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=-8.2 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,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 597E1C33CA2 for ; Fri, 10 Jan 2020 05:34:50 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 0A99A2072E for ; Fri, 10 Jan 2020 05:34:49 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 0A99A2072E Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 813FE8E0005; Fri, 10 Jan 2020 00:34:49 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 79D7A8E0001; Fri, 10 Jan 2020 00:34:49 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 665158E0005; Fri, 10 Jan 2020 00:34:49 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0042.hostedemail.com [216.40.44.42]) by kanga.kvack.org (Postfix) with ESMTP id 457148E0001 for ; Fri, 10 Jan 2020 00:34:49 -0500 (EST) Received: from smtpin20.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay05.hostedemail.com (Postfix) with SMTP id C7A28181AC9C6 for ; Fri, 10 Jan 2020 05:34:48 +0000 (UTC) X-FDA: 76360610256.20.bells45_47015b5222d4f X-HE-Tag: bells45_47015b5222d4f X-Filterd-Recvd-Size: 14304 Received: from mga06.intel.com (mga06.intel.com [134.134.136.31]) by imf11.hostedemail.com (Postfix) with ESMTP for ; Fri, 10 Jan 2020 05:34:47 +0000 (UTC) X-Amp-Result: UNKNOWN X-Amp-Original-Verdict: FILE UNKNOWN X-Amp-File-Uploaded: False Received: from fmsmga004.fm.intel.com ([10.253.24.48]) by orsmga104.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 09 Jan 2020 21:34:46 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.69,415,1571727600"; d="scan'208";a="246892528" Received: from richard.sh.intel.com (HELO localhost) ([10.239.159.54]) by fmsmga004.fm.intel.com with ESMTP; 09 Jan 2020 21:34:44 -0800 Date: Fri, 10 Jan 2020 13:34:42 +0800 From: Wei Yang To: Li Xinhai Cc: "richardw.yang" , khlebnikov , "linux-mm@kvack.org" , akpm , "linux-kernel@vger.kernel.org" , Rik van Riel , "kirill.shutemov" Subject: Re: [PATCH v2 1/2] mm/rmap: fix and simplify reusing mergeable anon_vma as parent when fork Message-ID: <20200110053442.GA27846@richard> Reply-To: Wei Yang References: <157839239609.694.10268055713935919822.stgit@buzz> <20200108023211.GC13943@richard> <20200109025240.GA2000@richard> <20200110023029.GB16823@richard> <20200110112357351531132@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <20200110112357351531132@gmail.com> User-Agent: Mutt/1.9.4 (2018-02-28) 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 Fri, Jan 10, 2020 at 11:23:59AM +0800, Li Xinhai wrote: >On 2020-01-10=A0at 10:30=A0Wei Yang=A0wrote: >>On Thu, Jan 09, 2020 at 11:54:21AM +0300, Konstantin Khlebnikov wrote: >>> >>> >>>On 09/01/2020 05.52, Wei Yang wrote: >>>> On Wed, Jan 08, 2020 at 01:40:44PM +0300, Konstantin Khlebnikov wrot= e: >>>> > On 08/01/2020 05.32, Wei Yang wrote: >>>> > > On Tue, Jan 07, 2020 at 01:19:56PM +0300, Konstantin Khlebnikov = wrote: >>>> > > > This fixes some misconceptions in commit 4e4a9eb92133 ("mm/rma= p.c: reuse >>>> > > > mergeable anon_vma as parent when fork"). It merges anon-vma i= n unexpected >>>> > > > way but fortunately still produces valid anon-vma tree, so not= hing crashes. >>>> > > > >>>> > > > If in parent VMAs: SRC1 SRC2 .. SRCn share anon-vma ANON0, the= n after fork >>>> > > > before all patches in child process related VMAs: DST1 DST2 ..= DSTn will >>>> > > > fork indepndent anon-vmas: ANON1 ANON2 .. ANONn (each is child= of ANON0). >>>> > > > Before this patch only DST1 will fork new ANON1 and following = DST2 .. DSTn >>>> > > > will share parent's ANON0 (i.e. anon-vma tree is valid but isn= 't optimal). >>>> > > > With this patch DST1 will create new ANON1 and DST2 .. DSTn wi= ll share it. >>>> > > > >>>> > > > Root problem caused by initialization order in dup_mmap(): vma= ->vm_prev >>>> > > > is set after calling anon_vma_fork(). Thus in anon_vma_fork() = it points to >>>> > > > previous VMA in parent mm. >>>> > > > >>>> > > > Second problem is hidden behind first one: assumption "Parent = has vm_prev, >>>> > > > which implies we have vm_prev" is wrong if first VMA in parent= mm has set >>>> > > > flag VM_DONTCOPY. Luckily prev->anon_vma doesn't dereference N= ULL pointer >>>> > > > because in current code 'prev' actually is same as 'pprev'. >>>> > > > >>>> > > > Third hidden problem is linking between VMA and anon-vmas whos= e pages it >>>> > > > could contain. Loop in anon_vma_clone() attaches only parent's= anon-vmas, >>>> > > > shared anon-vma isn't attached. But every mapped page stays re= achable in >>>> > > > rmap because we erroneously share anon-vma from parent's previ= ous VMA. >>>> > > > >>>> > > > This patch moves sharing logic out of anon_vma_clone() into mo= re specific >>>> > > > anon_vma_fork() because this supposed to work only at fork() a= nd simply >>>> > > > reuses anon_vma from previous VMA if it is forked from the sam= e anon-vma. >>>> > > > >>>> > > > Signed-off-by: Konstantin Khlebnikov >>>> > > > Reported-by: Li Xinhai >>>> > > > Fixes: 4e4a9eb92133 ("mm/rmap.c: reuse mergeable anon_vma as p= arent when fork") >>>> > > > Link: https://lore.kernel.org/linux-mm/CALYGNiNzz+dxHX0g5-gNyp= UQc3B=3D8_Scp53-NTOh=3DzWsdUuHAw@mail.gmail.com/T/#t >>>> > > > --- >>>> > > > include/linux/rmap.h |=A0=A0=A0 3 ++- >>>> > > > kernel/fork.c=A0=A0=A0=A0=A0=A0=A0 |=A0=A0=A0 2 +- >>>> > > > mm/rmap.c=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |=A0=A0 23 ++++++++= +-------------- >>>> > > > 3 files changed, 12 insertions(+), 16 deletions(-) >>>> > > > >>>> > > > diff --git a/include/linux/rmap.h b/include/linux/rmap.h >>>> > > > index 988d176472df..560e4480dcd0 100644 >>>> > > > --- a/include/linux/rmap.h >>>> > > > +++ b/include/linux/rmap.h >>>> > > > @@ -143,7 +143,8 @@ void anon_vma_init(void); /* create anon_v= ma_cachep */ >>>> > > > int=A0 __anon_vma_prepare(struct vm_area_struct *); >>>> > > > void unlink_anon_vmas(struct vm_area_struct *); >>>> > > > int anon_vma_clone(struct vm_area_struct *, struct vm_area_str= uct *); >>>> > > > -int anon_vma_fork(struct vm_area_struct *, struct vm_area_str= uct *); >>>> > > > +int anon_vma_fork(struct vm_area_struct *vma, struct vm_area_= struct *pvma, >>>> > > > + =A0 struct vm_area_struct *prev); >>>> > > > >>>> > > > static inline int anon_vma_prepare(struct vm_area_struct *vma) >>>> > > > { >>>> > > > diff --git a/kernel/fork.c b/kernel/fork.c >>>> > > > index 2508a4f238a3..c33626993831 100644 >>>> > > > --- a/kernel/fork.c >>>> > > > +++ b/kernel/fork.c >>>> > > > @@ -556,7 +556,7 @@ static __latent_entropy int dup_mmap(struc= t mm_struct *mm, >>>> > > > tmp->anon_vma =3D NULL; >>>> > > > if (anon_vma_prepare(tmp)) >>>> > > > goto fail_nomem_anon_vma_fork; >>>> > > > - } else if (anon_vma_fork(tmp, mpnt)) >>>> > > > + } else if (anon_vma_fork(tmp, mpnt, prev)) >>>> > > > goto fail_nomem_anon_vma_fork; >>>> > > > tmp->vm_flags &=3D ~(VM_LOCKED | VM_LOCKONFAULT); >>>> > > > tmp->vm_next =3D tmp->vm_prev =3D NULL; >>>> > > > diff --git a/mm/rmap.c b/mm/rmap.c >>>> > > > index b3e381919835..3c1e04389291 100644 >>>> > > > --- a/mm/rmap.c >>>> > > > +++ b/mm/rmap.c >>>> > > > @@ -269,19 +269,6 @@ int anon_vma_clone(struct vm_area_struct = *dst, struct vm_area_struct *src) >>>> > > > { >>>> > > > struct anon_vma_chain *avc, *pavc; >>>> > > > struct anon_vma *root =3D NULL; >>>> > > > - struct vm_area_struct *prev =3D dst->vm_prev, *pprev =3D src= ->vm_prev; >>>> > > > - >>>> > > > - /* >>>> > > > - * If parent share anon_vma with its vm_prev, keep this shari= ng in in >>>> > > > - * child. >>>> > > > - * >>>> > > > - * 1. Parent has vm_prev, which implies we have vm_prev. >>>> > > > - * 2. Parent and its vm_prev have the same anon_vma. >>>> > > > - */ >>>> > > > - if (!dst->anon_vma && src->anon_vma && >>>> > > > - =A0=A0=A0 pprev && pprev->anon_vma =3D=3D src->anon_vma) >>>> > > > - dst->anon_vma =3D prev->anon_vma; >>>> > > > - >>>> > > > >>>> > > > list_for_each_entry_reverse(pavc, &src->anon_vma_chain, same_v= ma) { >>>> > > > struct anon_vma *anon_vma; >>>> > > > @@ -332,7 +319,8 @@ int anon_vma_clone(struct vm_area_struct *= dst, struct vm_area_struct *src) >>>> > > >=A0=A0=A0 * the corresponding VMA in the parent process is atta= ched to. >>>> > > >=A0=A0=A0 * Returns 0 on success, non-zero on failure. >>>> > > >=A0=A0=A0 */ >>>> > > > -int anon_vma_fork(struct vm_area_struct *vma, struct vm_area_= struct *pvma) >>>> > > > +int anon_vma_fork(struct vm_area_struct *vma, struct vm_area_= struct *pvma, >>>> > > > + =A0 struct vm_area_struct *prev) >>>> > > > { >>>> > > > struct anon_vma_chain *avc; >>>> > > > struct anon_vma *anon_vma; >>>> > > > @@ -342,6 +330,13 @@ int anon_vma_fork(struct vm_area_struct *= vma, struct vm_area_struct *pvma) >>>> > > > if (!pvma->anon_vma) >>>> > > > return 0; >>>> > > > >>>> > > > + /* Share anon_vma with previous VMA if it has the same paren= t. */ >>>> > > > + if (prev && prev->anon_vma && >>>> > > > + =A0=A0=A0 prev->anon_vma->parent =3D=3D pvma->anon_vma) { >>>> > > > + vma->anon_vma =3D prev->anon_vma; >>>> > > > + return anon_vma_clone(vma, prev); >>>> > > > + } >>>> > > > + >>>> > > >>>> > > I am afraid this one change the intended behavior. Let's put a c= hart to >>>> > > describe. >>>> > > >>>> > > Commit 4e4a9eb92133 ("mm/rmap.c: reusemergeable anon_vma as pare= nt when >>>> > > fork") tries to improve the following situation. >>>> > > >>>> > > Before the commit, the behavior is like this: >>>> > > >>>> > > Parent process: >>>> > > >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 +-----+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 | pav |<-----------------+--------------= --------+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 +-----+=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0 | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0 | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 += -----------+=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----------+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |= pprev=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0 |pvma=A0=A0=A0=A0=A0=A0= | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 += -----------+=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----------+ >>>> > > >>>> > > Child Process >>>> > > >>>> > > >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 +-----+=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 | av1 |<-----------------+=A0 | av2 |<--= ----------+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 +-----+=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0 |=A0 +-----+=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0 | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 += -----------+=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----------+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |= prev=A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0 |vma=A0=A0=A0=A0=A0=A0= =A0 | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 += -----------+=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----------+ >>>> > > >>>> > > >>>> > > Parent pprev and pvma share the same anon_vma due to >>>> > > find_mergeable_anon_vma(). While the anon_vma_clone() would pick= up different >>>> > > anon_vma for child process's vma. >>>> > > >>>> > > The purpose of my commit is to give child process the following = shape. >>>> > > >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 +-----+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 | av=A0 |<-----------------+------------= ----------+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0 +-----+=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0 | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0 | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 += -----------+=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----------+ >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |= prev=A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0 |vma=A0=A0=A0=A0=A0=A0= =A0 | >>>> > >=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 += -----------+=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----------+ >>>> > > >>>> > > After this, we reduce the extra "av2" for child process. But yes= , because of >>>> > > the two reasons you found, it didn't do the exact thing. >>>> > > >>>> > > While if my understanding is correct, the anon_vma_clone() would= pick up any >>>> > > anon_vma in its process tree, except parent's. If this fails to = get a reusable >>>> > > one, anon_vma_fork() would allocate one, whose parent is pvma->a= non_vma. >>>> > > >>>> > > Let me summarise original behavior: >>>> > > >>>> > >=A0=A0=A0=A0 * if anon_vma_clone succeed, it find one anon_vma in= the process tree, but >>>> > >=A0=A0=A0=A0=A0=A0 it could not be pvma->anon_vma >>>> > >=A0=A0=A0=A0 * if anon_vma_clone fail, it will allocate a new ano= n_vma and its parent is >>>> > >=A0=A0=A0=A0=A0=A0 pvma->anon_vam >>>> > > >>>> > > Then take a look into your code here. >>>> > > >>>> > > "prev->anon_vma->parent =3D=3D pvma->anon_vma" means prev->anon_= vma parent is >>>> > > pvma's anon_vma. If my understanding is correct, this just match= the second >>>> > > case. For "prev", we didn't find a reusable anon_vma and allocat= e a new one. >>>> > > >>>> > > But how about the first case? prev reuse an anon_vma in the proc= ess tree which >>>> > > is not parent's? >>>> > >>>> > If anon_vma_clone() pick old anon-vma for first vma in sharing cha= in (prev) >>>> > then second vma (vma) will fork new anon-vma (unless pick another = old anon-vma), >>>> > then third vma will share it. And so on. >>>> >>>> No, I am afraid you are not correct here. Or I don't understand your= sentence. >>>> >>>> This is my understanding about the behavior before my commit. Suppos= e av1 and >>>> av2 are both reused from old anon_vma. And if my understanding is co= rrect, >>>> they are different from pvma->anon_vma. Then how your code match thi= s >>>> situatioin? >>>> >>>>=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----+=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0 +-----+ >>>>=A0=A0=A0=A0=A0=A0=A0=A0=A0 | av1 |<-----------------+=A0 | av2 |<---= ---------+ >>>>=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----+=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0 |=A0 +-----+=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 | >>>>=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0= =A0=A0=A0=A0=A0=A0=A0=A0 | >>>>=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-= ----------+=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----------+ >>>>=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |p= rev=A0=A0=A0=A0=A0=A0 |=A0=A0=A0=A0=A0=A0=A0=A0=A0 |vma=A0=A0=A0=A0=A0=A0= =A0 | >>>>=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-= ----------+=A0=A0=A0=A0=A0=A0=A0=A0=A0 +-----------+ >>>> >>>> Would you explain your understanding the second and third vma in you= r >>>> sentence? Which case you are trying to illustrate? >>> >>>series of vma in parent with shared AV: >>> >>>SRC1 - AV0 >>>SRC2 - AV0 >>>SRC3 - AV0 >>>... >>>SRCn - AV0 >>> >>>in child after fork >>> >>>DST1 - AV_OLD_1 (some old vma, picked by anon_vma_clone) plus DST1 is = attached to same AVs as SRC1 >>>DST2 - AV_OLD_2 (other old vma) plus DST1 is attached to same AVs as S= RC2 >>>DST2 - AV1 prev AV parent does not match AV0, no old vma found for reu= sing -> allocate new one (child of AV0) >>>DST3 - AV1 - DST2->AV->parent =3D=3D SRC3->AV (AV0) -> share AV with p= rev >>>DST4 - AV1 - same thing >>>... >>>DSTn - AV1 >>> >> >>Yes, your code works for DST3..DSTn. They will pick up AV1 since >>(DST2->AV->parent =3D=3D SRC3->AV). >> >>My question is why DST1 and DST2 has different AV? The purpose of my pa= tch >>tries to make child has the same topology and parent. So the ideal look= of >>child is: >> >>DST1 - AV1 >>DST2 - AV1 >>DST2 - AV1 >>DST3 - AV1 >>DST4 - AV1 >> >>Would you mind putting more words on DST1 and DST2? I didn't fully unde= rstand >>the logic here. >> >>Thanks >>=20 > >I think that the first version is doing the work as you expected, but be= en >revised in second version, to limits the number of users of reused old >anon(which=A0is picked=A0in anon_vma_clone() and keep the tree structure= . > Any reason to reduce the reuse? Maybe I lost some point. >>-- >>Wei Yang >>Help you, Help me --=20 Wei Yang Help you, Help me