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=-13.6 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,NICE_REPLY_A, 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 E5771C433E6 for ; Tue, 1 Sep 2020 09:34:05 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 87688205CB for ; Tue, 1 Sep 2020 09:34:05 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 87688205CB 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 345446B005A; Tue, 1 Sep 2020 05:34:02 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 31B348E0001; Tue, 1 Sep 2020 05:34:02 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 2327B6B0074; Tue, 1 Sep 2020 05:34:02 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0090.hostedemail.com [216.40.44.90]) by kanga.kvack.org (Postfix) with ESMTP id 0C6776B005A for ; Tue, 1 Sep 2020 05:34:02 -0400 (EDT) Received: from smtpin20.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay02.hostedemail.com (Postfix) with ESMTP id A8548988F for ; Tue, 1 Sep 2020 09:34:01 +0000 (UTC) X-FDA: 77213981082.20.spade26_100234a27097 Received: from filter.hostedemail.com (10.5.16.251.rfc1918.com [10.5.16.251]) by smtpin20.hostedemail.com (Postfix) with ESMTP id 69124180BFBAF for ; Tue, 1 Sep 2020 09:34:01 +0000 (UTC) X-HE-Tag: spade26_100234a27097 X-Filterd-Recvd-Size: 9378 Received: from foss.arm.com (foss.arm.com [217.140.110.172]) by imf41.hostedemail.com (Postfix) with ESMTP for ; Tue, 1 Sep 2020 09:34:00 +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 041D430E; Tue, 1 Sep 2020 02:34:00 -0700 (PDT) Received: from [10.163.69.134] (unknown [10.163.69.134]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id 961733F71F; Tue, 1 Sep 2020 02:33:55 -0700 (PDT) Subject: Re: [PATCH v3 09/13] mm/debug_vm_pgtable/locks: Move non page table modifying test together To: "Aneesh Kumar K.V" , linux-mm@kvack.org, akpm@linux-foundation.org Cc: mpe@ellerman.id.au, linuxppc-dev@lists.ozlabs.org, linux-arm-kernel@lists.infradead.org, linux-s390@vger.kernel.org, linux-snps-arc@lists.infradead.org, x86@kernel.org, linux-arch@vger.kernel.org, Gerald Schaefer , Christophe Leroy , Vineet Gupta , Mike Rapoport , Qian Cai References: <20200827080438.315345-1-aneesh.kumar@linux.ibm.com> <20200827080438.315345-10-aneesh.kumar@linux.ibm.com> From: Anshuman Khandual Message-ID: <9b75b175-f319-d40e-a95e-b323b3db654a@arm.com> Date: Tue, 1 Sep 2020 15:03:23 +0530 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Language: en-US X-Rspamd-Queue-Id: 69124180BFBAF X-Spamd-Result: default: False [0.00 / 100.00] X-Rspamd-Server: rspam01 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 09/01/2020 12:08 PM, Aneesh Kumar K.V wrote: > On 9/1/20 9:11 AM, Anshuman Khandual wrote: >> >> >> On 08/27/2020 01:34 PM, Aneesh Kumar K.V wrote: >>> This will help in adding proper locks in a later patch >> >> It really makes sense to classify these tests here as static and dynam= ic. >> Static are the ones that test via page table entry values modification >> without changing anything on the actual page table itself. Where as th= e >> dynamic tests do change the page table. Static tests would probably ne= ver >> require any lock synchronization but the dynamic ones will do. >> >>> >>> Signed-off-by: Aneesh Kumar K.V >>> --- >>> =C2=A0 mm/debug_vm_pgtable.c | 52 ++++++++++++++++++++++++-----------= -------- >>> =C2=A0 1 file changed, 29 insertions(+), 23 deletions(-) >>> >>> diff --git a/mm/debug_vm_pgtable.c b/mm/debug_vm_pgtable.c >>> index 0ce5c6a24c5b..78c8af3445ac 100644 >>> --- a/mm/debug_vm_pgtable.c >>> +++ b/mm/debug_vm_pgtable.c >>> @@ -992,7 +992,7 @@ static int __init debug_vm_pgtable(void) >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 p4dp =3D p4d_alloc(mm, pgdp, vaddr); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pudp =3D pud_alloc(mm, p4dp, vaddr); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pmdp =3D pmd_alloc(mm, pudp, vaddr); >>> -=C2=A0=C2=A0=C2=A0 ptep =3D pte_alloc_map_lock(mm, pmdp, vaddr, &ptl= ); >>> +=C2=A0=C2=A0=C2=A0 ptep =3D pte_alloc_map(mm, pmdp, vaddr); >>> =C2=A0 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 /* >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 * Save all the page table page a= ddresses as the page table >>> @@ -1012,33 +1012,12 @@ static int __init debug_vm_pgtable(void) >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 p4d_basic_tests(p4d_aligned, prot); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pgd_basic_tests(pgd_aligned, prot); >>> =C2=A0 -=C2=A0=C2=A0=C2=A0 pte_clear_tests(mm, ptep, vaddr); >>> -=C2=A0=C2=A0=C2=A0 pmd_clear_tests(mm, pmdp); >>> -=C2=A0=C2=A0=C2=A0 pud_clear_tests(mm, pudp); >>> -=C2=A0=C2=A0=C2=A0 p4d_clear_tests(mm, p4dp); >>> -=C2=A0=C2=A0=C2=A0 pgd_clear_tests(mm, pgdp); >>> - >>> -=C2=A0=C2=A0=C2=A0 pte_advanced_tests(mm, vma, ptep, pte_aligned, va= ddr, prot); >>> -=C2=A0=C2=A0=C2=A0 pmd_advanced_tests(mm, vma, pmdp, pmd_aligned, va= ddr, prot, saved_ptep); >>> -=C2=A0=C2=A0=C2=A0 pud_advanced_tests(mm, vma, pudp, pud_aligned, va= ddr, prot); >>> -=C2=A0=C2=A0=C2=A0 hugetlb_advanced_tests(mm, vma, ptep, pte_aligned= , vaddr, prot); >>> - >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pmd_leaf_tests(pmd_aligned, prot); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pud_leaf_tests(pud_aligned, prot); >>> =C2=A0 -=C2=A0=C2=A0=C2=A0 pmd_huge_tests(pmdp, pmd_aligned, prot); >>> -=C2=A0=C2=A0=C2=A0 pud_huge_tests(pudp, pud_aligned, prot); >>> - >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pte_savedwrite_tests(pte_aligned, prot= none); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pmd_savedwrite_tests(pmd_aligned, prot= none); >>> =C2=A0 -=C2=A0=C2=A0=C2=A0 pte_unmap_unlock(ptep, ptl); >>> - >>> -=C2=A0=C2=A0=C2=A0 pmd_populate_tests(mm, pmdp, saved_ptep); >>> -=C2=A0=C2=A0=C2=A0 pud_populate_tests(mm, pudp, saved_pmdp); >>> -=C2=A0=C2=A0=C2=A0 p4d_populate_tests(mm, p4dp, saved_pudp); >>> -=C2=A0=C2=A0=C2=A0 pgd_populate_tests(mm, pgdp, saved_p4dp); >>> - >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pte_special_tests(pte_aligned, prot); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pte_protnone_tests(pte_aligned, protno= ne); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pmd_protnone_tests(pmd_aligned, protno= ne); >>> @@ -1056,11 +1035,38 @@ static int __init debug_vm_pgtable(void) >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pmd_swap_tests(pmd_aligned, prot); >>> =C2=A0 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 swap_migration_tests(); >>> -=C2=A0=C2=A0=C2=A0 hugetlb_basic_tests(pte_aligned, prot); >>> =C2=A0 =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pmd_thp_tests(pmd_aligned, prot= ); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pud_thp_tests(pud_aligned, prot); >>> =C2=A0 +=C2=A0=C2=A0=C2=A0 /* >>> +=C2=A0=C2=A0=C2=A0=C2=A0 * Page table modifying tests >>> +=C2=A0=C2=A0=C2=A0=C2=A0 */ >> >> In this comment, please do add some more details about how these tests >> need proper locking mechanism unlike the previous static ones. Also ad= d >> a similar comment section for the static tests that dont really change >> page table and need not require corresponding locking mechanism. Both >> comment sections will make this classification clear. >> >>> +=C2=A0=C2=A0=C2=A0 pte_clear_tests(mm, ptep, vaddr); >>> +=C2=A0=C2=A0=C2=A0 pmd_clear_tests(mm, pmdp); >>> +=C2=A0=C2=A0=C2=A0 pud_clear_tests(mm, pudp); >>> +=C2=A0=C2=A0=C2=A0 p4d_clear_tests(mm, p4dp); >>> +=C2=A0=C2=A0=C2=A0 pgd_clear_tests(mm, pgdp); >>> + >>> +=C2=A0=C2=A0=C2=A0 ptep =3D pte_alloc_map_lock(mm, pmdp, vaddr, &ptl= ); >>> +=C2=A0=C2=A0=C2=A0 pte_advanced_tests(mm, vma, ptep, pte_aligned, va= ddr, prot); >>> +=C2=A0=C2=A0=C2=A0 pmd_advanced_tests(mm, vma, pmdp, pmd_aligned, va= ddr, prot, saved_ptep); >>> +=C2=A0=C2=A0=C2=A0 pud_advanced_tests(mm, vma, pudp, pud_aligned, va= ddr, prot); >>> +=C2=A0=C2=A0=C2=A0 hugetlb_advanced_tests(mm, vma, ptep, pte_aligned= , vaddr, prot); >>> + >>> + >>> +=C2=A0=C2=A0=C2=A0 pmd_huge_tests(pmdp, pmd_aligned, prot); >>> +=C2=A0=C2=A0=C2=A0 pud_huge_tests(pudp, pud_aligned, prot); >>> + >>> +=C2=A0=C2=A0=C2=A0 pte_unmap_unlock(ptep, ptl); >>> + >>> +=C2=A0=C2=A0=C2=A0 pmd_populate_tests(mm, pmdp, saved_ptep); >>> +=C2=A0=C2=A0=C2=A0 pud_populate_tests(mm, pudp, saved_pmdp); >>> +=C2=A0=C2=A0=C2=A0 p4d_populate_tests(mm, p4dp, saved_pudp); >>> +=C2=A0=C2=A0=C2=A0 pgd_populate_tests(mm, pgdp, saved_p4dp); >>> + >>> +=C2=A0=C2=A0=C2=A0 hugetlb_basic_tests(pte_aligned, prot); >> >> hugetlb_basic_tests() is a non page table modifying static test and >> should be classified accordingly. >> >>> + >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 p4d_free(mm, saved_p4dp); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pud_free(mm, saved_pudp); >>> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 pmd_free(mm, saved_pmdp); >>> >> >> Changes till this patch fails to boot on arm64 platform. This should b= e >> folded with the next patch. >> >> [=C2=A0=C2=A0 17.080644] ------------[ cut here ]------------ >> [=C2=A0=C2=A0 17.081342] kernel BUG at mm/pgtable-generic.c:164! >> [=C2=A0=C2=A0 17.082091] Internal error: Oops - BUG: 0 [#1] PREEMPT SM= P >> [=C2=A0=C2=A0 17.082977] Modules linked in: >> [=C2=A0=C2=A0 17.083481] CPU: 79 PID: 1 Comm: swapper/0 Tainted: G=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 W=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0 5.9.0-rc2-00105-g740e72cd6717 #14 >> [=C2=A0=C2=A0 17.084998] Hardware name: linux,dummy-virt (DT) >> [=C2=A0=C2=A0 17.085745] pstate: 60400005 (nZCv daif +PAN -UAO BTYPE=3D= --) >> [=C2=A0=C2=A0 17.086645] pc : pgtable_trans_huge_deposit+0x58/0x60 >> [=C2=A0=C2=A0 17.087462] lr : debug_vm_pgtable+0x4dc/0x8f0 >> [=C2=A0=C2=A0 17.088168] sp : ffff80001219bcf0 >> [=C2=A0=C2=A0 17.088710] x29: ffff80001219bcf0 x28: ffff8000114ac000 >> [=C2=A0=C2=A0 17.089574] x27: ffff8000114ac000 x26: 0020000000000fd3 >> [=C2=A0=C2=A0 17.090431] x25: 0020000081400fd3 x24: fffffe00175c12c0 >> [=C2=A0=C2=A0 17.091286] x23: ffff0005df04d1a8 x22: 0000f18d6e035000 >> [=C2=A0=C2=A0 17.092143] x21: ffff0005dd790000 x20: ffff0005df18e1a8 >> [=C2=A0=C2=A0 17.093003] x19: ffff0005df04cb80 x18: 0000000000000014 >> [=C2=A0=C2=A0 17.093859] x17: 00000000b76667d0 x16: 00000000fd4e6611 >> [=C2=A0=C2=A0 17.094716] x15: 0000000000000001 x14: 0000000000000002 >> [=C2=A0=C2=A0 17.095572] x13: 000000000055d966 x12: fffffe001755e400 >> [=C2=A0=C2=A0 17.096429] x11: 0000000000000008 x10: ffff0005fcb6e210 >> [=C2=A0=C2=A0 17.097292] x9 : ffff0005fcb6e210 x8 : 0020000081590fd3 >> [=C2=A0=C2=A0 17.098149] x7 : ffff0005dd71e0c0 x6 : ffff0005df18e1a8 >> [=C2=A0=C2=A0 17.099006] x5 : 0020000081590fd3 x4 : 0020000081590fd3 >> [=C2=A0=C2=A0 17.099862] x3 : ffff0005df18e1a8 x2 : fffffe00175c12c0 >> [=C2=A0=C2=A0 17.100718] x1 : fffffe00175c1300 x0 : 0000000000000000 >> [=C2=A0=C2=A0 17.101583] Call trace: >> [=C2=A0=C2=A0 17.101993]=C2=A0 pgtable_trans_huge_deposit+0x58/0x60 >> [=C2=A0=C2=A0 17.102754]=C2=A0 do_one_initcall+0x74/0x1cc >> [=C2=A0=C2=A0 17.103381]=C2=A0 kernel_init_freeable+0x1d0/0x238 >> [=C2=A0=C2=A0 17.104089]=C2=A0 kernel_init+0x14/0x118 >> [=C2=A0=C2=A0 17.104658]=C2=A0 ret_from_fork+0x10/0x34 >> [=C2=A0=C2=A0 17.105251] Code: f9000443 f9000843 f9000822 d65f03c0 (d4= 210000) >> [=C2=A0=C2=A0 17.106303] ---[ end trace e63471e00f8c147f ]--- >> >=20 > IIUC, this should happen even without the series right? This is >=20 > =C2=A0=C2=A0=C2=A0=C2=A0assert_spin_locked(pmd_lockptr(mm, pmdp)); Crash does not happen without this series. A previous patch [PATCH v3 08/= 13] added pgtable_trans_huge_deposit/withdraw() in pmd_advanced_tests(). arm6= 4 does not define __HAVE_ARCH_PGTABLE_DEPOSIT and hence falls back on gener= ic pgtable_trans_huge_deposit() which the asserts the lock.