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 D5FA0C2D0C9 for ; Thu, 12 Dec 2019 15:18:43 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 932482467A for ; Thu, 12 Dec 2019 15:18:43 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 932482467A Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=arm.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 321FA8E0005; Thu, 12 Dec 2019 10:18:43 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 2D1818E0001; Thu, 12 Dec 2019 10:18:43 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 1E7E78E0005; Thu, 12 Dec 2019 10:18:43 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0049.hostedemail.com [216.40.44.49]) by kanga.kvack.org (Postfix) with ESMTP id 095D78E0001 for ; Thu, 12 Dec 2019 10:18:43 -0500 (EST) Received: from smtpin14.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay02.hostedemail.com (Postfix) with SMTP id ABB4B4427 for ; Thu, 12 Dec 2019 15:18:42 +0000 (UTC) X-FDA: 76256846484.14.lock08_86222381a4018 X-HE-Tag: lock08_86222381a4018 X-Filterd-Recvd-Size: 10439 Received: from foss.arm.com (foss.arm.com [217.140.110.172]) by imf36.hostedemail.com (Postfix) with ESMTP for ; Thu, 12 Dec 2019 15:18:41 +0000 (UTC) Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 68A6530E; Thu, 12 Dec 2019 07:18:40 -0800 (PST) Received: from [10.37.9.115] (unknown [10.37.9.115]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id BB9B63F6CF; Thu, 12 Dec 2019 07:18:35 -0800 (PST) Subject: Re: [PATCH v16 11/25] mm: pagewalk: Add p4d_entry() and pgd_entry() To: =?UTF-8?Q?Thomas_Hellstr=c3=b6m_=28VMware=29?= Cc: Mark Rutland , Dave Hansen , Andy Lutomirski , Arnd Bergmann , Ard Biesheuvel , Peter Zijlstra , Catalin Marinas , x86@kernel.org, linux-kernel@vger.kernel.org, "linux-mm@kvack.org" , =?UTF-8?B?SsOpcsO0bWUgR2xpc3Nl?= , Ingo Molnar , Borislav Petkov , Zong Li , "H. Peter Anvin" , James Morse , Thomas Gleixner , Will Deacon , Andrew Morton , linux-arm-kernel@lists.infradead.org, "Liang, Kan" References: <20191206135316.47703-1-steven.price@arm.com> <20191206135316.47703-12-steven.price@arm.com> <13280f9e-6f03-e1fd-659a-31462ba185b0@shipmail.org> <7fd20e9f-822a-897d-218e-bddf135fd33d@shipmail.org> <16b2ecbc-316a-33f8-ace2-e54cd8001b24@shipmail.org> From: Steven Price Message-ID: <2cef327b-a16b-f76d-4c3f-bd894332c736@arm.com> Date: Thu, 12 Dec 2019 15:18:33 +0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.2.1 MIME-Version: 1.0 In-Reply-To: <16b2ecbc-316a-33f8-ace2-e54cd8001b24@shipmail.org> Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-GB 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 12/12/2019 14:04, Thomas Hellstr=C3=B6m (VMware) wrote: > On 12/12/19 2:15 PM, Steven Price wrote: >> On 12/12/2019 11:33, Thomas Hellstr=C3=B6m (VMware) wrote: >>> On 12/12/19 12:23 PM, Thomas Hellstr=C3=B6m (VMware) wrote: >>>> On 12/6/19 2:53 PM, Steven Price wrote: >>>>> pgd_entry() and pud_entry() were removed by commit 0b1fbfe50006c410 >>>>> ("mm/pagewalk: remove pgd_entry() and pud_entry()") because there w= ere >>>>> no users. We're about to add users so reintroduce them, along with >>>>> p4d_entry() as we now have 5 levels of tables. >>>>> >>>>> Note that commit a00cc7d9dd93d66a ("mm, x86: add support for >>>>> PUD-sized transparent hugepages") already re-added pud_entry() but = with >>>>> different semantics to the other callbacks. Since there have never >>>>> been upstream users of this, revert the semantics back to match the >>>>> other callbacks. This means pud_entry() is called for all entries, = not >>>>> just transparent huge pages. >> >> When I wrote that there were no upstream users, which sadly shows how >> long ago that was :( >> >>>> Actually, there are two users of pud_entry(), in hmm.c and since 5.5= rc1 also mapping_dirty_helpers.c. The latter one is unproblematic and req= uires no attention but the one in hmm.c is probably largely untested, and= seems to assume it was called outside of the spinlock. >>>> >>>> The problem with the current patch is that the hmm pud_entry will tr= averse also pmds, so that will be done twice now. >>>> >>>> In another thread we were discussing a means of rerunning the level = (in case of a race), or continuing after a level, based on the return val= ue after the callback. The change was fairly invasive, >>>> >>> Hmm. Forgot to remove the above text that appears twice. :(. The corr= ect one is inline below. >>> >>>> >>>>> Tested-by: Zong Li >>>>> Signed-off-by: Steven Price >>>>> --- >>>>> =C2=A0 include/linux/pagewalk.h | 19 +++++++++++++------ >>>>> =C2=A0 mm/pagewalk.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 | 27 ++++++++++++++++----------- >>>>> =C2=A0 2 files changed, 29 insertions(+), 17 deletions(-) >>>>> >>>>> diff --git a/include/linux/pagewalk.h b/include/linux/pagewalk.h >>>>> index 6ec82e92c87f..06790f23957f 100644 >>>>> --- a/include/linux/pagewalk.h >>>>> +++ b/include/linux/pagewalk.h >>>>> @@ -8,15 +8,15 @@ struct mm_walk; >>>>> =C2=A0 =C2=A0 /** >>>>> =C2=A0=C2=A0 * mm_walk_ops - callbacks for walk_page_range >>>>> - * @pud_entry:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if set, c= alled for each non-empty PUD (2nd-level) entry >>>>> - *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= this handler should only handle pud_trans_huge() puds. >>>>> - *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= the pmd_entry or pte_entry callbacks will be used for >>>>> - *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= regular PUDs. >>>>> - * @pmd_entry:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if set, c= alled for each non-empty PMD (3rd-level) entry >>>>> + * @pgd_entry:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if set, c= alled for each non-empty PGD (top-level) entry >>>>> + * @p4d_entry:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if set, c= alled for each non-empty P4D entry >>>>> + * @pud_entry:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if set, c= alled for each non-empty PUD entry >>>>> + * @pmd_entry:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if set, c= alled for each non-empty PMD entry >>>>> =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 this handler is required to be able to handle >>>>> =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 pmd_trans_huge() pmds.=C2=A0 They may simply choose to >>>>> =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 split_huge_page() instead of handling it explicitly. >>>>> - * @pte_entry:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if set, c= alled for each non-empty PTE (4th-level) entry >>>>> + * @pte_entry:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if set, c= alled for each non-empty PTE (lowest-level) >>>>> + *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= entry >>>>> =C2=A0=C2=A0 * @pte_hole:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= if set, called for each hole at all levels >>>>> =C2=A0=C2=A0 * @hugetlb_entry:=C2=A0=C2=A0=C2=A0 if set, called for= each hugetlb entry >>>>> =C2=A0=C2=A0 * @test_walk:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= caller specific callback function to determine whether >>>>> @@ -27,8 +27,15 @@ struct mm_walk; >>>>> =C2=A0=C2=A0 * @pre_vma:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0 if set, called before starting walk on a non-null v= ma. >>>>> =C2=A0=C2=A0 * @post_vma:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 if set, called after a walk on a non-null vma, provide= d >>>>> =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=C2=A0=C2=A0 = that @pre_vma and the vma walk succeeded. >>>>> + * >>>>> + * p?d_entry callbacks are called even if those levels are folded = on a >>>>> + * particular architecture/configuration. >>>>> =C2=A0=C2=A0 */ >>>>> =C2=A0 struct mm_walk_ops { >>>>> +=C2=A0=C2=A0=C2=A0 int (*pgd_entry)(pgd_t *pgd, unsigned long addr= , >>>>> +=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 next, struct mm_walk *walk); >>>>> +=C2=A0=C2=A0=C2=A0 int (*p4d_entry)(p4d_t *p4d, unsigned long addr= , >>>>> +=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 next, struct mm_walk *walk); >>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 int (*pud_entry)(pud_t *pud, unsigne= d long addr, >>>>> =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 next, struct mm_walk *walk); >>>>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 int (*pmd_entry)(pmd_t *pmd, unsigne= d long addr, >>>>> diff --git a/mm/pagewalk.c b/mm/pagewalk.c >>>>> index ea0b9e606ad1..c089786e7a7f 100644 >>>>> --- a/mm/pagewalk.c >>>>> +++ b/mm/pagewalk.c >>>>> @@ -94,15 +94,9 @@ static int walk_pud_range(p4d_t *p4d, unsigned l= ong addr, unsigned long end, >>>>> =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 if (o= ps->pud_entry) { >>>>> -=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= spinlock_t *ptl =3D pud_trans_huge_lock(pud, walk->vma); >>>>> - >>>>> -=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= if (ptl) { >>>>> -=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 err =3D ops->pud_entry(pud, addr, next, walk); >>>>> -=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 spin_unlock(ptl); >>>>> -=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 if (err) >>>>> -=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 break; >>>>> -=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 continue; >>>>> -=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=C2=A0= err =3D ops->pud_entry(pud, addr, next, walk); >>>>> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= if (err) >>>>> +=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 break; >>>> >>>> Actually, there are two current users of pud_entry(), in hmm.c and s= ince 5.5rc1 also mapping_dirty_helpers.c. The latter one is unproblematic= and requires no attention but the one in hmm.c is probably largely untes= ted, and seems to assume it was called outside of the spinlock. >> >> Thanks for pointing that out, I guess the simplest fix would be to >> squash in something like the below which should restore the old >> behaviour for hmm.c without affecting others. >> >> Steve=20 >=20 > I'm not fully sure that the old behaviour is the correct one, but defin= itely hmm's pud_entry needs some fixing. > I'm more concerned with the pagewalk code. With your patch it actually = splits all huge puds present in the page-table > on each page walk which is not what we want. Good catch - yes that's certainly not ideal. > One idea would be to add a new member to struct_mm_walk: >=20 > enum page_walk_ret_action { > =C2=A0=C2=A0=C2=A0=C2=A0ACTION_SUBTREE =3D 0, > =C2=A0=C2=A0=C2=A0=C2=A0ACTION_CONTINUE =3D 1, > =C2=A0=C2=A0=C2=A0=C2=A0ACTION_AGAIN =3D 2 /* Only for levels that tha= ve p?d_unstable */ > }; >=20 > struct mm_walk { > =C2=A0=C2=A0=C2=A0=C2=A0... > =C2=A0=C2=A0=C2=A0=C2=A0enum page_walk_ret_action action; /* or perhap= s as an enum */ > }; >=20 >=20 > if (ops->pud_entry) { > =C2=A0=C2=A0=C2=A0=C2=A0walk->action =3D ACTION_SUBTREE; > =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=A0if (walk->action =3D=3D ACTION_AGAIN)=C2=A0 /*= Callback tried to split huge entry, but failed */ > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 goto again; > =C2=A0=C2=A0=C2=A0=C2=A0else if (walk->action =3D=3D ACTION_CONTINUE) = /* Done with this subtree. Probably huge entry handled. */ > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 continue; > =C2=A0=C2=A0=C2=A0=C2=A0/* ACTION_SUBTREE falls through */ > } I'll have a go at implementing the above - this might also allow removing= the test_p?d() callbacks as they can simply return ACTION_CONTINUE. Steve > we discussed something similar before on linux-mm, but the idea then wa= s to redefine > the positive return value of the callback to the action, but that meant= changing those existing callbacks that relied on > a positive return value. The above would be helpful also for pmd_entry. >=20 > /Thomas >=20 >=20 >=20 >=20 >=20 > _______________________________________________ > linux-arm-kernel mailing list > linux-arm-kernel@lists.infradead.org > http://lists.infradead.org/mailman/listinfo/linux-arm-kernel