From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id A94113CA48E for ; Sun, 3 May 2026 13:20:41 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1777814441; cv=none; b=gk9N/S1tje7CgElYWegCJ6moM20Z5aSbUBPBRgNmE7N7jjrj2D/7ZFWihSizaZW34ISGwK4BhT7Bu3RfdmmdY6F9bHJ39QhNf09822Tq7Xrcqr7vM3vKgHLW877dEZyfwUVz6tV6r0C++6/gNLsiHJ+basj7M5jvp2kqZq1E2uA= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1777814441; c=relaxed/simple; bh=672ZsjRBQ4tLmM9sg3nMeDblbAAebi8px4GL8tAIOkc=; h=Date:To:From:Subject:Message-Id; b=YFtFe2yiVQ6vAuvGR2XTFCtScKFh91vgBlvsrLF0fylBrt0faOJcFjmoeO2KQDQXOoJ4qhJMtsbvI16jv8+bSYlTtmt62UuXd8vMb46aTBl4uRQv/Ou6Kq1fcld9qQ2tY/aREkEjRSmtaabHDNpVHiWYLqBxepi9aeE/j9h+pLM= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux-foundation.org header.i=@linux-foundation.org header.b=Kse0vnm0; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux-foundation.org header.i=@linux-foundation.org header.b="Kse0vnm0" Received: by smtp.kernel.org (Postfix) with ESMTPSA id A2FA9C2BCB4; Sun, 3 May 2026 13:20:40 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linux-foundation.org; s=korg; t=1777814441; bh=672ZsjRBQ4tLmM9sg3nMeDblbAAebi8px4GL8tAIOkc=; h=Date:To:From:Subject:From; b=Kse0vnm0KVHsQ+y0XoTTSmdw6reJhZFWvsPsHk3c8qNlFIsG2x1VfRSWbd+iF6dRJ oWVVIDZUah+aiNdVg7x+/RLirAchQZdBbCx4IsWFYS2ubcOZ0BL2tWf1P/BJEzVGGo kmLEpM6ZIoTEkKXnu8Q4GiHFx9m5tiAF79Bviu3E= Date: Sun, 03 May 2026 06:20:38 -0700 To: mm-commits@vger.kernel.org,zokeefe@google.com,ziy@nvidia.com,ying.huang@linux.alibaba.com,yang@os.amperecomputing.com,willy@infradead.org,will@kernel.org,wangkefeng.wang@huawei.com,vishal.moola@gmail.com,vbabka@suse.cz,usama.arif@linux.dev,tiwai@suse.de,thomas.hellstrom@linux.intel.com,surenb@google.com,sunnanyong@huawei.com,shivankg@amd.com,ryan.roberts@arm.com,rppt@kernel.org,rostedt@goodmis.org,rientjes@google.com,richard.weiyang@gmail.com,rdunlap@infradead.org,raquini@redhat.com,rakie.kim@sk.com,pfalcato@suse.de,peterx@redhat.com,mhocko@suse.com,mhiramat@kernel.org,matthew.brost@intel.com,mathieu.desnoyers@efficios.com,ljs@kernel.org,liam@infradead.org,lance.yang@linux.dev,joshua.hahnjy@gmail.com,jannh@google.com,jack@suse.cz,jackmanb@google.com,hughd@google.com,hannes@cmpxchg.org,gourry@gourry.net,dev.jain@arm.com,david@kernel.org,corbet@lwn.net,catalin.marinas@arm.com,byungchul@sk.com,baolin.wang@linux.alibaba.com,baohua@kernel.org,bagasdotme@gmail.com,apopple@nvidia.com,anshuman.khandual@arm.com,aarcange@redhat.com,npache@redhat.com,akpm@linux-foundation.org From: Andrew Morton Subject: [to-be-updated] mm-khugepaged-generalize-__collapse_huge_page_-for-mthp-support.patch removed from -mm tree Message-Id: <20260503132040.A2FA9C2BCB4@smtp.kernel.org> Precedence: bulk X-Mailing-List: mm-commits@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: The quilt patch titled Subject: mm/khugepaged: generalize __collapse_huge_page_* for mTHP support has been removed from the -mm tree. Its filename was mm-khugepaged-generalize-__collapse_huge_page_-for-mthp-support.patch This patch was dropped because an updated version will be issued ------------------------------------------------------ From: Nico Pache Subject: mm/khugepaged: generalize __collapse_huge_page_* for mTHP support Date: Sun, 19 Apr 2026 12:57:41 -0600 Generalize the order of the __collapse_huge_page_* and collapse_max_* functions to support future mTHP collapse. The current mechanism for determining collapse with the khugepaged_max_ptes_none value is not designed with mTHP in mind. This raises a key design issue: if we support user defined max_pte_none values (even those scaled by order), a collapse of a lower order can introduces an feedback loop, or "creep", when max_ptes_none is set to a value greater than HPAGE_PMD_NR / 2. With this configuration, a successful collapse to order N will populate enough pages to satisfy the collapse condition on order N+1 on the next scan. This leads to unnecessary work and memory churn. To fix this issue introduce a helper function that will limit mTHP collapse support to two max_ptes_none values, 0 and HPAGE_PMD_NR - 1. This effectively supports two modes: - max_ptes_none=0: never introduce new none-pages for mTHP collapse. - max_ptes_none=511 (on 4k pagesz): Always collapse to the highest available mTHP order. This removes the possiblilty of "creep", while not modifying any uAPI expectations. A warning will be emitted if any non-supported max_ptes_none value is configured with mTHP enabled. mTHP collapse will not honor the khugepaged_max_ptes_shared or khugepaged_max_ptes_swap parameters, and will fail if it encounters a shared or swapped entry. No functional changes in this patch; however it defines future behavior for mTHP collapse. Link: https://lore.kernel.org/20260419185750.260784-5-npache@redhat.com Co-developed-by: Dev Jain Signed-off-by: Dev Jain Signed-off-by: Nico Pache Cc: Alistair Popple Cc: Andrea Arcangeli Cc: Anshuman Khandual Cc: Bagas Sanjaya Cc: Baolin Wang Cc: Barry Song Cc: Brendan Jackman Cc: Byungchul Park Cc: Catalin Marinas Cc: David Hildenbrand (Arm) Cc: David Rientjes Cc: Gregory Price Cc: "Huang, Ying" Cc: Hugh Dickins Cc: Jan Kara Cc: Jann Horn Cc: Johannes Weiner Cc: Jonathan Corbet Cc: Joshua Hahn Cc: Kefeng Wang Cc: Lance Yang Cc: Liam Howlett Cc: Lorenzo Stoakes Cc: "Masami Hiramatsu (Google)" Cc: Mathieu Desnoyers Cc: Matthew Brost Cc: Matthew Wilcox (Oracle) Cc: Michal Hocko Cc: Mike Rapoport Cc: Nanyong Sun Cc: Pedro Falcato Cc: Peter Xu Cc: Rafael Aquini Cc: Rakie Kim Cc: Randy Dunlap Cc: Ryan Roberts Cc: Shivank Garg Cc: Steven Rostedt Cc: Suren Baghdasaryan Cc: Takashi Iwai (SUSE) Cc: Thomas Hellström Cc: Usama Arif Cc: Vishal Moola (Oracle) Cc: Vlastimil Babka Cc: Wei Yang Cc: Will Deacon Cc: Yang Shi Cc: Zach O'Keefe Cc: Zi Yan Signed-off-by: Andrew Morton --- mm/khugepaged.c | 124 ++++++++++++++++++++++++++++++++-------------- 1 file changed, 88 insertions(+), 36 deletions(-) --- a/mm/khugepaged.c~mm-khugepaged-generalize-__collapse_huge_page_-for-mthp-support +++ a/mm/khugepaged.c @@ -352,51 +352,86 @@ static bool pte_none_or_zero(pte_t pte) * collapse_max_ptes_none - Calculate maximum allowed empty PTEs for collapse * @cc: The collapse control struct * @vma: The vma to check for userfaultfd + * @order: The folio order being collapsed to * * If we are not in khugepaged mode use HPAGE_PMD_NR to allow any - * empty page. + * empty page. For PMD-sized collapses (order == HPAGE_PMD_ORDER), use the + * configured khugepaged_max_ptes_none value. + * + * For mTHP collapses, we currently only support khugepaged_max_pte_none values + * of 0 or (KHUGEPAGED_MAX_PTES_LIMIT). Any other value will emit a warning and + * no mTHP collapse will be attempted * * Return: Maximum number of empty PTEs allowed for the collapse operation */ -static unsigned int collapse_max_ptes_none(struct collapse_control *cc, - struct vm_area_struct *vma) +static int collapse_max_ptes_none(struct collapse_control *cc, + struct vm_area_struct *vma, unsigned int order) { if (vma && userfaultfd_armed(vma)) return 0; if (!cc->is_khugepaged) return HPAGE_PMD_NR; - return khugepaged_max_ptes_none; + if (is_pmd_order(order)) + return khugepaged_max_ptes_none; + /* Zero/non-present collapse disabled. */ + if (!khugepaged_max_ptes_none) + return 0; + if (khugepaged_max_ptes_none == KHUGEPAGED_MAX_PTES_LIMIT) + return (1 << order) - 1; + + pr_warn_once("mTHP collapse only supports max_ptes_none values of 0 or %u\n", + KHUGEPAGED_MAX_PTES_LIMIT); + return -EINVAL; } /** * collapse_max_ptes_shared - Calculate maximum allowed shared PTEs for collapse * @cc: The collapse control struct + * @order: The folio order being collapsed to * * If we are not in khugepaged mode use HPAGE_PMD_NR to allow any * shared page. * + * For mTHP collapses, we currently dont support collapsing memory with + * shared memory. + * * Return: Maximum number of shared PTEs allowed for the collapse operation */ -static unsigned int collapse_max_ptes_shared(struct collapse_control *cc) +static unsigned int collapse_max_ptes_shared(struct collapse_control *cc, + unsigned int order) { if (!cc->is_khugepaged) return HPAGE_PMD_NR; + if (!is_pmd_order(order)) + return 0; + return khugepaged_max_ptes_shared; } /** * collapse_max_ptes_swap - Calculate maximum allowed swap PTEs for collapse * @cc: The collapse control struct + * @order: The folio order being collapsed to * * If we are not in khugepaged mode use HPAGE_PMD_NR to allow any * swap page. * + * For PMD-sized collapses (order == HPAGE_PMD_ORDER), use the configured + * khugepaged_max_ptes_swap value. + * + * For mTHP collapses, we currently dont support collapsing memory with + * swapped out memory. + * * Return: Maximum number of swap PTEs allowed for the collapse operation */ -static unsigned int collapse_max_ptes_swap(struct collapse_control *cc) +static unsigned int collapse_max_ptes_swap(struct collapse_control *cc, + unsigned int order) { if (!cc->is_khugepaged) return HPAGE_PMD_NR; + if (!is_pmd_order(order)) + return 0; + return khugepaged_max_ptes_swap; } @@ -590,18 +625,22 @@ static void release_pte_pages(pte_t *pte static enum scan_result __collapse_huge_page_isolate(struct vm_area_struct *vma, unsigned long start_addr, pte_t *pte, struct collapse_control *cc, - struct list_head *compound_pagelist) + unsigned int order, struct list_head *compound_pagelist) { + const unsigned long nr_pages = 1UL << order; struct page *page = NULL; struct folio *folio = NULL; unsigned long addr = start_addr; pte_t *_pte; int none_or_zero = 0, shared = 0, referenced = 0; enum scan_result result = SCAN_FAIL; - unsigned int max_ptes_none = collapse_max_ptes_none(cc, vma); - unsigned int max_ptes_shared = collapse_max_ptes_shared(cc); + int max_ptes_none = collapse_max_ptes_none(cc, vma, order); + unsigned int max_ptes_shared = collapse_max_ptes_shared(cc, order); + + if (max_ptes_none < 0) + return result; - for (_pte = pte; _pte < pte + HPAGE_PMD_NR; + for (_pte = pte; _pte < pte + nr_pages; _pte++, addr += PAGE_SIZE) { pte_t pteval = ptep_get(_pte); if (pte_none_or_zero(pteval)) { @@ -734,18 +773,18 @@ out: } static void __collapse_huge_page_copy_succeeded(pte_t *pte, - struct vm_area_struct *vma, - unsigned long address, - spinlock_t *ptl, - struct list_head *compound_pagelist) + struct vm_area_struct *vma, unsigned long address, + spinlock_t *ptl, unsigned int order, + struct list_head *compound_pagelist) { - unsigned long end = address + HPAGE_PMD_SIZE; + const unsigned long nr_pages = 1UL << order; + unsigned long end = address + (PAGE_SIZE << order); struct folio *src, *tmp; pte_t pteval; pte_t *_pte; unsigned int nr_ptes; - for (_pte = pte; _pte < pte + HPAGE_PMD_NR; _pte += nr_ptes, + for (_pte = pte; _pte < pte + nr_pages; _pte += nr_ptes, address += nr_ptes * PAGE_SIZE) { nr_ptes = 1; pteval = ptep_get(_pte); @@ -798,13 +837,11 @@ static void __collapse_huge_page_copy_su } static void __collapse_huge_page_copy_failed(pte_t *pte, - pmd_t *pmd, - pmd_t orig_pmd, - struct vm_area_struct *vma, - struct list_head *compound_pagelist) + pmd_t *pmd, pmd_t orig_pmd, struct vm_area_struct *vma, + unsigned int order, struct list_head *compound_pagelist) { + const unsigned long nr_pages = 1UL << order; spinlock_t *pmd_ptl; - /* * Re-establish the PMD to point to the original page table * entry. Restoring PMD needs to be done prior to releasing @@ -818,7 +855,7 @@ static void __collapse_huge_page_copy_fa * Release both raw and compound pages isolated * in __collapse_huge_page_isolate. */ - release_pte_pages(pte, pte + HPAGE_PMD_NR, compound_pagelist); + release_pte_pages(pte, pte + nr_pages, compound_pagelist); } /* @@ -838,16 +875,16 @@ static void __collapse_huge_page_copy_fa */ static enum scan_result __collapse_huge_page_copy(pte_t *pte, struct folio *folio, pmd_t *pmd, pmd_t orig_pmd, struct vm_area_struct *vma, - unsigned long address, spinlock_t *ptl, + unsigned long address, spinlock_t *ptl, unsigned int order, struct list_head *compound_pagelist) { + const unsigned long nr_pages = 1UL << order; unsigned int i; enum scan_result result = SCAN_SUCCEED; - /* * Copying pages' contents is subject to memory poison at any iteration. */ - for (i = 0; i < HPAGE_PMD_NR; i++) { + for (i = 0; i < nr_pages; i++) { pte_t pteval = ptep_get(pte + i); struct page *page = folio_page(folio, i); unsigned long src_addr = address + i * PAGE_SIZE; @@ -866,10 +903,10 @@ static enum scan_result __collapse_huge_ if (likely(result == SCAN_SUCCEED)) __collapse_huge_page_copy_succeeded(pte, vma, address, ptl, - compound_pagelist); + order, compound_pagelist); else __collapse_huge_page_copy_failed(pte, pmd, orig_pmd, vma, - compound_pagelist); + order, compound_pagelist); return result; } @@ -1040,12 +1077,12 @@ static enum scan_result check_pmd_still_ * Returns result: if not SCAN_SUCCEED, mmap_lock has been released. */ static enum scan_result __collapse_huge_page_swapin(struct mm_struct *mm, - struct vm_area_struct *vma, unsigned long start_addr, pmd_t *pmd, - int referenced) + struct vm_area_struct *vma, unsigned long start_addr, + pmd_t *pmd, int referenced, unsigned int order) { int swapped_in = 0; vm_fault_t ret = 0; - unsigned long addr, end = start_addr + (HPAGE_PMD_NR * PAGE_SIZE); + unsigned long addr, end = start_addr + (PAGE_SIZE << order); enum scan_result result; pte_t *pte = NULL; spinlock_t *ptl; @@ -1077,6 +1114,19 @@ static enum scan_result __collapse_huge_ pte_present(vmf.orig_pte)) continue; + /* + * TODO: Support swapin without leading to further mTHP + * collapses. Currently bringing in new pages via swapin may + * cause a future higher order collapse on a rescan of the same + * range. + */ + if (!is_pmd_order(order)) { + pte_unmap(pte); + mmap_read_unlock(mm); + result = SCAN_EXCEED_SWAP_PTE; + goto out; + } + vmf.pte = pte; vmf.ptl = ptl; ret = do_swap_page(&vmf); @@ -1196,7 +1246,7 @@ static enum scan_result collapse_huge_pa * that case. Continuing to collapse causes inconsistency. */ result = __collapse_huge_page_swapin(mm, vma, address, pmd, - referenced); + referenced, HPAGE_PMD_ORDER); if (result != SCAN_SUCCEED) goto out_nolock; } @@ -1244,6 +1294,7 @@ static enum scan_result collapse_huge_pa pte = pte_offset_map_lock(mm, &_pmd, address, &pte_ptl); if (pte) { result = __collapse_huge_page_isolate(vma, address, pte, cc, + HPAGE_PMD_ORDER, &compound_pagelist); spin_unlock(pte_ptl); } else { @@ -1274,6 +1325,7 @@ static enum scan_result collapse_huge_pa result = __collapse_huge_page_copy(pte, folio, pmd, _pmd, vma, address, pte_ptl, + HPAGE_PMD_ORDER, &compound_pagelist); pte_unmap(pte); if (unlikely(result != SCAN_SUCCEED)) @@ -1318,9 +1370,9 @@ static enum scan_result collapse_scan_pm unsigned long addr; spinlock_t *ptl; int node = NUMA_NO_NODE, unmapped = 0; - unsigned int max_ptes_none = collapse_max_ptes_none(cc, vma); - unsigned int max_ptes_shared = collapse_max_ptes_shared(cc); - unsigned int max_ptes_swap = collapse_max_ptes_swap(cc); + int max_ptes_none = collapse_max_ptes_none(cc, vma, HPAGE_PMD_ORDER); + unsigned int max_ptes_shared = collapse_max_ptes_shared(cc, HPAGE_PMD_ORDER); + unsigned int max_ptes_swap = collapse_max_ptes_swap(cc, HPAGE_PMD_ORDER); VM_BUG_ON(start_addr & ~HPAGE_PMD_MASK); @@ -2371,8 +2423,8 @@ static enum scan_result collapse_scan_fi int present, swap; int node = NUMA_NO_NODE; enum scan_result result = SCAN_SUCCEED; - unsigned int max_ptes_none = collapse_max_ptes_none(cc, NULL); - unsigned int max_ptes_swap = collapse_max_ptes_swap(cc); + int max_ptes_none = collapse_max_ptes_none(cc, NULL, HPAGE_PMD_ORDER); + unsigned int max_ptes_swap = collapse_max_ptes_swap(cc, HPAGE_PMD_ORDER); present = 0; swap = 0; _ Patches currently in -mm which might be from npache@redhat.com are mm-khugepaged-generalize-collapse_huge_page-for-mthp-collapse.patch mm-khugepaged-skip-collapsing-mthp-to-smaller-orders.patch mm-khugepaged-add-per-order-mthp-collapse-failure-statistics.patch mm-khugepaged-improve-tracepoints-for-mthp-orders.patch mm-khugepaged-introduce-collapse_allowable_orders-helper-function.patch mm-khugepaged-introduce-mthp-collapse-support.patch mm-khugepaged-avoid-unnecessary-mthp-collapse-attempts.patch documentation-mm-update-the-admin-guide-for-mthp-collapse.patch