From patchwork Tue Mar 11 12:25:07 2025 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Xu Lu X-Patchwork-Id: 14011782 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 5D221C282EC for ; Tue, 11 Mar 2025 12:41:57 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To: Message-Id:Date:Subject:Cc:To:From:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=ke0fTkZ3V0vWgv4w//PTnCcvOBf1dVU3B0puwV4WJT4=; b=KrTOjcPrHWu4el uIR+2bzQJOgl9ll09k0Bk0rEkuw+wTQqhVf14cU3hMhg3hcozJZTxD/19MxVTc75kqwJsEMgwA9Qn 3LRNKw5GsLE21LAY04vPzqDaKWf26lv2L94GqXtS0MEigvu2AcFGpc6ZOB3EjfhqopWmboRPeoBkV EoTYqvpCdx1H685jFaeoZi+rMw7Txb24kJsUWUolMoTwq1CntltrURDBpXGhQinRlN26vdHH1TIXN 3PjEIEHHwo08L0ZyRwuuFkhlMiaC+OBTkhuXgROfr9hIILVzT80EfuzdPkOuU89y3N+rrjvib2w4N K+Ki/Bi4VeUJfyJnWDmQ==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.98 #2 (Red Hat Linux)) id 1tryvX-00000005gwv-3toi; Tue, 11 Mar 2025 12:41:43 +0000 Received: from mail-pl1-x633.google.com ([2607:f8b0:4864:20::633]) by bombadil.infradead.org with esmtps (Exim 4.98 #2 (Red Hat Linux)) id 1tryfj-00000005d6j-1Wa3 for linux-riscv@lists.infradead.org; Tue, 11 Mar 2025 12:25:24 +0000 Received: by mail-pl1-x633.google.com with SMTP id d9443c01a7336-22359001f1aso127549345ad.3 for ; Tue, 11 Mar 2025 05:25:23 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=bytedance.com; s=google; t=1741695922; x=1742300722; darn=lists.infradead.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=8OMb2TInCE9Qb7KUVvNgWIsY4Uxx9EaJDiVCyGMWyMw=; b=fHi/f5ie5g+Raron2UT8ctucRsTpWdXbh5bc78KgqgOZeoH3xy3CojPxpdDKWCNIgO G8DkRn71T1IJlSWckodvIrm2EIGNV2hNQPuIejIW01GnueNgv0mBcMFlla3DR0A9L7Db aDN7g2F+IbNPw3JIZNImY6ZwUI3H9Tz418bX6ZTO0yejSS4ZQyTM2fat9fr7iQOo9dr4 RyjSqLWdWxaQjNbZqLHCO4Tm6NGZLFIVYD0DP7GuGqqCoAKmg1jZEwIqaB3EAR19Mrdk zJlC01ZSGIn7je9gZqgR35OOTwjcLA51CojNC94J6SXnC9I8ijxpWRdbfOseW3fft9rX QPHQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1741695922; x=1742300722; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=8OMb2TInCE9Qb7KUVvNgWIsY4Uxx9EaJDiVCyGMWyMw=; b=L8fGAmz6H3ZzogJ1K5DibUlk3wmjGdbzCBvsdezjdInyHexVfNSuO7nwUCsEhy8dLt Wna7ixqr074NDq6Lfx+PkrIurJgk4kQW4VAVzlGk1CqCyhFdybY0U7X/GBec3iyNbB3p GHcr6ODhdY07dQEpQyN/i427LMdqfcGEjUpo/00zXC7wZi3XMx4qwwbuZPOdzAchglGf Elu/TsmMS2KZdmcl5rrKIlL6PKGsIpF58rWPSGgKVc4fzrwS9/6NFfFGRX4LWIrpmXGm ODfTjHevEHmefJHLemTrO82FRAhcxbKYrRm0ZIQzDrSvxZRQqxjIJHgPt0dohY+uX9iz bcKg== X-Forwarded-Encrypted: i=1; AJvYcCUH259Tzytq4XKRHe1bxrtBE03ldhg+0TNehWPth4YO6v3Fdj2tpS9y4wkqkn/Z1pi05A0lpe2V7qssnQ==@lists.infradead.org X-Gm-Message-State: AOJu0YzwUQg2/SKkGtZmvSs17hXRKzusw6MgiTyO6mO3rTqX5LaivGNI lPzRkc35UgUFpkYZ/HeBkZIxZ3mBHO0O3RtkebZKxBDpb9Y8vYm6GxCPWe9n1JI= X-Gm-Gg: ASbGncumZ8BDZyJIsGVGOaAAQYSH2SSHlAtos/mxB4/4T7n6iEHuvVLhq8hTRuYKOnJ yHpiCY+EnZtMcvj6IZNqDWqPigxtX3G77n3FS0AecMB6CGXdkjYR8M3IUR1D3QidrKs3bIXzWf3 yfea/Y47ce8TKfAUPfYrOtyJ9ozeEGx2StY+w8p2KMR87Y6WfcrSkP1BOaPKnXw64F5Rfa27Nlv TKcTI1nMVj1eIxEpiCLVQHKnyaVcNB4JHiXB/VeZY14eMol3wSvPRRS1N/IOAgGTU617zD3PiqU Ng6zNd8oiaSCUv9RetjwJRZcQ9GRaEtrhPgdGw1nZhygMqdAjzo7P9HgRdI9bVZP2qDA/UIvgmV HgQiNYZzXVGxhF7eFxzIxrO88ajc= X-Google-Smtp-Source: AGHT+IF+AKaM7+LGh1MSeSzuaMnYSTp9Pfd35y7XVQv+2e9hDZ72MVIWeURa7iA/fxUl+/8M2l45Ow== X-Received: by 2002:a05:6a20:c891:b0:1f5:59e5:8ada with SMTP id adf61e73a8af0-1f559e58e8dmr18738163637.4.1741695922464; Tue, 11 Mar 2025 05:25:22 -0700 (PDT) Received: from J9GPGXL7NT.bytedance.net ([61.213.176.56]) by smtp.gmail.com with ESMTPSA id d2e1a72fcca58-73698450244sm10226621b3a.80.2025.03.11.05.25.18 (version=TLS1_3 cipher=TLS_CHACHA20_POLY1305_SHA256 bits=256/256); Tue, 11 Mar 2025 05:25:22 -0700 (PDT) From: Xu Lu To: akpm@linux-foundation.org, jhubbard@nvidia.com, kirill.shutemov@linux.intel.com, tjeznach@rivosinc.com, joro@8bytes.org, will@kernel.org, robin.murphy@arm.com Cc: lihangjing@bytedance.com, xieyongji@bytedance.com, linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org, Xu Lu Subject: [PATCH v2 1/4] mm/gup: Add huge pte handling logic in follow_page_pte() Date: Tue, 11 Mar 2025 20:25:07 +0800 Message-Id: <20250311122510.72934-2-luxu.kernel@bytedance.com> X-Mailer: git-send-email 2.39.5 (Apple Git-154) In-Reply-To: <20250311122510.72934-1-luxu.kernel@bytedance.com> References: <20250311122510.72934-1-luxu.kernel@bytedance.com> MIME-Version: 1.0 X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20250311_052523_414238_D4D68E71 X-CRM114-Status: GOOD ( 14.78 ) X-BeenThere: linux-riscv@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-riscv" Errors-To: linux-riscv-bounces+linux-riscv=archiver.kernel.org@lists.infradead.org Page mapped at pte level can also be huge page when ARM CONT_PTE or RISC-V SVNAPOT is applied. Lack of huge pte handling logic in follow_page_pte() may lead to both performance and correctness issues. For example, on RISC-V platform, pages in the same 64K huge page have the same pte value, which means follow_page_pte() will get the same page for all of them using pte_pfn(). Then __get_user_pages() will return an array of pages with the same pfn. Mapping these pages causes memory confusion. This error can be triggered by the following code: void *addr = mmap(NULL, 0x10000, PROT_READ | PROT_WRITE, MAP_ANONYMOUS | MAP_PRIVATE | MAP_HUGETLB | MAP_HUGE_64KB, -1, 0); struct vfio_iommu_type1_dma_map dmap_map = { .argsz = sizeof(dma_map), .flags = VFIO_DMA_MAP_FLAG_READ | VFIO_DMA_MAP_FLAG_WRITE, .vaddr = (uint64_t)addr, .size = 0x10000, }; ioctl(vfio_container_fd, VFIO_IOMMU_MAP_DMA, &dma_map); This commit supplies huge pte handling logic in follow_page_pte() to avoid such problems. Signed-off-by: Xu Lu --- arch/riscv/include/asm/pgtable.h | 6 ++++++ include/linux/pgtable.h | 8 ++++++++ mm/gup.c | 17 +++++++++++------ 3 files changed, 25 insertions(+), 6 deletions(-) diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h index 050fdc49b5ad7..40ae5979dd82c 100644 --- a/arch/riscv/include/asm/pgtable.h +++ b/arch/riscv/include/asm/pgtable.h @@ -800,6 +800,12 @@ static inline bool pud_user_accessible_page(pud_t pud) #endif #ifdef CONFIG_TRANSPARENT_HUGEPAGE +#define pte_trans_huge pte_trans_huge +static inline int pte_trans_huge(pte_t pte) +{ + return pte_huge(pte) && pte_napot(pte); +} + static inline int pmd_trans_huge(pmd_t pmd) { return pmd_leaf(pmd); diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h index 94d267d02372e..3f57ee6dcf017 100644 --- a/include/linux/pgtable.h +++ b/include/linux/pgtable.h @@ -1584,6 +1584,14 @@ static inline unsigned long my_zero_pfn(unsigned long addr) #ifdef CONFIG_MMU +#if (defined(CONFIG_TRANSPARENT_HUGEPAGE) && !defined(pte_trans_huge)) || \ + (!defined(CONFIG_TRANSPARENT_HUGEPAGE)) +static inline int pte_trans_huge(pte_t pte) +{ + return 0; +} +#endif + #ifndef CONFIG_TRANSPARENT_HUGEPAGE static inline int pmd_trans_huge(pmd_t pmd) { diff --git a/mm/gup.c b/mm/gup.c index 3883b307780ea..67981ee28df86 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -838,7 +838,7 @@ static inline bool can_follow_write_pte(pte_t pte, struct page *page, static struct page *follow_page_pte(struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, - struct dev_pagemap **pgmap) + struct follow_page_context *ctx) { struct mm_struct *mm = vma->vm_mm; struct folio *folio; @@ -879,8 +879,8 @@ static struct page *follow_page_pte(struct vm_area_struct *vma, * case since they are only valid while holding the pgmap * reference. */ - *pgmap = get_dev_pagemap(pte_pfn(pte), *pgmap); - if (*pgmap) + ctx->pgmap = get_dev_pagemap(pte_pfn(pte), ctx->pgmap); + if (ctx->pgmap) page = pte_page(pte); else goto no_page; @@ -940,6 +940,11 @@ static struct page *follow_page_pte(struct vm_area_struct *vma, */ folio_mark_accessed(folio); } + if (is_vm_hugetlb_page(vma) || pte_trans_huge(pte)) { + ctx->page_mask = (1 << folio_order(folio)) - 1; + page = folio_page(folio, 0) + + ((address & (folio_size(folio) - 1)) >> PAGE_SHIFT); + } out: pte_unmap_unlock(ptep, ptl); return page; @@ -975,7 +980,7 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma, return no_page_table(vma, flags, address); } if (likely(!pmd_leaf(pmdval))) - return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap); + return follow_page_pte(vma, address, pmd, flags, ctx); if (pmd_protnone(pmdval) && !gup_can_follow_protnone(vma, flags)) return no_page_table(vma, flags, address); @@ -988,14 +993,14 @@ static struct page *follow_pmd_mask(struct vm_area_struct *vma, } if (unlikely(!pmd_leaf(pmdval))) { spin_unlock(ptl); - return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap); + return follow_page_pte(vma, address, pmd, flags, ctx); } if (pmd_trans_huge(pmdval) && (flags & FOLL_SPLIT_PMD)) { spin_unlock(ptl); split_huge_pmd(vma, pmd, address); /* If pmd was left empty, stuff a page table in there quickly */ return pte_alloc(mm, pmd) ? ERR_PTR(-ENOMEM) : - follow_page_pte(vma, address, pmd, flags, &ctx->pgmap); + follow_page_pte(vma, address, pmd, flags, ctx); } page = follow_huge_pmd(vma, address, pmd, flags, ctx); spin_unlock(ptl);