Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754840AbaF3Lxu (ORCPT ); Mon, 30 Jun 2014 07:53:50 -0400 Received: from mta-out1.inet.fi ([62.71.2.199]:39242 "EHLO jenni2.inet.fi" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752372AbaF3Lxc (ORCPT ); Mon, 30 Jun 2014 07:53:32 -0400 Date: Mon, 30 Jun 2014 14:53:11 +0300 From: "Kirill A. Shutemov" To: Naoya Horiguchi Cc: linux-mm@kvack.org, Andrew Morton , Dave Hansen , Hugh Dickins , linux-kernel@vger.kernel.org, Naoya Horiguchi Subject: Re: [PATCH v3 02/13] pagewalk: improve vma handling Message-ID: <20140630115311.GR19833@node.dhcp.inet.fi> References: <1403295099-6407-1-git-send-email-n-horiguchi@ah.jp.nec.com> <1403295099-6407-3-git-send-email-n-horiguchi@ah.jp.nec.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1403295099-6407-3-git-send-email-n-horiguchi@ah.jp.nec.com> User-Agent: Mutt/1.5.22.1 (2013-10-16) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Jun 20, 2014 at 04:11:28PM -0400, Naoya Horiguchi wrote: > Current implementation of page table walker has a fundamental problem > in vma handling, which started when we tried to handle vma(VM_HUGETLB). > Because it's done in pgd loop, considering vma boundary makes code > complicated and bug-prone. > > From the users viewpoint, some user checks some vma-related condition to > determine whether the user really does page walk over the vma. > > In order to solve these, this patch moves vma check outside pgd loop and > introduce a new callback ->test_walk(). > > ChangeLog v3: > - drop walk->skip control > > Signed-off-by: Naoya Horiguchi > --- > include/linux/mm.h | 15 +++- > mm/pagewalk.c | 198 ++++++++++++++++++++++++++++++----------------------- > 2 files changed, 126 insertions(+), 87 deletions(-) > > diff --git v3.16-rc1.orig/include/linux/mm.h v3.16-rc1/include/linux/mm.h > index c5cb6394e6cb..489a63a06a4a 100644 > --- v3.16-rc1.orig/include/linux/mm.h > +++ v3.16-rc1/include/linux/mm.h > @@ -1107,10 +1107,16 @@ void unmap_vmas(struct mmu_gather *tlb, struct vm_area_struct *start_vma, > * @pte_entry: if set, called for each non-empty PTE (4th-level) entry > * @pte_hole: if set, called for each hole at all levels > * @hugetlb_entry: if set, called for each hugetlb entry > - * *Caution*: The caller must hold mmap_sem() if @hugetlb_entry > - * is used. > + * @test_walk: caller specific callback function to determine whether > + * we walk over the current vma or not. A positive returned > + * value means "do page table walk over the current vma," > + * and a negative one means "abort current page table walk > + * right now." 0 means "skip the current vma." > + * @mm: mm_struct representing the target process of page table walk > + * @vma: vma currently walked (NULL if walking outside vmas) > + * @private: private data for callbacks' usage > * > - * (see walk_page_range for more details) > + * (see the comment on walk_page_range() for more details) > */ > struct mm_walk { > int (*pmd_entry)(pmd_t *pmd, unsigned long addr, > @@ -1122,7 +1128,10 @@ struct mm_walk { > int (*hugetlb_entry)(pte_t *pte, unsigned long hmask, > unsigned long addr, unsigned long next, > struct mm_walk *walk); > + int (*test_walk)(unsigned long addr, unsigned long next, > + struct mm_walk *walk); > struct mm_struct *mm; > + struct vm_area_struct *vma; > void *private; > }; > > diff --git v3.16-rc1.orig/mm/pagewalk.c v3.16-rc1/mm/pagewalk.c > index 335690650b12..86d811202374 100644 > --- v3.16-rc1.orig/mm/pagewalk.c > +++ v3.16-rc1/mm/pagewalk.c > @@ -59,7 +59,7 @@ static int walk_pmd_range(pud_t *pud, unsigned long addr, unsigned long end, > continue; > > split_huge_page_pmd_mm(walk->mm, addr, pmd); > - if (pmd_none_or_trans_huge_or_clear_bad(pmd)) > + if (pmd_trans_unstable(pmd)) > goto again; > err = walk_pte_range(pmd, addr, next, walk); > if (err) > @@ -95,6 +95,32 @@ static int walk_pud_range(pgd_t *pgd, unsigned long addr, unsigned long end, > return err; > } > > +static int walk_pgd_range(unsigned long addr, unsigned long end, > + struct mm_walk *walk) > +{ > + pgd_t *pgd; > + unsigned long next; > + int err = 0; > + > + pgd = pgd_offset(walk->mm, addr); > + do { > + next = pgd_addr_end(addr, end); > + if (pgd_none_or_clear_bad(pgd)) { > + if (walk->pte_hole) > + err = walk->pte_hole(addr, next, walk); > + if (err) > + break; > + continue; > + } > + if (walk->pmd_entry || walk->pte_entry) > + err = walk_pud_range(pgd, addr, next, walk); > + if (err) > + break; > + } while (pgd++, addr = next, addr != end); > + > + return err; > +} > + > #ifdef CONFIG_HUGETLB_PAGE > static unsigned long hugetlb_entry_end(struct hstate *h, unsigned long addr, > unsigned long end) > @@ -103,10 +129,10 @@ static unsigned long hugetlb_entry_end(struct hstate *h, unsigned long addr, > return boundary < end ? boundary : end; > } > > -static int walk_hugetlb_range(struct vm_area_struct *vma, > - unsigned long addr, unsigned long end, > +static int walk_hugetlb_range(unsigned long addr, unsigned long end, > struct mm_walk *walk) > { > + struct vm_area_struct *vma = walk->vma; > struct hstate *h = hstate_vma(vma); > unsigned long next; > unsigned long hmask = huge_page_mask(h); > @@ -119,15 +145,14 @@ static int walk_hugetlb_range(struct vm_area_struct *vma, > if (pte && walk->hugetlb_entry) > err = walk->hugetlb_entry(pte, hmask, addr, next, walk); > if (err) > - return err; > + break; > } while (addr = next, addr != end); > > return 0; I guess it should be 'return err;', right? > } > > #else /* CONFIG_HUGETLB_PAGE */ > -static int walk_hugetlb_range(struct vm_area_struct *vma, > - unsigned long addr, unsigned long end, > +static int walk_hugetlb_range(unsigned long addr, unsigned long end, > struct mm_walk *walk) > { > return 0; > @@ -135,109 +160,114 @@ static int walk_hugetlb_range(struct vm_area_struct *vma, > > #endif /* CONFIG_HUGETLB_PAGE */ > > +/* > + * Decide whether we really walk over the current vma on [@start, @end) > + * or skip it via the returned value. Return 0 if we do walk over the > + * current vma, and return 1 if we skip the vma. Negative values means > + * error, where we abort the current walk. > + * > + * Default check (only VM_PFNMAP check for now) is used when the caller > + * doesn't define test_walk() callback. > + */ > +static int walk_page_test(unsigned long start, unsigned long end, > + struct mm_walk *walk) > +{ > + struct vm_area_struct *vma = walk->vma; > + > + if (walk->test_walk) > + return walk->test_walk(start, end, walk); > > + /* > + * Do not walk over vma(VM_PFNMAP), because we have no valid struct > + * page backing a VM_PFNMAP range. See also commit a9ff785e4437. > + */ > + if (vma->vm_flags & VM_PFNMAP) > + return 1; > + return 0; > +} > + > +static int __walk_page_range(unsigned long start, unsigned long end, > + struct mm_walk *walk) > +{ > + int err = 0; > + struct vm_area_struct *vma = walk->vma; > + > + if (vma && is_vm_hugetlb_page(vma)) { > + if (walk->hugetlb_entry) > + err = walk_hugetlb_range(start, end, walk); > + } else > + err = walk_pgd_range(start, end, walk); > + > + return err; > +} > > /** > - * walk_page_range - walk a memory map's page tables with a callback > - * @addr: starting address > - * @end: ending address > - * @walk: set of callbacks to invoke for each level of the tree > - * > - * Recursively walk the page table for the memory area in a VMA, > - * calling supplied callbacks. Callbacks are called in-order (first > - * PGD, first PUD, first PMD, first PTE, second PTE... second PMD, > - * etc.). If lower-level callbacks are omitted, walking depth is reduced. > + * walk_page_range - walk page table with caller specific callbacks > * > - * Each callback receives an entry pointer and the start and end of the > - * associated range, and a copy of the original mm_walk for access to > - * the ->private or ->mm fields. > + * Recursively walk the page table tree of the process represented by @walk->mm > + * within the virtual address range [@start, @end). During walking, we can do > + * some caller-specific works for each entry, by setting up pmd_entry(), > + * pte_entry(), and/or hugetlb_entry(). If you don't set up for some of these > + * callbacks, the associated entries/pages are just ignored. > + * The return values of these callbacks are commonly defined like below: > + * - 0 : succeeded to handle the current entry, and if you don't reach the > + * end address yet, continue to walk. > + * - >0 : succeeded to handle the current entry, and return to the caller > + * with caller specific value. > + * - <0 : failed to handle the current entry, and return to the caller > + * with error code. > * > - * Usually no locks are taken, but splitting transparent huge page may > - * take page table lock. And the bottom level iterator will map PTE > - * directories from highmem if necessary. > + * Before starting to walk page table, some callers want to check whether > + * they really want to walk over the current vma, typically by checking > + * its vm_flags. walk_page_test() and @walk->test_walk() are used for this > + * purpose. > * > - * If any callback returns a non-zero value, the walk is aborted and > - * the return value is propagated back to the caller. Otherwise 0 is returned. > + * struct mm_walk keeps current values of some common data like vma and pmd, > + * which are useful for the access from callbacks. If you want to pass some > + * caller-specific data to callbacks, @walk->private should be helpful. > * > - * walk->mm->mmap_sem must be held for at least read if walk->hugetlb_entry > - * is !NULL. > + * Locking: > + * Callers of walk_page_range() and walk_page_vma() should hold > + * @walk->mm->mmap_sem, because these function traverse vma list and/or > + * access to vma's data. > */ > -int walk_page_range(unsigned long addr, unsigned long end, > +int walk_page_range(unsigned long start, unsigned long end, > struct mm_walk *walk) > { > - pgd_t *pgd; > - unsigned long next; > int err = 0; > + unsigned long next; > > - if (addr >= end) > - return err; > + if (start >= end) > + return -EINVAL; > > if (!walk->mm) > return -EINVAL; > > VM_BUG_ON(!rwsem_is_locked(&walk->mm->mmap_sem)); > > - pgd = pgd_offset(walk->mm, addr); > do { > - struct vm_area_struct *vma = NULL; > + struct vm_area_struct *vma; > > - next = pgd_addr_end(addr, end); > + vma = find_vma(walk->mm, start); > + if (!vma) { /* after the last vma */ > + walk->vma = NULL; > + next = end; > + } else if (start < vma->vm_start) { /* outside the found vma */ > + walk->vma = NULL; > + next = vma->vm_start; Is there a reason why we shoul go for __walk_page_range() for these two cases if walkj->pte_hole() is not defined? Otherwise, looks okay. Acked-by: Kirill A. Shutemov -- Kirill A. Shutemov -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/