From mboxrd@z Thu Jan 1 00:00:00 1970 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 50DA2C369DC for ; Mon, 28 Apr 2025 13:47:17 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender:List-Subscribe:List-Help :List-Post:List-Archive:List-Unsubscribe:List-Id:Content-Transfer-Encoding: Content-Type:In-Reply-To:References:Cc:To:From:Subject:MIME-Version:Date: Message-ID:Reply-To:Content-ID:Content-Description:Resent-Date:Resent-From: Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=ZDpYjstdd/DhgHEufcBds9bwNo3Lzx1GtED2qA5h/CM=; b=AtIeDt+Xf2YPcHDtAfbNQm8SI6 rtZyuPMn8WIzVtQtOzFllW653Jy9k5Z2FTm5bHmRutq/auYTWBWVZHvfgDO5eiXAvOV1Et4Ooblhz ANTiDPbVDupPmpEFFfe5d3NaLG2MI/s/OIZt8dbzTju3OksHX69UcWp7avYrE+LS4DgtM1gR/A6dK LQZr4dUMrZYdNgova86BA1GENv2BmkCoJuNMvbv0mblHC6ff7hw1k5bzWERyZDwd/pJQFkeDSnrNq Mny196mkGjjPUZpQsTqCbJzfu7Z/FRZEgoOJCj60jjvTlyXHW5bg0qwZVl//jtW4Znbb9beRhsJdH Je3mF+bA==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.98.2 #2 (Red Hat Linux)) id 1u9Op6-00000006WxG-2GIM; Mon, 28 Apr 2025 13:47:05 +0000 Received: from out-188.mta0.migadu.com ([91.218.175.188]) by bombadil.infradead.org with esmtps (Exim 4.98.2 #2 (Red Hat Linux)) id 1u9OLp-00000006Qrg-1VHB for linux-arm-kernel@lists.infradead.org; Mon, 28 Apr 2025 13:16:54 +0000 Message-ID: <29c70c06-42c2-4bc0-a56e-443a1200fde0@linux.dev> DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.dev; s=key1; t=1745846196; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=ZDpYjstdd/DhgHEufcBds9bwNo3Lzx1GtED2qA5h/CM=; b=IRbRJrdeO0WfMYiAZuGndCEUziv3aIyli9DwyUUonpjfGgCWLCpIDGDqWqQdrOhRbsjA4D t8wNN1hTguxQA+36bugJrqOpBYh1m9X+HQIAQhQMMBjkALjwcztCFD80XfmQUmQ1x3oXSh mYWN98Eim6j2eqABxeIac0bJhrWG8Nk= Date: Mon, 28 Apr 2025 21:16:16 +0800 MIME-Version: 1.0 Subject: Re: [PATCH 6/7] mm: Batch around can_change_pte_writable() Content-Language: en-US X-Report-Abuse: Please report any abuse attempt to abuse@migadu.com and include these headers. From: Lance Yang To: Dev Jain , akpm@linux-foundation.org Cc: ryan.roberts@arm.com, david@redhat.com, willy@infradead.org, linux-mm@kvack.org, linux-kernel@vger.kernel.org, catalin.marinas@arm.com, will@kernel.org, Liam.Howlett@oracle.com, lorenzo.stoakes@oracle.com, vbabka@suse.cz, jannh@google.com, anshuman.khandual@arm.com, peterx@redhat.com, joey.gouly@arm.com, ioworker0@gmail.com, baohua@kernel.org, kevin.brodsky@arm.com, quic_zhenhuah@quicinc.com, christophe.leroy@csgroup.eu, yangyicong@hisilicon.com, linux-arm-kernel@lists.infradead.org, namit@vmware.com, hughd@google.com, yang@os.amperecomputing.com, ziy@nvidia.com References: <20250428120414.12101-1-dev.jain@arm.com> <20250428120414.12101-7-dev.jain@arm.com> In-Reply-To: Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-Migadu-Flow: FLOW_OUT X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20250428_061649_537029_237FD9CC X-CRM114-Status: GOOD ( 25.84 ) X-BeenThere: linux-arm-kernel@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-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org On 2025/4/28 20:50, Lance Yang wrote: > Hey Dev, > > On 2025/4/28 20:04, Dev Jain wrote: >> In preparation for patch 7, we need to properly batch around >> can_change_pte_writable(). We batch around pte_needs_soft_dirty_wp() by >> the corresponding fpb flag, we batch around the page-anon exclusive check >> using folio_maybe_mapped_shared(); modify_prot_start_ptes() collects the >> dirty and access bits across the batch, therefore batching across >> pte_dirty(): this is correct since the dirty bit on the PTE really >> is just an indication that the folio got written to, so even if >> the PTE is not actually dirty (but one of the PTEs in the batch is), >> the wp-fault optimization can be made. >> >> Signed-off-by: Dev Jain >> --- >>   include/linux/mm.h | 4 ++-- >>   mm/gup.c           | 2 +- >>   mm/huge_memory.c   | 4 ++-- >>   mm/memory.c        | 6 +++--- >>   mm/mprotect.c      | 9 ++++++--- >>   5 files changed, 14 insertions(+), 11 deletions(-) >> >> diff --git a/include/linux/mm.h b/include/linux/mm.h >> index 5eb0d77c4438..ffa02e15863f 100644 >> --- a/include/linux/mm.h >> +++ b/include/linux/mm.h >> @@ -2710,8 +2710,8 @@ int get_cmdline(struct task_struct *task, char >> *buffer, int buflen); >>   #define  MM_CP_UFFD_WP_ALL                 (MM_CP_UFFD_WP | \ >>                           MM_CP_UFFD_WP_RESOLVE) >> -bool can_change_pte_writable(struct vm_area_struct *vma, unsigned >> long addr, >> -                 pte_t pte); >> +bool can_change_ptes_writable(struct vm_area_struct *vma, unsigned >> long addr, >> +                 pte_t pte, struct folio *folio, unsigned int nr); >>   extern long change_protection(struct mmu_gather *tlb, >>                     struct vm_area_struct *vma, unsigned long start, >>                     unsigned long end, unsigned long cp_flags); >> diff --git a/mm/gup.c b/mm/gup.c >> index 84461d384ae2..6a605fc5f2cb 100644 >> --- a/mm/gup.c >> +++ b/mm/gup.c >> @@ -614,7 +614,7 @@ static inline bool can_follow_write_common(struct >> page *page, >>           return false; >>       /* >> -     * See can_change_pte_writable(): we broke COW and could map the >> page >> +     * See can_change_ptes_writable(): we broke COW and could map the >> page >>        * writable if we have an exclusive anonymous page ... >>        */ >>       return page && PageAnon(page) && PageAnonExclusive(page); >> diff --git a/mm/huge_memory.c b/mm/huge_memory.c >> index 28c87e0e036f..e5496c0d9e7e 100644 >> --- a/mm/huge_memory.c >> +++ b/mm/huge_memory.c >> @@ -2032,12 +2032,12 @@ static inline bool >> can_change_pmd_writable(struct vm_area_struct *vma, >>           return false; >>       if (!(vma->vm_flags & VM_SHARED)) { >> -        /* See can_change_pte_writable(). */ >> +        /* See can_change_ptes_writable(). */ >>           page = vm_normal_page_pmd(vma, addr, pmd); >>           return page && PageAnon(page) && PageAnonExclusive(page); >>       } >> -    /* See can_change_pte_writable(). */ >> +    /* See can_change_ptes_writable(). */ >>       return pmd_dirty(pmd); >>   } >> diff --git a/mm/memory.c b/mm/memory.c >> index b9e8443aaa86..b1fda3de8d27 100644 >> --- a/mm/memory.c >> +++ b/mm/memory.c >> @@ -750,7 +750,7 @@ static void restore_exclusive_pte(struct >> vm_area_struct *vma, >>           pte = pte_mkuffd_wp(pte); >>       if ((vma->vm_flags & VM_WRITE) && >> -        can_change_pte_writable(vma, address, pte)) { >> +        can_change_ptes_writable(vma, address, pte, NULL, 1)) { >>           if (folio_test_dirty(folio)) >>               pte = pte_mkdirty(pte); >>           pte = pte_mkwrite(pte, vma); >> @@ -5767,7 +5767,7 @@ static void numa_rebuild_large_mapping(struct >> vm_fault *vmf, struct vm_area_stru >>               ptent = pte_modify(ptent, vma->vm_page_prot); >>               writable = pte_write(ptent); >>               if (!writable && pte_write_upgrade && >> -                can_change_pte_writable(vma, addr, ptent)) >> +                can_change_ptes_writable(vma, addr, ptent, NULL, 1)) >>                   writable = true; >>           } >> @@ -5808,7 +5808,7 @@ static vm_fault_t do_numa_page(struct vm_fault >> *vmf) >>        */ >>       writable = pte_write(pte); >>       if (!writable && pte_write_upgrade && >> -        can_change_pte_writable(vma, vmf->address, pte)) >> +        can_change_ptes_writable(vma, vmf->address, pte, NULL, 1)) >>           writable = true; >>       folio = vm_normal_folio(vma, vmf->address, pte); >> diff --git a/mm/mprotect.c b/mm/mprotect.c >> index 33eabc995584..362fd7e5457d 100644 >> --- a/mm/mprotect.c >> +++ b/mm/mprotect.c >> @@ -40,8 +40,8 @@ >>   #include "internal.h" >> -bool can_change_pte_writable(struct vm_area_struct *vma, unsigned >> long addr, >> -                 pte_t pte) >> +bool can_change_ptes_writable(struct vm_area_struct *vma, unsigned >> long addr, >> +                  pte_t pte, struct folio *folio, unsigned int nr) >>   { >>       struct page *page; >> @@ -67,6 +67,9 @@ bool can_change_pte_writable(struct vm_area_struct >> *vma, unsigned long addr, >>            * write-fault handler similarly would map them writable >> without >>            * any additional checks while holding the PT lock. >>            */ >> +        if (unlikely(nr != 1)) >> +            return !folio_maybe_mapped_shared(folio); >> + >>           page = vm_normal_page(vma, addr, pte); >>           return page && PageAnon(page) && PageAnonExclusive(page); >>       } > > IIUC, As mentioned in the comment above, we should do the same anonymous > check > to large folios. And folio_maybe_mapped_shared() already handles both > order-0 > and large folios nicely, so we could simplify the logic as follows: Forget to add: Note that the exclusive flag is set only for non-large folios or the head page of large folios during mapping, so PageAnonExclusive() will always return false for tail pages of large folios, IIUC. Thanks, Lance > > diff --git a/mm/mprotect.c b/mm/mprotect.c > index 1605e89349d2..df56a30bb241 100644 > --- a/mm/mprotect.c > +++ b/mm/mprotect.c > @@ -43,8 +43,6 @@ >  bool can_change_ptes_writable(struct vm_area_struct *vma, unsigned > long addr, >                               pte_t pte, struct folio *folio, unsigned > int nr) >  { > -       struct page *page; > - >         if (WARN_ON_ONCE(!(vma->vm_flags & VM_WRITE))) >                 return false; > > @@ -67,11 +65,7 @@ bool can_change_ptes_writable(struct vm_area_struct > *vma, unsigned long addr, >                  * write-fault handler similarly would map them > writable without >                  * any additional checks while holding the PT lock. >                  */ > -               if (unlikely(nr != 1)) > -                       return !folio_maybe_mapped_shared(folio); > - > -               page = vm_normal_page(vma, addr, pte); > -               return page && PageAnon(page) && PageAnonExclusive(page); > +               return folio_test_anon(folio) && ! > folio_maybe_mapped_shared(folio); >         } > >         VM_WARN_ON_ONCE(is_zero_pfn(pte_pfn(pte)) && pte_dirty(pte)); > -- > > Thanks, > Lance > >> @@ -222,7 +225,7 @@ static long change_pte_range(struct mmu_gather *tlb, >>                */ >>               if ((cp_flags & MM_CP_TRY_CHANGE_WRITABLE) && >>                   !pte_write(ptent) && >> -                can_change_pte_writable(vma, addr, ptent)) >> +                can_change_ptes_writable(vma, addr, ptent, folio, 1)) >>                   ptent = pte_mkwrite(ptent, vma); >>               ptep_modify_prot_commit(vma, addr, pte, oldpte, ptent); >