{"id":2223482,"url":"http://patchwork.ozlabs.org/api/1.2/patches/2223482/?format=json","web_url":"http://patchwork.ozlabs.org/project/linuxppc-dev/patch/20260415111412.1003526-6-songmuchun@bytedance.com/","project":{"id":2,"url":"http://patchwork.ozlabs.org/api/1.2/projects/2/?format=json","name":"Linux PPC development","link_name":"linuxppc-dev","list_id":"linuxppc-dev.lists.ozlabs.org","list_email":"linuxppc-dev@lists.ozlabs.org","web_url":"https://github.com/linuxppc/wiki/wiki","scm_url":"https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git","webscm_url":"https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git/","list_archive_url":"https://lore.kernel.org/linuxppc-dev/","list_archive_url_format":"https://lore.kernel.org/linuxppc-dev/{}/","commit_url_format":"https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git/commit/?id={}"},"msgid":"<20260415111412.1003526-6-songmuchun@bytedance.com>","list_archive_url":"https://lore.kernel.org/linuxppc-dev/20260415111412.1003526-6-songmuchun@bytedance.com/","date":"2026-04-15T11:14:11","name":"[v2,5/6] mm/sparse-vmemmap: Fix missing architecture-specific page table sync","commit_ref":null,"pull_url":null,"state":"handled-elsewhere","archived":false,"hash":"8a7fe2185f86cb2e815c8be20658fa446f05943d","submitter":{"id":78930,"url":"http://patchwork.ozlabs.org/api/1.2/people/78930/?format=json","name":"Muchun Song","email":"songmuchun@bytedance.com"},"delegate":null,"mbox":"http://patchwork.ozlabs.org/project/linuxppc-dev/patch/20260415111412.1003526-6-songmuchun@bytedance.com/mbox/","series":[{"id":499970,"url":"http://patchwork.ozlabs.org/api/1.2/series/499970/?format=json","web_url":"http://patchwork.ozlabs.org/project/linuxppc-dev/list/?series=499970","date":"2026-04-15T11:14:06","name":"mm: Fix vmemmap optimization accounting and initialization","version":2,"mbox":"http://patchwork.ozlabs.org/series/499970/mbox/"}],"comments":"http://patchwork.ozlabs.org/api/patches/2223482/comments/","check":"pending","checks":"http://patchwork.ozlabs.org/api/patches/2223482/checks/","tags":{},"related":[],"headers":{"Return-Path":"\n ","X-Original-To":["incoming@patchwork.ozlabs.org","linuxppc-dev@lists.ozlabs.org"],"Delivered-To":"patchwork-incoming@legolas.ozlabs.org","Authentication-Results":["legolas.ozlabs.org;\n\tdkim=pass (2048-bit key;\n unprotected) header.d=bytedance.com header.i=@bytedance.com\n header.a=rsa-sha256 header.s=google header.b=cQN6iqDh;\n\tdkim-atps=neutral","legolas.ozlabs.org;\n spf=pass (sender SPF authorized) smtp.mailfrom=lists.ozlabs.org\n (client-ip=2404:9400:21b9:f100::1; helo=lists.ozlabs.org;\n envelope-from=linuxppc-dev+bounces-19768-incoming=patchwork.ozlabs.org@lists.ozlabs.org;\n receiver=patchwork.ozlabs.org)","lists.ozlabs.org;\n arc=none smtp.remote-ip=\"2607:f8b0:4864:20::536\"","lists.ozlabs.org;\n dmarc=pass (p=quarantine dis=none) header.from=bytedance.com","lists.ozlabs.org;\n\tdkim=pass (2048-bit key;\n unprotected) header.d=bytedance.com header.i=@bytedance.com\n header.a=rsa-sha256 header.s=google header.b=cQN6iqDh;\n\tdkim-atps=neutral","lists.ozlabs.org;\n spf=pass (sender SPF authorized) smtp.mailfrom=bytedance.com\n (client-ip=2607:f8b0:4864:20::536; helo=mail-pg1-x536.google.com;\n envelope-from=songmuchun@bytedance.com; receiver=lists.ozlabs.org)"],"Received":["from lists.ozlabs.org (lists.ozlabs.org\n [IPv6:2404:9400:21b9:f100::1])\n\t(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n\t key-exchange x25519)\n\t(No client certificate requested)\n\tby legolas.ozlabs.org (Postfix) with ESMTPS id 4fwdpc2RS8z1yCv\n\tfor ; Wed, 15 Apr 2026 21:15:04 +1000 (AEST)","from boromir.ozlabs.org (localhost [127.0.0.1])\n\tby lists.ozlabs.org (Postfix) with ESMTP id 4fwdpR2D6Tz30TL;\n\tWed, 15 Apr 2026 21:14:55 +1000 (AEST)","from mail-pg1-x536.google.com (mail-pg1-x536.google.com\n [IPv6:2607:f8b0:4864:20::536])\n\t(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n\t key-exchange x25519 server-signature RSA-PSS (2048 bits) server-digest\n SHA256)\n\t(No client certificate requested)\n\tby lists.ozlabs.org (Postfix) with ESMTPS id 4fwdpQ3vKgz2yvG\n\tfor ; Wed, 15 Apr 2026 21:14:54 +1000 (AEST)","by mail-pg1-x536.google.com with SMTP id\n 41be03b00d2f7-c6e2355739dso2985899a12.2\n for ;\n Wed, 15 Apr 2026 04:14:54 -0700 (PDT)","from n232-176-004.byted.org ([36.110.163.96])\n by smtp.gmail.com with ESMTPSA id\n d9443c01a7336-2b477fd3724sm19509485ad.0.2026.04.15.04.14.47\n (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256);\n Wed, 15 Apr 2026 04:14:52 -0700 (PDT)"],"ARC-Seal":"i=1; a=rsa-sha256; d=lists.ozlabs.org; s=201707; t=1776251695;\n\tcv=none;\n b=gkWMRWqGmOtBM5nYRZyVYmzvdjutEPetPsXrwcA05T3rcqMzfrirQPyWT/HuMtCM9WPOsrQtiXk39tu24fG+KWS7VOMOoFRSi3XoWtpwiMHy8M/AO3T+paGUa9/ozNEPppwDawS/2vgLx5+NLtLT2QrNdGNTTRZrsopGIwvPT8RLgQu1PcOj+mUKPXctvAe5amEJ8GofCIX05ppJlVFJpOovboQdZ8RxjJ/lmJkEQpdGmpuUF/cB4vBa/QwG/F8ALMxVYK2Uqa2eKxURn5GikLAJ4dtnijegza1JbLT8rhrvr8SJ83iLjXDKKv80YlMBcktD0oQ8PshL4vqiP+a3KA==","ARC-Message-Signature":"i=1; a=rsa-sha256; d=lists.ozlabs.org; s=201707;\n\tt=1776251695; c=relaxed/relaxed;\n\tbh=UUwLFe4sJwTmvK33kSzO0AkxCRTRozJKSvcgK9DIV10=;\n\th=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References:\n\t MIME-Version;\n b=JGb4anTThmFqJEZ8HfQS39iYtGkYdM+ddRnqmhHXHNhtl5G6mXVLCntwUg7+mN19huPqa8JtkyUVbfC0xf15/tXaYOzmTkI4M1puGZUpZ3JUxgmTmXnvL2+HhwZ4pDgMWHJ77mRSDOWcplxR6LnwgeBjjhyxhLsYtZuKYjpwWb2/UpfITaQ5KG3Lh/qAH2rGvktiAMnOcvtk7Q7qfSdw6N/5LBxsPZFebRTGAv5rRw1JItNUR3XVgMlrhY9uwe8ER79hHSVAJTA8djv0/ZCPBqA2TWMQWNEFG72UBxXE5bdvg8WipaKOZEJ9virxX/RhJhKh/IOcYivZLko8xxfEog==","ARC-Authentication-Results":"i=1; lists.ozlabs.org;\n dmarc=pass (p=quarantine dis=none) header.from=bytedance.com;\n dkim=pass (2048-bit key;\n unprotected) header.d=bytedance.com header.i=@bytedance.com\n header.a=rsa-sha256 header.s=google header.b=cQN6iqDh; dkim-atps=neutral;\n spf=pass (client-ip=2607:f8b0:4864:20::536; helo=mail-pg1-x536.google.com;\n envelope-from=songmuchun@bytedance.com;\n receiver=lists.ozlabs.org) smtp.mailfrom=bytedance.com","DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed;\n d=bytedance.com; s=google; t=1776251693; x=1776856493;\n darn=lists.ozlabs.org;\n h=content-transfer-encoding:mime-version:references:in-reply-to\n :message-id:date:subject:cc:to:from:from:to:cc:subject:date\n :message-id:reply-to;\n bh=UUwLFe4sJwTmvK33kSzO0AkxCRTRozJKSvcgK9DIV10=;\n b=cQN6iqDhKj0c/ObCuxXautvQ1q4xtpNq+Qy4ELWcyS4UhTBL78KkzbZ9KQOHMm9lpH\n jI3pRd359uicuPe+mAA5O7QfB4gzCkbNN9XF3G7kHMn3L9XjT2YqFo06fml21+KfRO7d\n 3QP3ATE4aTkT6r8nDEzRZBf2deQExkVsyh0cdw2Em+KmuaWotj2p2JR38E03VvojOTqo\n FtA5VN/1800yo0mzRq58Rmn/vVR2gL0hbUAO58Zkcn+j8w27wPE4WsNRakWEz81AAITO\n Kov0Qaj9h1LrAeKYjDUuK6RSaVMtWLYoA0WjIL7f5SFIpvV1SYZEpMDT5dzQFVOvm+nO\n 6SEA==","X-Google-DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed;\n d=1e100.net; s=20251104; t=1776251693; x=1776856493;\n h=content-transfer-encoding:mime-version:references:in-reply-to\n :message-id:date:subject:cc:to:from:x-gm-gg:x-gm-message-state:from\n :to:cc:subject:date:message-id:reply-to;\n bh=UUwLFe4sJwTmvK33kSzO0AkxCRTRozJKSvcgK9DIV10=;\n b=aGPHBqBTbVHtdqCStFvkese9E/ZFPRO7AUFNec6vqou6hnqc0cOHKpoK1AK0zHzso3\n +cfQ7Vh/YbQLvBKyS5kQY7JBZDGvHLyEm/FAjEdSWnXYX+RaaujlWBbTyq7kz6/d142I\n sY4nfrjjF1ch90FUQEVN5HTNPDV9338Nr/8sLLb7tkzMOIiU0msgJCJEVo+B+WlZyHl3\n 4z2xrseZk191iOohh/h/nIOnP49ZHixEBmOWU3z8Yk+dxfeqiuK/BSuwuWlP8NlIIkkW\n K0LPbZhF9QdSbO9bv7/+yUhKLfIlA0MIQOCclZzEwRxRYR6k4FQjVTij3ucnfjteVv6Z\n Vckw==","X-Forwarded-Encrypted":"i=1;\n AFNElJ8sgtdJmoEERfV1DJo/TMvEbT8UaTzEAHblzl1bzXX8yxMRsMc679R0PD+m7t58SajxkPP5t8vIpwK4auk=@lists.ozlabs.org","X-Gm-Message-State":"AOJu0YxpGoDeZRB+t1dzgGf1HMsfJRgVJKmAJKQvilnNBCoT6uDQ99D/\n\t+yRl2Sum68GOWh3H59htvMaaZNUyxZNaJZAniyAjzKIut5baQTtymWxhGuqq35UIUlA=","X-Gm-Gg":"AeBDietUyGxJiMZFLzX2INE9Vc7QL1CwWUuEegHFDBRhw7Z5ikUdH2LK8f0aI3Wpuhp\n\t0LOcqJdzwqPgCvZd0/kQscPdOh8lj1tZsx3UwHhXs49YZuv8ezpXr+5Ahg0ySYcdlIcKrOQhcKZ\n\t3RSslyi10Ccf0l/q0z7s4Ca9IHop0NZ1rEESul/d29ZFf8FXQMNzLzDlWXb26AS2IF0CDx99vBl\n\tzeJ7dcCGt7/O4mop6BHIGR6syn4BoRCg3vGzP7c6hzhGMy1aAgwP8/zrR3grMdFJLajTP2SkmU4\n\te9FzVLOUzw1cjfAEpqw0mwO+1qjzaSO+v+lx4iaXX+ObVnpHtQ1uN8Hvh04bK+zM0PGmBEuh3Ay\n\t+jHIBTq9ZKU2BVnyWyEhjJha4qQINpsGuYBvWV4NDhzDLEKtA4VGdmJnQF4N50RikLUBnGGxxJk\n\t/QxX9DrgPjkSVH4eIL/tKAv5LljAGnzAyxjbgwfCCLZn4=","X-Received":"by 2002:a17:903:37c6:b0:2b2:49a7:a5bc with SMTP id\n d9443c01a7336-2b2d5a60603mr217754215ad.39.1776251692524;\n Wed, 15 Apr 2026 04:14:52 -0700 (PDT)","From":"Muchun Song ","To":"Andrew Morton ,\n\tDavid Hildenbrand ,\n\tMuchun Song ,\n\tOscar Salvador ,\n\tMichael Ellerman ,\n\tMadhavan Srinivasan ","Cc":"Lorenzo Stoakes ,\n\t\"Liam R . Howlett\" ,\n\tVlastimil Babka ,\n\tMike Rapoport ,\n\tSuren Baghdasaryan ,\n\tMichal Hocko ,\n\tNicholas Piggin ,\n\tChristophe Leroy ,\n\taneesh.kumar@linux.ibm.com,\n\tjoao.m.martins@oracle.com,\n\tlinux-mm@kvack.org,\n\tlinuxppc-dev@lists.ozlabs.org,\n\tlinux-kernel@vger.kernel.org,\n\tMuchun Song ","Subject":"[PATCH v2 5/6] mm/sparse-vmemmap: Fix missing architecture-specific\n page table sync","Date":"Wed, 15 Apr 2026 19:14:11 +0800","Message-Id":"<20260415111412.1003526-6-songmuchun@bytedance.com>","X-Mailer":"git-send-email 2.20.1","In-Reply-To":"<20260415111412.1003526-1-songmuchun@bytedance.com>","References":"<20260415111412.1003526-1-songmuchun@bytedance.com>","X-Mailing-List":"linuxppc-dev@lists.ozlabs.org","List-Id":"","List-Help":"","List-Owner":"","List-Post":"","List-Archive":",\n ","List-Subscribe":",\n ,\n ","List-Unsubscribe":"","Precedence":"list","MIME-Version":"1.0","Content-Transfer-Encoding":"8bit","X-Spam-Status":"No, score=-0.2 required=3.0 tests=DKIM_SIGNED,DKIM_VALID,\n\tDKIM_VALID_AU,DKIM_VALID_EF,RCVD_IN_DNSWL_NONE,SPF_HELO_NONE,SPF_PASS\n\tautolearn=disabled version=4.0.1 OzLabs 8","X-Spam-Checker-Version":"SpamAssassin 4.0.1 (2024-03-25) on lists.ozlabs.org"},"content":"On x86-64, vmemmap_populate() normally calls sync_global_pgds() to\nkeep the page tables in sync. However, when vmemmap optimization for\ncompound devmaps is enabled, vmemmap_populate_compound_pages() is called\ndirectly from __populate_section_memmap(), bypassing the architecture-\nspecific vmemmap_populate() entirely. This skips the sync on x86-64\nand can later trigger vmemmap-access faults.\n\nFix this by moving the vmemmap_can_optimize() dispatch from\n__populate_section_memmap() into the generic helpers --\nvmemmap_populate_basepages() and vmemmap_populate_hugepages(). This way,\nthe architecture vmemmap_populate() is always invoked first, ensuring\nany arch-specific post-population steps (e.g. sync_global_pgds()) are\nexecuted before returning.\n\nArchitectures that override vmemmap_populate() (e.g. powerpc) handle\nthe optimization dispatch in their own implementation instead.\n\nFixes: 4917f55b4ef9 (\"mm/sparse-vmemmap: improve memory savings for compound devmaps\")\nSigned-off-by: Muchun Song \n---\n arch/powerpc/include/asm/book3s/64/radix.h | 6 ------\n arch/powerpc/mm/book3s64/radix_pgtable.c | 16 ++++++++++-----\n mm/sparse-vmemmap.c | 24 +++++++++++-----------\n 3 files changed, 23 insertions(+), 23 deletions(-)","diff":"diff --git a/arch/powerpc/include/asm/book3s/64/radix.h b/arch/powerpc/include/asm/book3s/64/radix.h\nindex bde07c6f900f..2600defa2dc2 100644\n--- a/arch/powerpc/include/asm/book3s/64/radix.h\n+++ b/arch/powerpc/include/asm/book3s/64/radix.h\n@@ -357,11 +357,5 @@ int radix__remove_section_mapping(unsigned long start, unsigned long end);\n #define vmemmap_can_optimize vmemmap_can_optimize\n bool vmemmap_can_optimize(struct vmem_altmap *altmap, struct dev_pagemap *pgmap);\n #endif\n-\n-#define vmemmap_populate_compound_pages vmemmap_populate_compound_pages\n-int __meminit vmemmap_populate_compound_pages(unsigned long start_pfn,\n-\t\t\t\t\t unsigned long start,\n-\t\t\t\t\t unsigned long end, int node,\n-\t\t\t\t\t struct dev_pagemap *pgmap);\n #endif /* __ASSEMBLER__ */\n #endif\ndiff --git a/arch/powerpc/mm/book3s64/radix_pgtable.c b/arch/powerpc/mm/book3s64/radix_pgtable.c\nindex 568500343e5f..21fece355fbb 100644\n--- a/arch/powerpc/mm/book3s64/radix_pgtable.c\n+++ b/arch/powerpc/mm/book3s64/radix_pgtable.c\n@@ -1109,7 +1109,10 @@ static inline pte_t *vmemmap_pte_alloc(pmd_t *pmdp, int node,\n \treturn pte_offset_kernel(pmdp, address);\n }\n \n-\n+static int __meminit vmemmap_populate_compound_pages(unsigned long start_pfn,\n+\t\t\t\t\t\t unsigned long start,\n+\t\t\t\t\t\t unsigned long end, int node,\n+\t\t\t\t\t\t struct dev_pagemap *pgmap);\n \n int __meminit radix__vmemmap_populate(unsigned long start, unsigned long end, int node,\n \t\t\t\t struct vmem_altmap *altmap, struct dev_pagemap *pgmap)\n@@ -1122,6 +1125,9 @@ int __meminit radix__vmemmap_populate(unsigned long start, unsigned long end, in\n \tpmd_t *pmd;\n \tpte_t *pte;\n \n+\tif (vmemmap_can_optimize(altmap, pgmap))\n+\t\treturn vmemmap_populate_compound_pages(page_to_pfn((struct page *)start),\n+\t\t\t\t\t\t start, end, node, pgmap);\n \t/*\n \t * If altmap is present, Make sure we align the start vmemmap addr\n \t * to PAGE_SIZE so that we calculate the correct start_pfn in\n@@ -1303,10 +1309,10 @@ static pte_t * __meminit vmemmap_compound_tail_page(unsigned long addr,\n \treturn pte;\n }\n \n-int __meminit vmemmap_populate_compound_pages(unsigned long start_pfn,\n-\t\t\t\t\t unsigned long start,\n-\t\t\t\t\t unsigned long end, int node,\n-\t\t\t\t\t struct dev_pagemap *pgmap)\n+static int __meminit vmemmap_populate_compound_pages(unsigned long start_pfn,\n+\t\t\t\t\t\t unsigned long start,\n+\t\t\t\t\t\t unsigned long end, int node,\n+\t\t\t\t\t\t struct dev_pagemap *pgmap)\n {\n \t/*\n \t * we want to map things as base page size mapping so that\ndiff --git a/mm/sparse-vmemmap.c b/mm/sparse-vmemmap.c\nindex f5245647afee..7f684ed3479e 100644\n--- a/mm/sparse-vmemmap.c\n+++ b/mm/sparse-vmemmap.c\n@@ -296,10 +296,16 @@ static int __meminit vmemmap_populate_range(unsigned long start,\n \treturn 0;\n }\n \n+static int __meminit vmemmap_populate_compound_pages(unsigned long start,\n+\t\t\t\t\t\t unsigned long end, int node,\n+\t\t\t\t\t\t struct dev_pagemap *pgmap);\n+\n int __meminit vmemmap_populate_basepages(unsigned long start, unsigned long end,\n \t\t\t\t\t int node, struct vmem_altmap *altmap,\n \t\t\t\t\t struct dev_pagemap *pgmap)\n {\n+\tif (vmemmap_can_optimize(altmap, pgmap))\n+\t\treturn vmemmap_populate_compound_pages(start, end, node, pgmap);\n \treturn vmemmap_populate_range(start, end, node, altmap, -1, 0);\n }\n \n@@ -411,6 +417,9 @@ int __meminit vmemmap_populate_hugepages(unsigned long start, unsigned long end,\n \tpud_t *pud;\n \tpmd_t *pmd;\n \n+\tif (vmemmap_can_optimize(altmap, pgmap))\n+\t\treturn vmemmap_populate_compound_pages(start, end, node, pgmap);\n+\n \tfor (addr = start; addr < end; addr = next) {\n \t\tnext = pmd_addr_end(addr, end);\n \n@@ -453,7 +462,6 @@ int __meminit vmemmap_populate_hugepages(unsigned long start, unsigned long end,\n \treturn 0;\n }\n \n-#ifndef vmemmap_populate_compound_pages\n /*\n * For compound pages bigger than section size (e.g. x86 1G compound\n * pages with 2M subsection size) fill the rest of sections as tail\n@@ -491,14 +499,14 @@ static pte_t * __meminit compound_section_tail_page(unsigned long addr)\n \treturn pte;\n }\n \n-static int __meminit vmemmap_populate_compound_pages(unsigned long start_pfn,\n-\t\t\t\t\t\t unsigned long start,\n+static int __meminit vmemmap_populate_compound_pages(unsigned long start,\n \t\t\t\t\t\t unsigned long end, int node,\n \t\t\t\t\t\t struct dev_pagemap *pgmap)\n {\n \tunsigned long size, addr;\n \tpte_t *pte;\n \tint rc;\n+\tunsigned long start_pfn = page_to_pfn((struct page *)start);\n \n \tif (reuse_compound_section(start_pfn, pgmap)) {\n \t\tpte = compound_section_tail_page(start);\n@@ -544,26 +552,18 @@ static int __meminit vmemmap_populate_compound_pages(unsigned long start_pfn,\n \treturn 0;\n }\n \n-#endif\n-\n struct page * __meminit __populate_section_memmap(unsigned long pfn,\n \t\tunsigned long nr_pages, int nid, struct vmem_altmap *altmap,\n \t\tstruct dev_pagemap *pgmap)\n {\n \tunsigned long start = (unsigned long) pfn_to_page(pfn);\n \tunsigned long end = start + nr_pages * sizeof(struct page);\n-\tint r;\n \n \tif (WARN_ON_ONCE(!IS_ALIGNED(pfn, PAGES_PER_SUBSECTION) ||\n \t\t!IS_ALIGNED(nr_pages, PAGES_PER_SUBSECTION)))\n \t\treturn NULL;\n \n-\tif (vmemmap_can_optimize(altmap, pgmap))\n-\t\tr = vmemmap_populate_compound_pages(pfn, start, end, nid, pgmap);\n-\telse\n-\t\tr = vmemmap_populate(start, end, nid, altmap, pgmap);\n-\n-\tif (r < 0)\n+\tif (vmemmap_populate(start, end, nid, altmap, pgmap))\n \t\treturn NULL;\n \n \treturn pfn_to_page(pfn);\n","prefixes":["v2","5/6"]}