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 X-Spam-Level: X-Spam-Status: No, score=-15.3 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,NICE_REPLY_A,SPF_HELO_NONE,SPF_PASS,USER_AGENT_SANE_1 autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E548FC433B4 for ; Fri, 30 Apr 2021 01:57:24 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 7B86D613CA for ; Fri, 30 Apr 2021 01:57:24 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 7B86D613CA Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=huawei.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id ED2FB6B0073; Thu, 29 Apr 2021 21:57:23 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id E83556B0074; Thu, 29 Apr 2021 21:57:23 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id D236D6B0075; Thu, 29 Apr 2021 21:57:23 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0226.hostedemail.com [216.40.44.226]) by kanga.kvack.org (Postfix) with ESMTP id B9FAC6B0073 for ; Thu, 29 Apr 2021 21:57:23 -0400 (EDT) Received: from smtpin31.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay05.hostedemail.com (Postfix) with ESMTP id 68C84181AEF3E for ; Fri, 30 Apr 2021 01:57:23 +0000 (UTC) X-FDA: 78087371166.31.E095FC0 Received: from szxga06-in.huawei.com (szxga06-in.huawei.com [45.249.212.32]) by imf21.hostedemail.com (Postfix) with ESMTP id 920F0E000119 for ; Fri, 30 Apr 2021 01:57:18 +0000 (UTC) Received: from DGGEMS402-HUB.china.huawei.com (unknown [172.30.72.58]) by szxga06-in.huawei.com (SkyGuard) with ESMTP id 4FWb6W10hBzlbls; Fri, 30 Apr 2021 09:55:14 +0800 (CST) Received: from [10.174.176.174] (10.174.176.174) by DGGEMS402-HUB.china.huawei.com (10.3.19.202) with Microsoft SMTP Server id 14.3.498.0; Fri, 30 Apr 2021 09:57:16 +0800 Subject: Re: [PATCH v2 3/5] mm/huge_memory.c: add missing read-only THP checking in transparent_hugepage_enabled() To: David Hildenbrand , CC: , , , , , , , , , , , , References: <20210429132648.305447-1-linmiaohe@huawei.com> <20210429132648.305447-4-linmiaohe@huawei.com> <68c8c4a8-c4f8-83db-7326-dabeea74239c@redhat.com> From: Miaohe Lin Message-ID: <9b511ad9-0ba1-e896-4eb5-0e91ca4b97ab@huawei.com> Date: Fri, 30 Apr 2021 09:57:16 +0800 User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101 Thunderbird/78.6.0 MIME-Version: 1.0 In-Reply-To: <68c8c4a8-c4f8-83db-7326-dabeea74239c@redhat.com> Content-Type: text/plain; charset="utf-8" Content-Language: en-US X-Originating-IP: [10.174.176.174] X-CFilter-Loop: Reflected X-Stat-Signature: b6pbyoyr38z9f5tsfhxch6m8bkytby8s X-Rspamd-Server: rspam02 X-Rspamd-Queue-Id: 920F0E000119 Authentication-Results: imf21.hostedemail.com; dkim=none; spf=pass (imf21.hostedemail.com: domain of linmiaohe@huawei.com designates 45.249.212.32 as permitted sender) smtp.mailfrom=linmiaohe@huawei.com; dmarc=pass (policy=none) header.from=huawei.com Received-SPF: none (huawei.com>: No applicable sender policy available) receiver=imf21; identity=mailfrom; envelope-from=""; helo=szxga06-in.huawei.com; client-ip=45.249.212.32 X-HE-DKIM-Result: none/none X-HE-Tag: 1619747838-763300 Content-Transfer-Encoding: quoted-printable X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: On 2021/4/29 22:57, David Hildenbrand wrote: > On 29.04.21 15:26, Miaohe Lin wrote: >> Since commit 99cb0dbd47a1 ("mm,thp: add read-only THP support for >> (non-shmem) FS"), read-only THP file mapping is supported. But it >> forgot to add checking for it in transparent_hugepage_enabled(). >> To fix it, we add checking for read-only THP file mapping and also >> introduce helper transhuge_vma_enabled() to check whether thp is >> enabled for specified vma to reduce duplicated code. >> >> Fixes: 99cb0dbd47a1 ("mm,thp: add read-only THP support for (non-shmem= ) FS") >> Signed-off-by: Miaohe Lin >> --- >> =C2=A0 include/linux/huge_mm.h | 21 +++++++++++++++++---- >> =C2=A0 mm/huge_memory.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2= =A0 6 ++++++ >> =C2=A0 mm/khugepaged.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= |=C2=A0 4 +--- >> =C2=A0 mm/shmem.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 3 +-- >> =C2=A0 4 files changed, 25 insertions(+), 9 deletions(-) >> >> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h >> index 0a526f211fec..f460b74619fc 100644 >> --- a/include/linux/huge_mm.h >> +++ b/include/linux/huge_mm.h >> @@ -115,6 +115,16 @@ extern struct kobj_attribute shmem_enabled_attr; >> =C2=A0 =C2=A0 extern unsigned long transparent_hugepage_flags; >> =C2=A0 +static inline bool transhuge_vma_enabled(struct vm_area_struct= *vma, >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 unsigned long v= m_flags) >=20 > You're passing the vma already, why do you pass vma->vm_flags separatel= y? It's sufficient to pass in the vma only. >=20 Many thanks for comment! IMO, vm_flags may not always equal to vma->vm_fl= ags. When hugepage_vma_check() is called from collapse_pte_mapped_thp, vma_flags =3D vma->vm_flags | VM_= HUGEPAGE. So I think we should pass vm_flags here. >> +{ >> +=C2=A0=C2=A0=C2=A0 /* Explicitly disabled through madvise. */ >> +=C2=A0=C2=A0=C2=A0 if ((vm_flags & VM_NOHUGEPAGE) || >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 test_bit(MMF_DISABLE_THP, = &vma->vm_mm->flags)) >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> +=C2=A0=C2=A0=C2=A0 return true; >> +} >> + >> =C2=A0 /* >> =C2=A0=C2=A0 * to be used on vmas which are known to support THP. >> =C2=A0=C2=A0 * Use transparent_hugepage_enabled otherwise >> @@ -128,15 +138,12 @@ static inline bool __transparent_hugepage_enable= d(struct vm_area_struct *vma) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (transparent_hugepage_flags & (1 << = TRANSPARENT_HUGEPAGE_NEVER_DAX)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0 -=C2=A0=C2=A0=C2=A0 if (vma->vm_flags & VM_NOHUGEPAGE) >> +=C2=A0=C2=A0=C2=A0 if (!transhuge_vma_enabled(vma, vma->vm_flags)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (vma_is_temporary_stack(vma)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0 -=C2=A0=C2=A0=C2=A0 if (test_bit(MMF_DISABLE_THP, &vma->vm_mm->= flags)) >> -=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> - >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (transparent_hugepage_flags & (1 << = TRANSPARENT_HUGEPAGE_FLAG)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return true; >> =C2=A0 @@ -362,6 +369,12 @@ static inline bool transhuge_vma_suitable(= struct vm_area_struct *vma, >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0 } >> =C2=A0 +static inline bool transhuge_vma_enabled(struct vm_area_struct= *vma, >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 unsigned long v= m_flags) >> +{ >> +=C2=A0=C2=A0=C2=A0 return false; >> +} >> + >> =C2=A0 static inline void prep_transhuge_page(struct page *page) {} >> =C2=A0 =C2=A0 static inline bool is_transparent_hugepage(struct page *= page) >> diff --git a/mm/huge_memory.c b/mm/huge_memory.c >> index 76ca1eb2a223..e24a96de2e37 100644 >> --- a/mm/huge_memory.c >> +++ b/mm/huge_memory.c >> @@ -68,12 +68,18 @@ bool transparent_hugepage_enabled(struct vm_area_s= truct *vma) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 /* The addr is used to check if the vma= size fits */ >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 unsigned long addr =3D (vma->vm_end & H= PAGE_PMD_MASK) - HPAGE_PMD_SIZE; >> =C2=A0 +=C2=A0=C2=A0=C2=A0 if (!transhuge_vma_enabled(vma, vma->vm_fla= gs)) >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (!transhuge_vma_suitable(vma, addr)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (vma_is_anonymous(vma)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return __transp= arent_hugepage_enabled(vma); >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (vma_is_shmem(vma)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return shmem_hu= ge_enabled(vma); >> +=C2=A0=C2=A0=C2=A0 if (IS_ENABLED(CONFIG_READ_ONLY_THP_FOR_FS) && vma= ->vm_file && >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 !inode_is_open_for_write(v= ma->vm_file->f_inode) && >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 (vma->vm_flags & VM_EXEC)) >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return true; >=20 > Nit: I'm really wondering why we have 3 different functions that sound = like they are doing the same thing >=20 > transparent_hugepage_enabled(vma) > transhuge_vma_enabled() > transhuge_vma_suitable() >=20 > Which check belongs where? Does it really have to be that complicated? >=20 IMO, transhuge_vma_suitable() checks whether pgoff , vm_start and vm_end = is possible for thp. transhuge_vma_enabled() checks whether thp is explicitly disabled through= madvise. And transparent_hugepage_enabled() use these helpers to get the conclusio= n whether thp is enabled for specified vma. Any suggestions? Thanks again! >> =C2=A0 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0 } >> diff --git a/mm/khugepaged.c b/mm/khugepaged.c >> index 6c0185fdd815..d97b20fad6e8 100644 >> --- a/mm/khugepaged.c >> +++ b/mm/khugepaged.c >> @@ -442,9 +442,7 @@ static inline int khugepaged_test_exit(struct mm_s= truct *mm) >> =C2=A0 static bool hugepage_vma_check(struct vm_area_struct *vma, >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 unsigned long vm_flags) >> =C2=A0 { >> -=C2=A0=C2=A0=C2=A0 /* Explicitly disabled through madvise. */ >> -=C2=A0=C2=A0=C2=A0 if ((vm_flags & VM_NOHUGEPAGE) || >> -=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 test_bit(MMF_DISABLE_THP, = &vma->vm_mm->flags)) >> +=C2=A0=C2=A0=C2=A0 if (!transhuge_vma_enabled(vma, vm_flags)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 /* Enabled via shmem mount optio= ns or sysfs settings. */ >> diff --git a/mm/shmem.c b/mm/shmem.c >> index a08cedefbfaa..1dcbec313c70 100644 >> --- a/mm/shmem.c >> +++ b/mm/shmem.c >> @@ -4032,8 +4032,7 @@ bool shmem_huge_enabled(struct vm_area_struct *v= ma) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 loff_t i_size; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pgoff_t off; >> =C2=A0 -=C2=A0=C2=A0=C2=A0 if ((vma->vm_flags & VM_NOHUGEPAGE) || >> -=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 test_bit(MMF_DISABLE_THP, = &vma->vm_mm->flags)) >> +=C2=A0=C2=A0=C2=A0 if (!transhuge_vma_enabled(vma, vma->vm_flags)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return false; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (shmem_huge =3D=3D SHMEM_HUGE_FORCE) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 return true; >> >=20 >=20