already have a folio; pass it in instead of doing the conversion each time. Link: https://lkml.kernel.org/r/20240403171456.1445117-4-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Replaces three calls to compound_head() with one. Shrinks the function from 2614 bytes to 1112 bytes in an allmodconfig build. Link: https://lkml.kernel.org/r/20240403171456.1445117-3-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Patch series "Use folio APIs in procfs". We're down to very few users of the PageFoo macros, with proc being a major user. After this patchset and another patchset I have for khugepaged, we can get rid of PageActive, PageReadahead and PageSwapBacked. This patchset has the usual advantages in its own right of removing hidden calls to compound_head(). We have the page table lock, so the mapcount & refcount are stable and there can't be any races with folios suddenly becoming tail pages. This patch (of 4): Replaces six calls to compound_head() with one. Shrinks the function from 5054 bytes to 1756 bytes in an allmodconfig build. Link: https://lkml.kernel.org/r/20240403171456.1445117-1-willy@infradead.org Link: https://lkml.kernel.org/r/20240403171456.1445117-2-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
If CONFIG_PAGE_IDLE_FLAG is not set, we can use FOLIO_FLAG_FALSE() to generate these definitions. Link: https://lkml.kernel.org/r/20240402201252.917342-5-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
All users have now been converted to the folio equivalents, so remove the page wrappers. Link: https://lkml.kernel.org/r/20240402201252.917342-4-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Replace seven calls to compound_head() with one. Link: https://lkml.kernel.org/r/20240402201252.917342-3-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Patch series "Remove page_idle and page_young wrappers". There are only a couple of places left using the page wrappers for idle & young tracking. Convert the two users in proc and then we can remove the wrappers. That enables the further simplification of autogenerating the definitions when CONFIG_PAGE_IDLE_FLAG is disabled. This patch (of 4): Replaces four calls to compound_head() with two. Link: https://lkml.kernel.org/r/20240402201252.917342-1-willy@infradead.org Link: https://lkml.kernel.org/r/20240402201252.917342-2-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Replace the use of pages with folios. Saves a few calls to compound_head() and removes some uses of obsolete functions. Link: https://lkml.kernel.org/r/20240403171838.1445826-8-willy@infradead.orgSigned-off-by:
Vishal Moola (Oracle) <vishal.moola@gmail.com> Signed-off-by:
-
Matthew Wilcox (Oracle) authored
Pull folios from the page cache instead of pages. Half of this work had been done already, but we were still operating on pages for a large chunk of this function. There is no attempt in this patch to handle large folios that are smaller than a THP; that will have to wait for a future patch. [willy@infradead.org: the unlikely() is embedded in IS_ERR()] Link: https://lkml.kernel.org/r/ZhIWX8K0E2tSyMSr@casper.infradead.org Link: https://lkml.kernel.org/r/20240403171838.1445826-7-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Use new_folio throughout where we had been using hpage. Link: https://lkml.kernel.org/r/20240403171838.1445826-6-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Simplify the body of __collapse_huge_page_copy() while I'm looking at it. Link: https://lkml.kernel.org/r/20240403171838.1445826-5-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Work purely in terms of the folio. Removes a call to compound_head() in put_page(). Link: https://lkml.kernel.org/r/20240403171838.1445826-4-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Both callers want to deal with a folio, so return a folio from this function. Link: https://lkml.kernel.org/r/20240403171838.1445826-3-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Patch series "khugepaged folio conversions". We've been kind of hacking piecemeal at converting khugepaged to use folios instead of compound pages, and so this patchset is a little larger than it should be as I undo some of our wrong moves in the past. In particular, collapse_file() now consistently uses 'new_folio' for the freshly allocated folio and 'folio' for the one that's currently in use. This patch (of 7): This function has one caller, and the combined function is simpler to read, reason about and modify. Link: https://lkml.kernel.org/r/20240403171838.1445826-1-willy@infradead.org Link: https://lkml.kernel.org/r/20240403171838.1445826-2-willy@infradead.orgSigned-off-by:
-
Dev Jain authored
Enforce consistency across files by avoiding two separate functions to parse /proc/self/maps, replacing them with a simple sscanf(). Link: https://lkml.kernel.org/r/20240330173557.2697684-4-dev.jain@arm.comSigned-off-by:
Dev Jain <dev.jain@arm.com> Cc: Anshuman Khandual <anshuman.khandual@arm.com> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Kalesh Singh <kaleshsingh@google.com> Cc: Shuah Khan <shuah@kernel.org> Signed-off-by:
-
Dev Jain authored
Mismatch index is currently being checked by a brute force iteration over the buffer. Instead, break the comparison into O(sqrt(n)) number of chunks, with the chunk size of this order only, where n is the size of the buffer. Do a brute-force iteration to print to stdout only when the highly optimized memcmp() library function returns a mismatch in the chunk. The time complexity of this algorithm is O(sqrt(n)) * t, where t is the time taken by memcmp(); for our test conditions, it is safe to assume t to be small. Link: https://lkml.kernel.org/r/20240330173557.2697684-3-dev.jain@arm.comSigned-off-by:
-
Dev Jain authored
Patch series "selftests/mm: mremap_test: Optimizations and style fixes". The mremap_test, in a worst case controlled by the -t flag, does a for loop iteration in orders of GB. Without compromising on the stdout report, the aim is to reduce this time. A pre-filled random buffer is allocated based on the seed, replacing repetitive rand() calls. The byte pattern in the memory locations is set through memcpy() from the random buffer. Replacing the loop for printing the mismatch index to stdout, employ an efficient algorithm by breaking the comparison into chunks, use the highly optimized memcmp() library function, and when a mismatch does occur, only then do a brute force iteration. Also, use sscanf() to parse /proc/self/maps for consistency across files. Execution time results (x86 system): ./mremap_test Original: 3 seconds After change: 0.8 seconds ./mremap_test -t100 Original: 17 seconds After change: 2 seconds ./mremap_test -t0 (worst case): Original: 9:40 minutes After change: 45 seconds This patch (of 3): Allocate a pre-filled random buffer using the seed. Replace iterative copying of the random sequence to buffers using the highly optimized library function memcpy(). Link: https://lkml.kernel.org/r/20240330173557.2697684-1-dev.jain@arm.com Link: https://lkml.kernel.org/r/20240330173557.2697684-2-dev.jain@arm.comSigned-off-by:
-
Joel Granados authored
This commit comes at the tail end of a greater effort to remove the empty elements at the end of the ctl_table arrays (sentinels) which will reduce the overall build time size of the kernel and run time memory bloat by ~64 bytes per sentinel (further information Link : https://lore.kernel.org/all/ZO5Yx5JFogGi%2FcBo@bombadil.infradead.org/) Remove sentinel from all files under mm/ that register a sysctl table. Link: https://lkml.kernel.org/r/20240328-jag-sysctl_remset_misc-v1-1-47c1463b3af2@samsung.comSigned-off-by:
Joel Granados <j.granados@samsung.com> Reviewed-by:
Muchun Song <muchun.song@linux.dev> Reviewed-by:
Miaohe Lin <linmiaohe@huawei.com> Signed-off-by:
-
Matthew Wilcox (Oracle) authored
With all callers converted, we can use the nice shorter name. Take this opportunity to reorder the arguments to the logical order (larger object first). Link: https://lkml.kernel.org/r/20240328225831.1765286-4-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Convert the three remaining callers to call vma_pgoff_address() directly. This removes an ambiguity where we'd check just one page if passed a tail page and all N pages if passed a head page. Also add better kernel-doc for vma_pgoff_address(). Link: https://lkml.kernel.org/r/20240328225831.1765286-3-willy@infradead.orgSigned-off-by:
-
Matthew Wilcox (Oracle) authored
Patch series "Unify vma_address and vma_pgoff_address". The current vma_address() pretends that the ambiguity between head & tail page is an advantage. If you pass a head page to vma_address(), it will operate on all pages in the folio, while if you pass a tail page, it will operate on a single page. That's not what any of the callers actually want, so first convert all callers to use vma_pgoff_address() and then rename vma_pgoff_address() to vma_address(). This patch (of 3): If 'page' is the first page of a large folio then vma_address() will scan for any page in the entire folio. This can lead to page_mapped_in_vma() returning true if some of the tail pages are mapped and the head page is not. This could lead to memory failure choosing to kill a task unnecessarily. Link: https://lkml.kernel.org/r/20240328225831.1765286-1-willy@infradead.org Link: https://lkml.kernel.org/r/20240328225831.1765286-2-willy@infradead.orgSigned-off-by:
-
Baolin Wang authored
Now the mTHP can also be split or added into the deferred list, so add folio_test_pmd_mappable() validation for PMD mapped THP, to avoid confusion with PMD mapped THP related statistics. [baolin.wang@linux.alibaba.com: check THP earlier in case folio is split, per Lance] Link: https://lkml.kernel.org/r/b99f8cb14bc85fdb6ab43721d1331cb5ebed2581.1713771041.git.baolin.wang@linux.alibaba.com Link: https://lkml.kernel.org/r/a5341defeef27c9ac7b85c97f030f93e4368bbc1.1711694852.git.baolin.wang@linux.alibaba.comSigned-off-by:
Lance Yang <ioworker0@gmail.com> Cc: Muchun Song <muchun.song@linux.dev> Signed-off-by:
-
Baolin Wang authored
Now the anonymous page allocation already supports multi-size THP (mTHP), but the numa balancing still prohibits mTHP migration even though it is an exclusive mapping, which is unreasonable. Allow scanning mTHP: Commit 859d4adc ("mm: numa: do not trap faults on shared data section pages") skips shared CoW pages' NUMA page migration to avoid shared data segment migration. In addition, commit 80d47f5d ("mm: don't try to NUMA-migrate COW pages that have other uses") change to use page_count() to avoid GUP pages migration, that will also skip the mTHP numa scanning. Theoretically, we can use folio_maybe_dma_pinned() to detect the GUP issue, although there is still a GUP race, the issue seems to have been resolved by commit 80d47f5d. Meanwhile, use the folio_likely_mapped_shared() to skip shared CoW pages though this is not a precise sharers count. To check if the folio is shared, ideally we want to make sure every page is mapped to the same process, but doing that seems expensive and using the estimated mapcount seems can work when running autonuma benchmark. Allow migrating mTHP: As mentioned in the previous thread[1], large folios (including THP) are more susceptible to false sharing issues among threads than 4K base page, leading to pages ping-pong back and forth during numa balancing, which is currently not easy to resolve. Therefore, as a start to support mTHP numa balancing, we can follow the PMD mapped THP's strategy, that means we can reuse the 2-stage filter in should_numa_migrate_memory() to check if the mTHP is being heavily contended among threads (through checking the CPU id and pid of the last access) to avoid false sharing at some degree. Thus, we can restore all PTE maps upon the first hint page fault of a large folio to follow the PMD mapped THP's strategy. In the future, we can continue to optimize the NUMA balancing algorithm to avoid the false sharing issue with large folios as much as possible. Performance data: Machine environment: 2 nodes, 128 cores Intel(R) Xeon(R) Platinum Base: 2024-03-25 mm-unstable branch Enable mTHP to run autonuma-benchmark mTHP:16K Base Patched numa01 numa01 224.70 143.48 numa01_THREAD_ALLOC numa01_THREAD_ALLOC 118.05 47.43 numa02 numa02 13.45 9.29 numa02_SMT numa02_SMT 14.80 7.50 mTHP:64K Base Patched numa01 numa01 216.15 114.40 numa01_THREAD_ALLOC numa01_THREAD_ALLOC 115.35 47.41 numa02 numa02 13.24 9.25 numa02_SMT numa02_SMT 14.67 7.34 mTHP:128K Base Patched numa01 numa01 205.13 144.45 numa01_THREAD_ALLOC numa01_THREAD_ALLOC 112.93 41.88 numa02 numa02 13.16 9.18 numa02_SMT numa02_SMT 14.81 7.49 [1] https://lore.kernel.org/all/20231117100745.fnpijbk4xgmals3k@techsingularity.net/ [baolin.wang@linux.alibaba.com: v3] Link: https://lkml.kernel.org/r/c33a5c0b0a0323b1f8ed53772f50501f4b196e25.1712132950.git.baolin.wang@linux.alibaba.com Link: https://lkml.kernel.org/r/d28d276d599c26df7f38c9de8446f60e22dd1950.1711683069.git.baolin.wang@linux.alibaba.comSigned-off-by:
-
Baolin Wang authored
Patch series "support multi-size THP numa balancing", v2. This patchset tries to support mTHP numa balancing, as a simple solution to start, the NUMA balancing algorithm for mTHP will follow the THP strategy as the basic support. Please find details in each patch. This patch (of 2): To support large folio's numa balancing, factor out the numa mapping rebuilding into a new helper as a preparation. Link: https://lkml.kernel.org/r/cover.1712132950.git.baolin.wang@linux.alibaba.com Link: https://lkml.kernel.org/r/cover.1711683069.git.baolin.wang@linux.alibaba.com Link: https://lkml.kernel.org/r/8bc2586bdd8dbbe6d83c09b77b360ec8fcac3736.1711683069.git.baolin.wang@linux.alibaba.comSigned-off-by:
-
Barry Song authored
Fallback rates surpassing 90% have been observed on phones utilizing 64KiB CONT-PTE mTHP. In these scenarios, when one out of every 16 PTEs fails to allocate large folios, the remaining 15 PTEs fallback. Consequently, invoking vma_thp_gfp_mask seems redundant in such cases. Furthermore, abstaining from its use can also contribute to improved code readability. Link: https://lkml.kernel.org/r/20240329073750.20012-1-21cnbao@gmail.comSigned-off-by:
Zi Yan <ziy@nvidia.com> Acked-by:
Yu Zhao <yuzhao@google.com> Reviewed-by:
-
Sergey Senozhatsky authored
Introduce "max_pages" param to recompress device attribute which sets an upper limit on the number of entries (pages) zram attempts to recompress (in this particular recompression call). S/W recompression can be quite expensive so limiting the number of pages recompress touches can be quite helpful. Link: https://lkml.kernel.org/r/20240329094050.2815699-1-senozhatsky@chromium.orgSigned-off-by:
Sergey Senozhatsky <senozhatsky@chromium.org> Acked-by:
Brian Geffon <bgeffon@google.com> Cc: Minchan Kim <minchan@kernel.org> Signed-off-by:
-
York Jasper Niebuhr authored
Implements the "init_mlocked_on_free" boot option. When this boot option is enabled, any mlock'ed pages are zeroed on free. If the pages are munlock'ed beforehand, no initialization takes place. This boot option is meant to combat the performance hit of "init_on_free" as reported in commit 6471384a ("mm: security: introduce init_on_alloc=1 and init_on_free=1 boot options"). With "init_mlocked_on_free=1" only relevant data is freed while everything else is left untouched by the kernel. Correspondingly, this patch introduces no performance hit for unmapping non-mlock'ed memory. The unmapping overhead for purely mlocked memory was measured to be approximately 13%. Realistically, most systems mlock only a fraction of the total memory so the real-world system overhead should be close to zero. Optimally, userspace programs clear any key material or other confidential memory before exit and munlock the according memory regions. If a program crashes, userspace key managers fail to do this job. Accordingly, no munlock operations are performed so the data is caught and zeroed by the kernel. Should the program not crash, all memory will ideally be munlocked so no overhead is caused. CONFIG_INIT_MLOCKED_ON_FREE_DEFAULT_ON can be set to enable "init_mlocked_on_free" by default. Link: https://lkml.kernel.org/r/20240329145605.149917-1-yjnworkstation@gmail.comSigned-off-by:
York Jasper Niebuhr <yjnworkstation@gmail.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: York Jasper Niebuhr <yjnworkstation@gmail.com> Cc: Kees Cook <keescook@chromium.org> Signed-off-by:
-
Jinjiang Tu authored
This extends test_prctl_fork() and test_prctl_fork_exec() to make sure that deduplication really happens, instead of only testing the MMF_VM_MERGE_ANY flag is set. [colin.i.king@gmail.com: fix spelling mistake in ksft_test_result_skip message] Link: https://lkml.kernel.org/r/20240402081537.1365939-1-colin.i.king@gmail.com Link: https://lkml.kernel.org/r/20240328111010.1502191-4-tujinjiang@huawei.comSigned-off-by:
Jinjiang Tu <tujinjiang@huawei.com> Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Suggested-by:
-
Jinjiang Tu authored
In order to extend test_prctl_fork() and test_prctl_fork_exec() to make sure that deduplication really happens, mmap_and_merge_range() needs to be refactored. Firstly, mmap_and_merge_range() will be called with no need to call enable KSM by madvise or prctl. So, switch the 'bool use_prctl' parameter to enum ksm_merge_mode. Secondly, mmap_and_merge_range() will be called in child process in the two testcases, it isn't appropriate to call ksft_test_result_{fail, skip}, because the global variables ksft_{fail, skip} aren't consistent with the parent process. Thus, convert calls of ksft_test_result_{fail, skip} to ksft_print_msg(), return differrent error according to the two cases, and rename mmap_and_merge_range() to __mmap_and_merge_range(). For existing callers, introduce new mmap_and_merge_range() to handle different return values of __mmap_and_merge_range(). Link: https://lkml.kernel.org/r/20240328111010.1502191-3-tujinjiang@huawei.comSigned-off-by: -
Jinjiang Tu authored
Patch series "mm/ksm: fix ksm exec support for prctl", v4. commit 3c6f33b7 ("mm/ksm: support fork/exec for prctl") inherits MMF_VM_MERGE_ANY flag when a task calls execve(). However, it doesn't create the mm_slot, so ksmd will not try to scan this task. The first patch fixes the issue. The second patch refactors to prepare for the third patch. The third patch extends the selftests of ksm to verfity the deduplication really happens after fork/exec inherits ths KSM setting. This patch (of 3): commit 3c6f33b7 ("mm/ksm: support fork/exec for prctl") inherits MMF_VM_MERGE_ANY flag when a task calls execve(). Howerver, it doesn't create the mm_slot, so ksmd will not try to scan this task. To fix it, allocate and add the mm_slot to ksm_mm_head in __bprm_mm_init() when the mm has MMF_VM_MERGE_ANY flag. Link: https://lkml.kernel.org/r/20240328111010.1502191-1-tujinjiang@huawei.com Link: https://lkml.kernel.org/r/20240328111010.1502191-2-tujinjiang@huawei.com Fixes: 3c6f33b7 ("mm/ksm: support fork/exec for prctl") Signed-off-by:
-
Rick Edgecombe authored
The existing shadow stack test for guard gaps just checks that new mappings are not placed in an existing mapping's guard gap. Add one that checks that new mappings are not placed such that preexisting mappings are in the new mappings guard gap. Link: https://lkml.kernel.org/r/20240326021656.202649-15-rick.p.edgecombe@intel.comSigned-off-by:
Rick Edgecombe <rick.p.edgecombe@intel.com> Cc: Alexei Starovoitov <ast@kernel.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Aneesh Kumar K.V <aneesh.kumar@kernel.org> Cc: Borislav Petkov (AMD) <bp@alien8.de> Cc: Christophe Leroy <christophe.leroy@csgroup.eu> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: Deepak Gupta <debug@rivosinc.com> Cc: Guo Ren <guoren@kernel.org> Cc: Helge Deller <deller@gmx.de> Cc: H. Peter Anvin (Intel) <hpa@zytor.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com> Cc: Kees Cook <keescook@chromium.org> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Liam R. Howlett <Liam.Howlett@oracle.com> Cc: Mark Brown <broonie@kernel.org> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Naveen N. Rao <naveen.n.rao@linux.ibm.com> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
