From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by smtp.subspace.kernel.org (Postfix) with ESMTP id 6F69F606A33D for ; Sat, 21 Nov 2026 01:09:22 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.a=rsa-sha256 header.s=k20260515 header.b=iS5n6k7R Received: by kanga.kvack.org (Postfix) id 961226B05B6; Fri, 20 Nov 2026 20:09:20 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 911AB6B05B7; Fri, 20 Nov 2026 20:09:20 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 8003F6B05B8; Fri, 20 Nov 2026 20:09:20 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0012.hostedemail.com [216.40.44.12]) by kanga.kvack.org (Postfix) with ESMTP id 5478F6B05B6 for ; Fri, 20 Nov 2026 20:09:20 -0500 (EST) Received: from smtpin12.hostedemail.com (lb01a-stub [10.200.18.249]) by unirelay03.hostedemail.com (Postfix) with ESMTP id B3C62A02A5 for ; Sat, 21 Nov 2026 01:09:19 +0000 (UTC) X-FDA: 85457342838.12.B256961 Received: from tor.source.kernel.org (tor.source.kernel.org [172.105.4.254]) by imf29.hostedemail.com (Postfix) with ESMTP id 1ECF2120003 for ; Sat, 21 Nov 2026 01:09:18 +0000 (UTC) Authentication-Results: imf29.hostedemail.com; dkim=pass header.d=kernel.org header.s=k20260515 header.b=iS5n6k7R; dmarc=pass (policy=quarantine) header.from=kernel.org; spf=pass (imf29.hostedemail.com: domain of sashiko-bot@kernel.org designates 172.105.4.254 as permitted sender) smtp.mailfrom=sashiko-bot@kernel.org ARC-Seal: i=1; a=rsa-sha256; d=hostedemail.com; s=arc-20220608; cv=none; t=1795223358; b=x7p6B9QHYSW34DATVqdUWqKrff00G4dXlhNdckrytZ254t/zAcyU3az/wuUfBtaFlADxLa cZG2MGZywr2s7y2P6xmfBKkVxGnxBmxoy9o+c7tRSYN/UaIsfm/seMU6lICnZ9mu8TFnnp jtEJVz+uFijM5NZoUeKwHrVArvojdpI= ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1795223358; h=from:from:sender:reply-to:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version: content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=ZUQvwLmTlji/bJ0Fu6NxTY+CyxNdQLCUrKCBR4Qtz8g=; b=0Dd2wYeWhplzIzzwvbwb/xik2WUaOxMU6mZAjfpV6BSt11rx9PE1OGEMLYJ0vlGUyc/zDK gOZ6fOOtM5nIObA79cnlT9jAxvlOgBoVAKqWDfSVEIOyqllhby6AnxUg23Yrs3bDjZjSYv SS5fML2YB/qg+OoFS/SXImKDADPOTk4= ARC-Authentication-Results: i=1; imf29.hostedemail.com; dkim=pass header.d=kernel.org header.s=k20260515 header.b=iS5n6k7R; dmarc=pass (policy=quarantine) header.from=kernel.org; spf=pass (imf29.hostedemail.com: domain of sashiko-bot@kernel.org designates 172.105.4.254 as permitted sender) smtp.mailfrom=sashiko-bot@kernel.org Received: from smtp.kernel.org (quasi.space.kernel.org [100.103.45.18]) by tor.source.kernel.org (Postfix) with ESMTP id 43D09612FF; Sat, 21 Nov 2026 01:09:17 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id C10D21F000CA; Sat, 21 Nov 2026 01:09:14 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1795223356; bh=ZUQvwLmTlji/bJ0Fu6NxTY+CyxNdQLCUrKCBR4Qtz8g=; h=From:Subject:Reply-To:To:Cc:In-Reply-To:References:Date; b=iS5n6k7RgsLK3C+3X//Rnclg2fGxJ+m4s4+WvjH08tvjB3e6IKKI0QG/dNfUu5Y5L gGy4rVLMArWJbfdocdDDmgZ646AvUwwwlx3uXAj5xDsIzM076+xy546eg+kG7u0u/K AhhLqZDFtsNXgXQpqcono3nQcDBiBjBl/B3rC8Bw2OpMTh4u1Dc2LeevIE8NpSKHJh ch7Cmmcva0wgFbKjfSaZ3JzJExgMhzBaT7T7d8qhR+dQ3APm6ROFJNomXxkee/uZJb VZ/35DL4NXAMUHJRrv8WHt6bmQHSuArS/Mkz77ez9D903jrgVMzYQ0ae6zdPmVUXLe yKfIbdEUqEJ/g== Message-ID: From: sashiko-bot@kernel.org Subject: Re: [RFC PATCH v3 01/25] mm: Introduce CONFIG_PISA and metadata table Reply-To: sashiko-reviews@lists.linux.dev To: "Luigi Rizzo" , "Andrew Morton" , "David Hildenbrand" , "Johannes Weiner" , "Vlastimil Babka" , "Joerg Roedel" , "Will Deacon" , "Robin Murphy" , "Marek Szyprowski" , "Jens Axboe" , "David S . Miller" , "Eric Dumazet" , "Jakub Kicinski" , "Paolo Abeni" , "Luigi Rizzo" Cc: "Lorenzo Stoakes" , "Liam R . Howlett" , "Mike Rapoport" , "Suren Baghdasaryan" , "Michal Hocko" , "Brendan Jackman" , "Zi Yan" , "Harry Yoo" , "Christoph Lameter" , "David Rientjes" , "Roman Gushchin" , "Hao Li" , "Matthew Wilcox" , "Jan Kara" , "Jann Horn" , "Catalin Marinas" , "Greg Kroah-Hartman" , "Rafael J . Wysocki" , "Danilo Krummrich" , "Jonathan Corbet" , "Simon Horman" , "Jesper Dangaard Brouer" , "Ilias Apalodimas" , "Willem de Bruijn" , "Kuniyuki Iwashima" , "Joshua Washington" , "Harshitha Ramamurthy" , "Alexander Lobakin" , "Christoph Hellwig" , "Jason Gunthorpe" , "Dragos Tatulea" , "Naman Gulati" , "Pranjal Shrivastava" , "Mostafa Saleh" , "Randy Dunlap" , linux-mm@kvack.org In-Reply-To: <20261120010655.467825-2-lrizzo@google.com> References: <20261109182533.3594984-1-lrizzo@google.com> <20261120010655.467825-1-lrizzo@google.com> <20261120010655.467825-2-lrizzo@google.com> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Date: Sat, 21 Nov 2026 01:09:14 +0000 X-Rspam-User: X-Rspamd-Server: rspam12 X-Spam-Status: No, score=-10.8 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_PASS,SPF_PASS autolearn=ham autolearn_force=no version=4.0.1 X-Rspamd-Queue-Id: 1ECF2120003 X-Stat-Signature: tw1n3hm6epzzopdx8hfqajww693yt31z X-HE-Tag: 1795223357-906504 X-HE-Meta: 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 bAi8TlFM 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 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: List-Subscribe: X-Spam-Checker-Version: SpamAssassin 4.0.1 (2024-03-26) on aws-us-west-2-korg-listserv-alma10-1 X-Archived-By: Thank you for your contribution! Sashiko AI review found 3 potential issues to consider: - [Medium] mm: unmapped KMSAN vmalloc shadow in pisa_pfn2slot and pisa_meta_array - [Medium] mm: freeing init_mm page tables in pisa_meta_init() without init_mm.mmap_lock - [Low] mm: early pisa_enabled() return in pisa_meta_init() during jump_label_update() -- commit e2c1da1898da520bb5d1c1e05212fb580b78c49f Author: Luigi Rizzo mm: Introduce CONFIG_PISA and metadata table This commit adds CONFIG_PISA, the two-level metadata tables (pisa_pfn2slot and pisa_meta_array) with on-demand page table management and memory hotplug support, and a built-in KUnit test suite. > diff --git a/mm/pisa-meta.c b/mm/pisa-meta.c > new file mode 100644 > index 0000000000000..734f49320047a > --- /dev/null > +++ b/mm/pisa-meta.c > @@ -0,0 +1,603 @@ [ ... ] > +static int pisa_pfn2slot_init_zero(unsigned long start) > +{ > + unsigned long end =3D start + PISA_PFN2SLOT_SIZE, addr; > + int ret =3D -ENOMEM; > + > + for (addr =3D start; addr < end; addr +=3D PUD_SIZE) { > + pgd_t *pgd =3D pgd_offset_k(addr); > + p4d_t *p4d =3D p4d_alloc(&init_mm, pgd, addr); > + pud_t *pud =3D p4d ? pud_alloc(&init_mm, p4d, addr) : NULL; > + > + if (!pud) > + goto undo; > + /* An earlier user of this vmalloc range may have left page tables. */ > + if (!pud_none(pudp_get(pud)) && !pud_free_pmd_page(pud, addr)) { > + ret =3D -EBUSY; > + goto undo; > + } > + pud_populate(&init_mm, pud, pisa_zero_pmds); > + } > + return 0; [ ... ] > +static void pisa_pfn2slot_free(struct vm_struct *vm) > +{ > + unsigned long start =3D ALIGN((unsigned long)vm->addr, PUD_SIZE); > + unsigned long end =3D start + PISA_PFN2SLOT_SIZE; > + > + for (unsigned long addr =3D start; addr < end; addr +=3D PUD_SIZE) { > + pgd_t *pgd =3D pgd_offset_k(addr); > + p4d_t *p4d =3D pgd_none(pgdp_get(pgd)) ? NULL : p4d_offset(pgd, addr); > + pud_t *pud =3D (!p4d || p4d_none(p4dp_get(p4d))) ? NULL : pud_offset(p= 4d, addr); > + pmd_t *pmds; > + > + if (!pud || pud_none(pudp_get(pud))) > + continue; > + pmds =3D pud_pgtable(pudp_get(pud)); > + pud_clear(pud); > + if (pmds =3D=3D pisa_zero_pmds) > + continue; > + > + for (unsigned int i =3D 0; i < PTRS_PER_PMD; i++) { > + pte_t *ptes =3D (pte_t *)pmd_page_vaddr(pmdp_get(&pmds[i])); > + > + if (ptes =3D=3D pisa_zero_ptes) > + continue; > + for (unsigned int j =3D 0; j < PTRS_PER_PTE; j++) { > + pte_t pte =3D ptep_get(&ptes[j]); > + > + if (pte_write(pte)) > + __free_page(pte_page(pte)); > + } > + pte_free_kernel(&init_mm, ptes); > + } > + pmd_free(&init_mm, pmds); > + } > + free_vm_area(vm); > +} [Severity: Medium] Can freeing lower-level init_mm page tables in pisa_pfn2slot_init_zero() and pisa_pfn2slot_free() without holding init_mm.mmap_lock cause a use-after-free during a concurrent ptdump walk? ptdump_walk_pgd() walks init_mm page tables under mmap_write_lock(&init_mm), and vmalloc paths such as vmap_try_huge_pud() take mmap_read_lock(&init_mm) around pud_free_pmd_page(). Because pisa_meta_init() runs at runtime without init_mm.mmap_lock, a concurrent /sys/kernel/debug/page_tables/kernel read can race with table teardown: CPU0 CPU1 ptdump_walk_pgd(&init_mm) pisa_meta_init() mmap_write_lock(&init_mm) pisa_pfn2slot_init_zero() walk_pud_level() pud_free_pmd_page(pud, addr) walk_pmd_level() pud_clear(pud) free_pte_table() pmd_free(&init_mm, pmd_table) walk_pte_level() <-- use-after-free The same race applies if pisa_meta_init() fails and pisa_pfn2slot_free() frees private PMD and PTE tables via pte_free_kernel() and pmd_free(). [ ... ] > + /* Lazily back the PTE page covering this 16-slot chunk of @pisa_meta_a= rray. */ > + meta =3D &pisa_meta_array[slot]; > + ptep =3D pisa_meta_pte((unsigned long)meta); > + if (pte_none(ptep_get(ptep))) { > + unsigned long addr =3D (unsigned long)meta & PAGE_MASK; > + > + if (!new_page) { > + /* Allocate outside the lock, this opens a race so retry. */ > + spin_unlock_irqrestore(&pisa_slot_lock, flags); > + new_page =3D alloc_pages_node(pfn_to_nid(pfn), gfp, 0); > + if (new_page) > + goto retry; > + return -ENOMEM; > + } > + set_pte_at(&init_mm, addr, ptep, pfn_pte(page_to_pfn(new_page), PAGE_K= ERNEL)); > + flush_cache_vmap(addr, addr + PAGE_SIZE); > + new_page =3D NULL; > + } > + > + memset(meta, 0, sizeof(*meta)); [Severity: Medium] Does accessing pisa_meta_array or pisa_pfn2slot fault on unmapped KMSAN vmalloc shadow memory when CONFIG_KMSAN=3Dy? pisa_meta_init() reserves pfn2slot_vm and meta_vm from the vmalloc range via get_vm_area(), and pisa_pfn2slot_init_zero(), pisa_pfn2slot_populate(), and pisa_meta_ensure_pfn() install page table entries directly with pud_populate() and set_pte_at() instead of going through vmap_pages_range_noflush() -> kmsan_vmap_pages_range_noflush(). Because CONFIG_PISA is enabled by default on x86_64 without depending on !KMSAN, KMSAN-instrumented accesses such as memset(meta, 0, sizeof(*meta)) or reads from pisa_pfn2slot translate the vmalloc address into KMSAN_VMALLOC_SHADOW_START / KMSAN_VMALLOC_ORIGIN_START, where no shadow or origin pages are mapped, causing an unhandled page fault during the built-in KUnit test at boot. [ ... ] > +/* Reserve @pisa_pfn2slot and @pisa_meta_array and populate page tables.= */ > +int pisa_meta_init(void) > +{ > + struct vm_struct *pfn2slot_vm, *meta_vm; > + unsigned long meta_size; > + unsigned long *bitmap; > + struct pisa_meta *meta; > + u32 *pfn2slot; > + static int pisa_init_err; > + int ret =3D 0; > + > + if (pisa_enabled()) > + return 0; > + > + mutex_lock(&pisa_meta_mutex); > + if (pisa_init_err) { > + /* A previous init attempt failed. */ > + mutex_unlock(&pisa_meta_mutex); > + return pisa_init_err; > + } > + if (pisa_pfn2slot) { > + /* The previous caller completed init while we waited on the mutex. */ > + mutex_unlock(&pisa_meta_mutex); > + goto out_enable; > + } [ ... ] > + pr_info("pisa: %lu GB pfn2slot KVA, %u slots (%lu KB meta KVA)\n", > + PISA_PFN2SLOT_SIZE >> 30, pisa_nr_slots, meta_size >> 10); > + mutex_unlock(&pisa_meta_mutex); > +out_enable: > + static_branch_enable(&pisa_key); > + return 0; [Severity: Low] Can a concurrent caller of pisa_meta_init() return 0 from the early pisa_enabled() check while static_branch_enable(&pisa_key) is still patching other jump sites? jump_label_update() patches jump entries one by one before setting key->enabled to 1: CPU0 CPU1 pisa_meta_init() ... mutex_unlock(&pisa_meta_mutex) static_branch_enable(&pisa_key) jump_label_update() /* patches pisa_meta_init() site */ pisa_meta_init() if (pisa_enabled()) return 0; pisa_meta_ensure_pfn() if (!pisa_enabled()) return -ERANGE; <-- unpatched si= te /* patches pisa_meta_ensure_pfn() site */ Testing static_key_enabled(&pisa_key) for the fast path in pisa_meta_init() would wait until jump_label_update() finishes patching all call sites before returning early. --=20 Sashiko AI review =C2=B7 https://sashiko.dev/#/patchset/20261120010655.4678= 25-1-lrizzo@google.com?part=3D1