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 mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 482F7C433EF for ; Tue, 5 Oct 2021 01:34:22 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 2C07C61406 for ; Tue, 5 Oct 2021 01:34:22 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232937AbhJEBgK (ORCPT ); Mon, 4 Oct 2021 21:36:10 -0400 Received: from mx0a-00069f02.pphosted.com ([205.220.165.32]:16860 "EHLO mx0a-00069f02.pphosted.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232447AbhJEBeD (ORCPT ); Mon, 4 Oct 2021 21:34:03 -0400 Received: from pps.filterd (m0246629.ppops.net [127.0.0.1]) by mx0b-00069f02.pphosted.com (8.16.1.2/8.16.1.2) with SMTP id 19508Mdj004481; Tue, 5 Oct 2021 01:31:30 GMT DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=oracle.com; h=from : to : cc : subject : date : message-id : references : in-reply-to : content-type : content-transfer-encoding : mime-version; s=corp-2021-07-09; bh=yRr616GsCMj/hPjbZO+PyiN1iBxYOeZqEMJF6VDNbGA=; b=j1WgyHA1yVy2LsQByjmGdvMq1Rk90lcUCLPhkFnJC1yJup8G7y2FlH6LaNPgmwoA8/lr fxNSxOydEgy8Mcxdq/FZGM2jPrv+m9bACVd4HwgYcnv9PLJeVohutG13nd3bSqL9bgPj x+O0FeY1q7WIS/Ochh+ILhSVkgdqYZo/M0mS2kJ08QEnKZn7512gIRRh1QIvgTIrEZ8N NGcmQ/x/v2u8IDhGxwwFkyBLiad+g2/bYMf5xWVHycg1DE/OG6g+LWRoWbI/e96A1vPb kC0M2wJx4PHs3y1ojOIbcc2hrdTknHSO9VcEOw+spuK3C7Kn2WP6+4jrP4T5IKn0gKWY MA== Received: from userp3020.oracle.com (userp3020.oracle.com [156.151.31.79]) by mx0b-00069f02.pphosted.com with ESMTP id 3bg43dumjm-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Tue, 05 Oct 2021 01:31:28 +0000 Received: from pps.filterd (userp3020.oracle.com [127.0.0.1]) by userp3020.oracle.com (8.16.1.2/8.16.1.2) with SMTP id 1951UdFx141646; Tue, 5 Oct 2021 01:30:43 GMT Received: from nam11-bn8-obe.outbound.protection.outlook.com (mail-bn8nam11lp2173.outbound.protection.outlook.com [104.47.58.173]) by userp3020.oracle.com with ESMTP id 3bf16s9duq-3 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Tue, 05 Oct 2021 01:30:43 +0000 ARC-Seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=TvqBJhZbfXvll2TYLLOocYJiiHcnnNpYt0mOj1rFHZLXqPfPXRl+k1mqc+Cai7Lh6NVmwnMuWJtgWDcrku1fh9awMQMv9vgNvJYikZX2x4/o1pxD7bVYUyE2SGVIaFpMDWr9b1mlEptOWFizZyUHbyycM36A91USIO44frkSsMlaund24PP/hd6WkYFXGs4A7dau9fx0jZZj9MHqSlykB2kPj6j3InoxK/Ff/KAupvqXYhjLAA6JhcBPEzvCV8Yxlkmn6PeCN0Z26bvzRclL4swEAp95T8CB4wYcxPchLwusB67rYwlvvgZMp6G1SMD5m/Mt1EzLzQZYnlH6PNMn5g== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=yRr616GsCMj/hPjbZO+PyiN1iBxYOeZqEMJF6VDNbGA=; b=Di8C8X/Hm1q/JLBWgWCNBl6APYrC9QwfVCpS15faYW/lnNgUtfxyI3N66pyIGeMf4LzeN3U57jf8D0FLEDY5gwJ6Raks4xcMlGWsvzvftTUPttXvm6iFGm5c+6i6NKYc1toaADF9Y4Py18LzfUPHHxQbNQEBWTPLsgobedgk8tMv3nwl8P/b0ehUxaFyEG4hisHGoJ+8GkbMm8FzH5MSWrlP72wEiEu0o3C/Aj9tT5BVfHiShC+3e0bcxHxzucJNndMzCTEUBmkY1wDR/wPRCaAmcuEBWp8+XkOnz8j0UG9uM5Ij/j/TLjPyC7nZZ1E+gVNa/SKcp20HsXBFGAXPcQ== ARC-Authentication-Results: i=1; mx.microsoft.com 1; spf=pass smtp.mailfrom=oracle.com; dmarc=pass action=none header.from=oracle.com; dkim=pass header.d=oracle.com; arc=none DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=oracle.onmicrosoft.com; s=selector2-oracle-onmicrosoft-com; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=yRr616GsCMj/hPjbZO+PyiN1iBxYOeZqEMJF6VDNbGA=; b=MWHIFQLnPfSznpzCoTJ14soJ3sWlpOlgD8bKOsZ4/61sWopxeC8mN9lns4KWZ1i5Y5iy6BzPrFvVOlGLlBxYDXyplxmgZQZ2DuEWhBUud4hYBjW0ozYDOpry46rChkUnTfLRJrcou7x9BS+cQEdxMnShjlAQL2bj2pGVUDDLtd4= Received: from BL0PR10MB3011.namprd10.prod.outlook.com (2603:10b6:208:7e::29) by BLAPR10MB5044.namprd10.prod.outlook.com (2603:10b6:208:326::15) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.4566.15; Tue, 5 Oct 2021 01:30:35 +0000 Received: from BL0PR10MB3011.namprd10.prod.outlook.com ([fe80::6d61:54c2:40f0:93a]) by BL0PR10MB3011.namprd10.prod.outlook.com ([fe80::6d61:54c2:40f0:93a%5]) with mapi id 15.20.4566.017; Tue, 5 Oct 2021 01:30:35 +0000 From: Liam Howlett To: "maple-tree@lists.infradead.org" , "linux-mm@kvack.org" , "linux-kernel@vger.kernel.org" , Andrew Morton , David Hildenbrand , Douglas Gilbert CC: Song Liu , Davidlohr Bueso , "Paul E . McKenney" , Matthew Wilcox , David Rientjes , Axel Rasmussen , Suren Baghdasaryan , Vlastimil Babka , Rik van Riel , Peter Zijlstra Subject: [PATCH v3 11/66] mm: Remove rb tree. Thread-Topic: [PATCH v3 11/66] mm: Remove rb tree. Thread-Index: AQHXuYiYqoxw43whD0eItCvC0jCZqg== Date: Tue, 5 Oct 2021 01:30:35 +0000 Message-ID: <20211005012959.1110504-12-Liam.Howlett@oracle.com> References: <20211005012959.1110504-1-Liam.Howlett@oracle.com> In-Reply-To: <20211005012959.1110504-1-Liam.Howlett@oracle.com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-mailer: git-send-email 2.30.2 authentication-results: lists.infradead.org; dkim=none (message not signed) header.d=none;lists.infradead.org; dmarc=none action=none header.from=oracle.com; x-ms-publictraffictype: Email x-ms-office365-filtering-correlation-id: 3574ecb3-6a07-4b09-360d-08d9879fbb1c x-ms-traffictypediagnostic: BLAPR10MB5044: x-microsoft-antispam-prvs: x-ms-oob-tlc-oobclassifiers: OLM:551; x-ms-exchange-senderadcheck: 1 x-ms-exchange-antispam-relay: 0 x-microsoft-antispam: BCL:0; x-microsoft-antispam-message-info: GdX6VAYi+RLLRjHS5v1MOs20zQTnMxY4wtazX9pSAVBCAShWJL7u+UrMB2GpxCu37Aq3/iWUGiOHkFVCMIHxZzieC8aZvCC5sUC1GMKQvEM+T59x1BqW48kHZyZxd1GIXV/H6WqLLUNwY8n7KcTFJGQPRele70JJwMFAdnSRDYmlo8unzzkRl7+A97u47Wpbnf1G3xVYPHKrnqgEBA/hlBgTgAJ1KkwS7exiv1lrX+Mwom+WJkx3q+OmH5zBeKAg0ycXI0H09eLsUxIYZilRrjZbSkP4sj8xNdyajhM0Wy9LoK8pZq9mJoXKdxWPktCiyqoWVVmIzevr29Wb6eC3m1P/UMpPGVhhDKd1Ib1sunISBxpHq7nzaqKXFv8Emyr0IVjJ4IoDx8sPyOomgg6kjtV09Hyv4wsPV0waJy8BaOypvVx50lrFg+c+3zX9v9nneqAMRW+ODnbxyK+Z2PDXLiCD1VWGfDaWBgqedUO8d7nRtgEEeAQ+TB4Rl22SSQy8z+j2lTcRtHi0knjjb0p6Sy2NCBN3a08FXew6Nv+2LWFP8hRBbtyYCx2GA3Uf31CTNhrMo0CYdPF6pOgOvRPoVbC5zTTo20mmEu1+sin3pyjBeKycN4GKm1QKWhsg3reS6AO5/MS3Qtj76KDZ63dY96VG/9yRU7t1GEHj5mTBrNx9IhQ+W8L+V/g0n6H9xUBUJiCShaiH1dJfWAwco9XzFw== x-forefront-antispam-report: CIP:255.255.255.255;CTRY:;LANG:en;SCL:1;SRV:;IPV:NLI;SFV:NSPM;H:BL0PR10MB3011.namprd10.prod.outlook.com;PTR:;CAT:NONE;SFS:(366004)(6486002)(66446008)(122000001)(26005)(7416002)(2616005)(4326008)(508600001)(1076003)(38100700002)(91956017)(64756008)(66556008)(66946007)(5660300002)(66476007)(44832011)(38070700005)(76116006)(86362001)(110136005)(8676002)(71200400001)(316002)(8936002)(54906003)(2906002)(83380400001)(186003)(6512007)(6506007)(36756003)(30864003)(559001)(579004);DIR:OUT;SFP:1101; x-ms-exchange-antispam-messagedata-chunkcount: 1 x-ms-exchange-antispam-messagedata-0: =?iso-8859-1?Q?l6PnOPeZE+JIPKfIngb0E60J+XZw0O48OShvhGMMCb1qANF+XAe0OwRsrF?= =?iso-8859-1?Q?bWZSfhSSpKI/0rov4lK5GlnPqILvQ+/403dTZSoxos7hLssywUDLZ/iUzM?= =?iso-8859-1?Q?lmWCG5VpHU7Ioq8E3fiTIxgaaY/Kc6Bnsmh7fNMk2u68zsaZf4GfEW9l0q?= =?iso-8859-1?Q?MPPukvGaKnDPQjJYqXBSjTxNthMZ82nMuQfXkDh6VPBR+Gb5UuZ0J88YPx?= =?iso-8859-1?Q?vbDIYgU0lnF6UI4/7JeyENgjmMydjSgYKxabEVM/vNyOaoHhgYbw2sPQE5?= =?iso-8859-1?Q?fDvQWh7E8AMsWiW4b6y120jgk1ibHHxzyeNnE0SRpGu6hoaWk75BXDx+Wa?= =?iso-8859-1?Q?k97uH8NdQf7jxcbiptvHOOdUyJL9VD/cfHG3S/EdEZhHZPJFXkyXpxNSLO?= =?iso-8859-1?Q?Dl88n8KYKZZ5LgTti0vi87BY/LMiW5Y/rX8PDW7GOVBhxptXP4ZvOkWlTz?= =?iso-8859-1?Q?uu8FAfc2VDIQyVlil6c/nkd79dWcDLsjZYGC9Qrr0ltBzollUgH7HoltV6?= =?iso-8859-1?Q?erJigtWqEyQJkxA0wvFmyveNaVh4lxh6p1rV/fPHArBLrVal7L2tL9wmD7?= =?iso-8859-1?Q?kbqIeC1imhma/fZAMRdLGFoaONhW7HUAC5JcBT6sd9OoMdVOa6VRguZME3?= =?iso-8859-1?Q?TwTn5NP6+Gy5MhTGdL6aPVIBLwdpaCZ7y9qTQnA17pYK60vgTrmwjj/fAi?= =?iso-8859-1?Q?ISznASIS+onqo762H8OimC/uZnGgEhEtGwdKVm2N6Xw72WxZShDlaPFblZ?= =?iso-8859-1?Q?PD05KwJ9fb5b2hfZyhJnqTql+xpHKIrkurBfxqScfsjsIl0uNZyrGmZvMw?= =?iso-8859-1?Q?qiPvxyEfZZ3E0LozmEU8XUR6HjHmO1GChtCCtG8CLMtOL9Ypo8VCstpcIs?= =?iso-8859-1?Q?bpf6PE3DZ87UxvdKyH4NGREgb2s5+cfigDIO0+U50/vBEGmvXFydPwOxF2?= =?iso-8859-1?Q?UeLVqtMbjCbIjhup0/ePe4SR6BlkSzhAid57uvLzX692oEEntLUIAVyIYw?= =?iso-8859-1?Q?G/ERBcfVSABYDNPD5p6eaAluF1MCRLEIZIID+7RFXTvddd8Bh20DFu6hZX?= =?iso-8859-1?Q?I2VE0fyR7mnQrLASqbrK/u0vqUWZ2lLaYOkviy7tvZ8DDkJdNX5XmAP6MK?= =?iso-8859-1?Q?L4wGcfJwXKR3enLF+xmz80CB+YPZItqk3wKExYQp7DoHwU9blCBgKa+Tf5?= =?iso-8859-1?Q?ZGlY6q/j1kXXNi49IK0EPlOjfSXQTQfLtjYG4jGYpTpddVluFmByUhehOI?= =?iso-8859-1?Q?9nXlMV9OKKbQ64Ly/gEZ8XgmBESo11SxsV/cfIWA+FcemR1qmBQ9t9dXDM?= =?iso-8859-1?Q?frqsXhIo0XQsP0Zd3s98hJ21BIj58wNdwKhmYxA6s4feUoYI5Ir+uGHM+f?= =?iso-8859-1?Q?mX+g9W2sdO?= x-ms-exchange-transport-forked: True Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 X-OriginatorOrg: oracle.com X-MS-Exchange-CrossTenant-AuthAs: Internal X-MS-Exchange-CrossTenant-AuthSource: BL0PR10MB3011.namprd10.prod.outlook.com X-MS-Exchange-CrossTenant-Network-Message-Id: 3574ecb3-6a07-4b09-360d-08d9879fbb1c X-MS-Exchange-CrossTenant-originalarrivaltime: 05 Oct 2021 01:30:35.6385 (UTC) X-MS-Exchange-CrossTenant-fromentityheader: Hosted X-MS-Exchange-CrossTenant-id: 4e2c6054-71cb-48f1-bd6c-3a9705aca71b X-MS-Exchange-CrossTenant-mailboxtype: HOSTED X-MS-Exchange-CrossTenant-userprincipalname: Eu1jCUPD9i9QHo/34Dl1C61g1ZrX9yU/wxPNiZrovRbsKGjpuro+GUXV60O86GofQAm5tDZDEPagGIsmtcjAhA== X-MS-Exchange-Transport-CrossTenantHeadersStamped: BLAPR10MB5044 X-Proofpoint-Virus-Version: vendor=nai engine=6300 definitions=10127 signatures=668683 X-Proofpoint-Spam-Details: rule=notspam policy=default score=0 suspectscore=0 adultscore=0 phishscore=0 malwarescore=0 bulkscore=0 mlxlogscore=999 mlxscore=0 spamscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.12.0-2109230001 definitions=main-2110050007 X-Proofpoint-GUID: JaK4EUglOPLIaxVBECcgWDDOK-Dgl8sS X-Proofpoint-ORIG-GUID: JaK4EUglOPLIaxVBECcgWDDOK-Dgl8sS Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: "Liam R. Howlett" Remove the RB tree and start using the maple tree for vm_area_struct tracking. Drop validate_mm() calls in expand_upwards() and expand_downwards() as the lock is not held. Signed-off-by: Liam R. Howlett --- arch/x86/kernel/tboot.c | 1 - drivers/firmware/efi/efi.c | 1 - fs/proc/task_nommu.c | 55 ++-- include/linux/mm.h | 9 +- include/linux/mm_types.h | 21 +- kernel/fork.c | 8 - mm/init-mm.c | 2 - mm/mmap.c | 526 ++++++++----------------------------- mm/nommu.c | 95 +++---- mm/util.c | 8 + 10 files changed, 182 insertions(+), 544 deletions(-) diff --git a/arch/x86/kernel/tboot.c b/arch/x86/kernel/tboot.c index 6f978f722dff..121f28bb2209 100644 --- a/arch/x86/kernel/tboot.c +++ b/arch/x86/kernel/tboot.c @@ -97,7 +97,6 @@ void __init tboot_probe(void) =20 static pgd_t *tboot_pg_dir; static struct mm_struct tboot_mm =3D { - .mm_rb =3D RB_ROOT, .mm_mt =3D MTREE_INIT(mm_mt, MAPLE_ALLOC_RANGE), .pgd =3D swapper_pg_dir, .mm_users =3D ATOMIC_INIT(2), diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c index 1aaecca36e23..2ac01505592a 100644 --- a/drivers/firmware/efi/efi.c +++ b/drivers/firmware/efi/efi.c @@ -54,7 +54,6 @@ static unsigned long __initdata mem_reserve =3D EFI_INVAL= ID_TABLE_ADDR; static unsigned long __initdata rt_prop =3D EFI_INVALID_TABLE_ADDR; =20 struct mm_struct efi_mm =3D { - .mm_rb =3D RB_ROOT, .mm_mt =3D MTREE_INIT(mm_mt, MAPLE_ALLOC_RANGE), .mm_users =3D ATOMIC_INIT(2), .mm_count =3D ATOMIC_INIT(1), diff --git a/fs/proc/task_nommu.c b/fs/proc/task_nommu.c index a6d21fc0033c..8691a1216d1c 100644 --- a/fs/proc/task_nommu.c +++ b/fs/proc/task_nommu.c @@ -22,15 +22,13 @@ void task_mem(struct seq_file *m, struct mm_struct *mm) { struct vm_area_struct *vma; struct vm_region *region; - struct rb_node *p; unsigned long bytes =3D 0, sbytes =3D 0, slack =3D 0, size; - =20 - mmap_read_lock(mm); - for (p =3D rb_first(&mm->mm_rb); p; p =3D rb_next(p)) { - vma =3D rb_entry(p, struct vm_area_struct, vm_rb); + MA_STATE(mas, &mm->mm_mt, 0, 0); =20 + mmap_read_lock(mm); + rcu_read_lock(); + mas_for_each(&mas, vma, ULONG_MAX) { bytes +=3D kobjsize(vma); - region =3D vma->vm_region; if (region) { size =3D kobjsize(region); @@ -53,7 +51,7 @@ void task_mem(struct seq_file *m, struct mm_struct *mm) sbytes +=3D kobjsize(mm); else bytes +=3D kobjsize(mm); -=09 + if (current->fs && current->fs->users > 1) sbytes +=3D kobjsize(current->fs); else @@ -77,20 +75,21 @@ void task_mem(struct seq_file *m, struct mm_struct *mm) "Shared:\t%8lu bytes\n", bytes, slack, sbytes); =20 + rcu_read_unlock(); mmap_read_unlock(mm); } =20 unsigned long task_vsize(struct mm_struct *mm) { struct vm_area_struct *vma; - struct rb_node *p; unsigned long vsize =3D 0; + MA_STATE(mas, &mm->mm_mt, 0, 0); =20 mmap_read_lock(mm); - for (p =3D rb_first(&mm->mm_rb); p; p =3D rb_next(p)) { - vma =3D rb_entry(p, struct vm_area_struct, vm_rb); + rcu_read_lock(); + mas_for_each(&mas, vma, ULONG_MAX) vsize +=3D vma->vm_end - vma->vm_start; - } + rcu_read_unlock(); mmap_read_unlock(mm); return vsize; } @@ -101,12 +100,12 @@ unsigned long task_statm(struct mm_struct *mm, { struct vm_area_struct *vma; struct vm_region *region; - struct rb_node *p; unsigned long size =3D kobjsize(mm); + MA_STATE(mas, &mm->mm_mt, 0, 0); =20 mmap_read_lock(mm); - for (p =3D rb_first(&mm->mm_rb); p; p =3D rb_next(p)) { - vma =3D rb_entry(p, struct vm_area_struct, vm_rb); + rcu_read_lock(); + mas_for_each(&mas, vma, ULONG_MAX) { size +=3D kobjsize(vma); region =3D vma->vm_region; if (region) { @@ -119,6 +118,7 @@ unsigned long task_statm(struct mm_struct *mm, >> PAGE_SHIFT; *data =3D (PAGE_ALIGN(mm->start_stack) - (mm->start_data & PAGE_MASK)) >> PAGE_SHIFT; + rcu_read_unlock(); mmap_read_unlock(mm); size >>=3D PAGE_SHIFT; size +=3D *text + *data; @@ -190,17 +190,20 @@ static int nommu_vma_show(struct seq_file *m, struct = vm_area_struct *vma) */ static int show_map(struct seq_file *m, void *_p) { - struct rb_node *p =3D _p; - - return nommu_vma_show(m, rb_entry(p, struct vm_area_struct, vm_rb)); + return nommu_vma_show(m, _p); } =20 static void *m_start(struct seq_file *m, loff_t *pos) { struct proc_maps_private *priv =3D m->private; struct mm_struct *mm; - struct rb_node *p; - loff_t n =3D *pos; + struct vm_area_struct *vma; + unsigned long addr =3D *pos; + MA_STATE(mas, &priv->mm->mm_mt, addr, addr); + + /* See m_next(). Zero at the start or after lseek. */ + if (addr =3D=3D -1UL) + return NULL; =20 /* pin the task and mm whilst we play with them */ priv->task =3D get_proc_task(priv->inode); @@ -216,14 +219,12 @@ static void *m_start(struct seq_file *m, loff_t *pos) return ERR_PTR(-EINTR); } =20 - /* start from the Nth VMA */ - for (p =3D rb_first(&mm->mm_rb); p; p =3D rb_next(p)) - if (n-- =3D=3D 0) - return p; + /* start the next element from addr */ + vma =3D mas_find(&mas, ULONG_MAX); =20 mmap_read_unlock(mm); mmput(mm); - return NULL; + return vma; } =20 static void m_stop(struct seq_file *m, void *_vml) @@ -242,10 +243,10 @@ static void m_stop(struct seq_file *m, void *_vml) =20 static void *m_next(struct seq_file *m, void *_p, loff_t *pos) { - struct rb_node *p =3D _p; + struct vm_area_struct *vma =3D _p; =20 - (*pos)++; - return p ? rb_next(p) : NULL; + *pos =3D vma->vm_end; + return vma->vm_next; } =20 static const struct seq_operations proc_pid_maps_ops =3D { diff --git a/include/linux/mm.h b/include/linux/mm.h index 73a52aba448f..f2d608283a9a 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2555,8 +2555,6 @@ extern int __split_vma(struct mm_struct *, struct vm_= area_struct *, extern int split_vma(struct mm_struct *, struct vm_area_struct *, unsigned long addr, int new_below); extern int insert_vm_struct(struct mm_struct *, struct vm_area_struct *); -extern void __vma_link_rb(struct mm_struct *, struct vm_area_struct *, - struct rb_node **, struct rb_node *); extern void unlink_file_vma(struct vm_area_struct *); extern struct vm_area_struct *copy_vma(struct vm_area_struct **, unsigned long addr, unsigned long len, pgoff_t pgoff, @@ -2712,12 +2710,7 @@ struct vm_area_struct *find_vma_intersection(struct = mm_struct *mm, static inline struct vm_area_struct *vma_lookup(struct mm_struct *mm, unsigned long addr= ) { - struct vm_area_struct *vma =3D find_vma(mm, addr); - - if (vma && addr < vma->vm_start) - vma =3D NULL; - - return vma; + return mtree_load(&mm->mm_mt, addr); } =20 static inline unsigned long vm_start_gap(struct vm_area_struct *vma) diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 2fc27dff8b6c..f5de9734fcda 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -326,19 +326,6 @@ struct vm_area_struct { =20 /* linked list of VM areas per task, sorted by address */ struct vm_area_struct *vm_next, *vm_prev; - - struct rb_node vm_rb; - - /* - * Largest free memory gap in bytes to the left of this VMA. - * Either between this VMA and vma->vm_prev, or between one of the - * VMAs below us in the VMA rbtree and its ->vm_prev. This helps - * get_unmapped_area find a free area of the right size. - */ - unsigned long rb_subtree_gap; - - /* Second cache line starts here. */ - struct mm_struct *vm_mm; /* The address space we belong to. */ =20 /* @@ -369,13 +356,14 @@ struct vm_area_struct { =20 /* Function pointers to deal with this struct. */ const struct vm_operations_struct *vm_ops; - /* Information about our backing store: */ unsigned long vm_pgoff; /* Offset (within vm_file) in PAGE_SIZE - units */ - struct file * vm_file; /* File we map to (can be NULL). */ + * units + */ + struct file *vm_file; /* File we map to (can be NULL). */ void * vm_private_data; /* was vm_pte (shared mem) */ =20 + #ifdef CONFIG_SWAP atomic_long_t swap_readahead_info; #endif @@ -404,7 +392,6 @@ struct mm_struct { struct { struct vm_area_struct *mmap; /* list of VMAs */ struct maple_tree mm_mt; - struct rb_root mm_rb; u64 vmacache_seqnum; /* per-thread vmacache */ #ifdef CONFIG_MMU unsigned long (*get_unmapped_area) (struct file *filp, diff --git a/kernel/fork.c b/kernel/fork.c index b8b7063d94e8..2eeb51346821 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -490,7 +490,6 @@ static __latent_entropy int dup_mmap(struct mm_struct *= mm, struct mm_struct *oldmm) { struct vm_area_struct *mpnt, *tmp, *prev, **pprev; - struct rb_node **rb_link, *rb_parent; int retval; unsigned long charge =3D 0; MA_STATE(old_mas, &oldmm->mm_mt, 0, 0); @@ -517,8 +516,6 @@ static __latent_entropy int dup_mmap(struct mm_struct *= mm, mm->exec_vm =3D oldmm->exec_vm; mm->stack_vm =3D oldmm->stack_vm; =20 - rb_link =3D &mm->mm_rb.rb_node; - rb_parent =3D NULL; pprev =3D &mm->mmap; retval =3D ksm_fork(mm, oldmm); if (retval) @@ -612,10 +609,6 @@ static __latent_entropy int dup_mmap(struct mm_struct = *mm, tmp->vm_prev =3D prev; prev =3D tmp; =20 - __vma_link_rb(mm, tmp, rb_link, rb_parent); - rb_link =3D &tmp->vm_rb.rb_right; - rb_parent =3D &tmp->vm_rb; - /* Link the vma into the MT */ mas_lock(&mas); mas.index =3D tmp->vm_start; @@ -1062,7 +1055,6 @@ static struct mm_struct *mm_init(struct mm_struct *mm= , struct task_struct *p, struct user_namespace *user_ns) { mm->mmap =3D NULL; - mm->mm_rb =3D RB_ROOT; mt_init_flags(&mm->mm_mt, MAPLE_ALLOC_RANGE); mm->vmacache_seqnum =3D 0; atomic_set(&mm->mm_users, 1); diff --git a/mm/init-mm.c b/mm/init-mm.c index e4a058f7e8b6..e5b2c618cbb6 100644 --- a/mm/init-mm.c +++ b/mm/init-mm.c @@ -1,6 +1,5 @@ // SPDX-License-Identifier: GPL-2.0 #include -#include #include #include #include @@ -28,7 +27,6 @@ * and size this cpu_bitmask to NR_CPUS. */ struct mm_struct init_mm =3D { - .mm_rb =3D RB_ROOT, .mm_mt =3D MTREE_INIT(mm_mt, MAPLE_ALLOC_RANGE), .pgd =3D swapper_pg_dir, .mm_users =3D ATOMIC_INIT(2), diff --git a/mm/mmap.c b/mm/mmap.c index 3da639586715..c51d739d7411 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -38,7 +38,6 @@ #include #include #include -#include #include #include #include @@ -288,93 +287,6 @@ SYSCALL_DEFINE1(brk, unsigned long, brk) return origbrk; } =20 -static inline unsigned long vma_compute_gap(struct vm_area_struct *vma) -{ - unsigned long gap, prev_end; - - /* - * Note: in the rare case of a VM_GROWSDOWN above a VM_GROWSUP, we - * allow two stack_guard_gaps between them here, and when choosing - * an unmapped area; whereas when expanding we only require one. - * That's a little inconsistent, but keeps the code here simpler. - */ - gap =3D vm_start_gap(vma); - if (vma->vm_prev) { - prev_end =3D vm_end_gap(vma->vm_prev); - if (gap > prev_end) - gap -=3D prev_end; - else - gap =3D 0; - } - return gap; -} - -#ifdef CONFIG_DEBUG_VM_RB -static unsigned long vma_compute_subtree_gap(struct vm_area_struct *vma) -{ - unsigned long max =3D vma_compute_gap(vma), subtree_gap; - if (vma->vm_rb.rb_left) { - subtree_gap =3D rb_entry(vma->vm_rb.rb_left, - struct vm_area_struct, vm_rb)->rb_subtree_gap; - if (subtree_gap > max) - max =3D subtree_gap; - } - if (vma->vm_rb.rb_right) { - subtree_gap =3D rb_entry(vma->vm_rb.rb_right, - struct vm_area_struct, vm_rb)->rb_subtree_gap; - if (subtree_gap > max) - max =3D subtree_gap; - } - return max; -} - -static int browse_rb(struct mm_struct *mm) -{ - struct rb_root *root =3D &mm->mm_rb; - int i =3D 0, j, bug =3D 0; - struct rb_node *nd, *pn =3D NULL; - unsigned long prev =3D 0, pend =3D 0; - - for (nd =3D rb_first(root); nd; nd =3D rb_next(nd)) { - struct vm_area_struct *vma; - vma =3D rb_entry(nd, struct vm_area_struct, vm_rb); - if (vma->vm_start < prev) { - pr_emerg("vm_start %lx < prev %lx\n", - vma->vm_start, prev); - bug =3D 1; - } - if (vma->vm_start < pend) { - pr_emerg("vm_start %lx < pend %lx\n", - vma->vm_start, pend); - bug =3D 1; - } - if (vma->vm_start > vma->vm_end) { - pr_emerg("vm_start %lx > vm_end %lx\n", - vma->vm_start, vma->vm_end); - bug =3D 1; - } - spin_lock(&mm->page_table_lock); - if (vma->rb_subtree_gap !=3D vma_compute_subtree_gap(vma)) { - pr_emerg("free gap %lx, correct %lx\n", - vma->rb_subtree_gap, - vma_compute_subtree_gap(vma)); - bug =3D 1; - } - spin_unlock(&mm->page_table_lock); - i++; - pn =3D nd; - prev =3D vma->vm_start; - pend =3D vma->vm_end; - } - j =3D 0; - for (nd =3D pn; nd; nd =3D rb_prev(nd)) - j++; - if (i !=3D j) { - pr_emerg("backwards %d, forwards %d\n", j, i); - bug =3D 1; - } - return bug ? -1 : i; -} #if defined(CONFIG_DEBUG_MAPLE_TREE) extern void mt_validate(struct maple_tree *mt); extern void mt_dump(const struct maple_tree *mt); @@ -403,17 +315,25 @@ static void validate_mm_mt(struct mm_struct *mm) dump_stack(); #ifdef CONFIG_DEBUG_VM dump_vma(vma_mt); - pr_emerg("and next in rb\n"); + pr_emerg("and vm_next\n"); dump_vma(vma->vm_next); -#endif +#endif // CONFIG_DEBUG_VM pr_emerg("mt piv: %px %lu - %lu\n", vma_mt, mas.index, mas.last); pr_emerg("mt vma: %px %lu - %lu\n", vma_mt, vma_mt->vm_start, vma_mt->vm_end); - pr_emerg("rb vma: %px %lu - %lu\n", vma, + if (vma->vm_prev) { + pr_emerg("ll prev: %px %lu - %lu\n", + vma->vm_prev, vma->vm_prev->vm_start, + vma->vm_prev->vm_end); + } + pr_emerg("ll vma: %px %lu - %lu\n", vma, vma->vm_start, vma->vm_end); - pr_emerg("rb->next =3D %px %lu - %lu\n", vma->vm_next, - vma->vm_next->vm_start, vma->vm_next->vm_end); + if (vma->vm_next) { + pr_emerg("ll next: %px %lu - %lu\n", + vma->vm_next, vma->vm_next->vm_start, + vma->vm_next->vm_end); + } =20 mt_dump(mas.tree); if (vma_mt->vm_end !=3D mas.last + 1) { @@ -439,21 +359,6 @@ static void validate_mm_mt(struct mm_struct *mm) rcu_read_unlock(); mt_validate(&mm->mm_mt); } -#else -#define validate_mm_mt(root) do { } while (0) -#endif -static void validate_mm_rb(struct rb_root *root, struct vm_area_struct *ig= nore) -{ - struct rb_node *nd; - - for (nd =3D rb_first(root); nd; nd =3D rb_next(nd)) { - struct vm_area_struct *vma; - vma =3D rb_entry(nd, struct vm_area_struct, vm_rb); - VM_BUG_ON_VMA(vma !=3D ignore && - vma->rb_subtree_gap !=3D vma_compute_subtree_gap(vma), - vma); - } -} =20 static void validate_mm(struct mm_struct *mm) { @@ -462,6 +367,8 @@ static void validate_mm(struct mm_struct *mm) unsigned long highest_address =3D 0; struct vm_area_struct *vma =3D mm->mmap; =20 + validate_mm_mt(mm); + while (vma) { struct anon_vma *anon_vma =3D vma->anon_vma; struct anon_vma_chain *avc; @@ -486,80 +393,13 @@ static void validate_mm(struct mm_struct *mm) mm->highest_vm_end, highest_address); bug =3D 1; } - i =3D browse_rb(mm); - if (i !=3D mm->map_count) { - if (i !=3D -1) - pr_emerg("map_count %d rb %d\n", mm->map_count, i); - bug =3D 1; - } VM_BUG_ON_MM(bug, mm); } -#else -#define validate_mm_rb(root, ignore) do { } while (0) + +#else // !CONFIG_DEBUG_MAPLE_TREE #define validate_mm_mt(root) do { } while (0) #define validate_mm(mm) do { } while (0) -#endif - -RB_DECLARE_CALLBACKS_MAX(static, vma_gap_callbacks, - struct vm_area_struct, vm_rb, - unsigned long, rb_subtree_gap, vma_compute_gap) - -/* - * Update augmented rbtree rb_subtree_gap values after vma->vm_start or - * vma->vm_prev->vm_end values changed, without modifying the vma's positi= on - * in the rbtree. - */ -static void vma_gap_update(struct vm_area_struct *vma) -{ - /* - * As it turns out, RB_DECLARE_CALLBACKS_MAX() already created - * a callback function that does exactly what we want. - */ - vma_gap_callbacks_propagate(&vma->vm_rb, NULL); -} - -static inline void vma_rb_insert(struct vm_area_struct *vma, - struct rb_root *root) -{ - /* All rb_subtree_gap values must be consistent prior to insertion */ - validate_mm_rb(root, NULL); - - rb_insert_augmented(&vma->vm_rb, root, &vma_gap_callbacks); -} - -static void __vma_rb_erase(struct vm_area_struct *vma, struct rb_root *roo= t) -{ - /* - * Note rb_erase_augmented is a fairly large inline function, - * so make sure we instantiate it only once with our desired - * augmented rbtree callbacks. - */ - rb_erase_augmented(&vma->vm_rb, root, &vma_gap_callbacks); -} - -static __always_inline void vma_rb_erase_ignore(struct vm_area_struct *vma= , - struct rb_root *root, - struct vm_area_struct *ignore) -{ - /* - * All rb_subtree_gap values must be consistent prior to erase, - * with the possible exception of - * - * a. the "next" vma being erased if next->vm_start was reduced in - * __vma_adjust() -> __vma_unlink() - * b. the vma being erased in detach_vmas_to_be_unmapped() -> - * vma_rb_erase() - */ - validate_mm_rb(root, ignore); - - __vma_rb_erase(vma, root); -} - -static __always_inline void vma_rb_erase(struct vm_area_struct *vma, - struct rb_root *root) -{ - vma_rb_erase_ignore(vma, root, vma); -} +#endif // CONFIG_DEBUG_MAPLE_TREE =20 /* * vma has some anon_vma assigned, and is already inserted on that @@ -593,39 +433,26 @@ anon_vma_interval_tree_post_update_vma(struct vm_area= _struct *vma) anon_vma_interval_tree_insert(avc, &avc->anon_vma->rb_root); } =20 -static int find_vma_links(struct mm_struct *mm, unsigned long addr, - unsigned long end, struct vm_area_struct **pprev, - struct rb_node ***rb_link, struct rb_node **rb_parent) +/* + * range_has_overlap() - Check the @start - @end range for overlapping VMA= s and + * sets up a pointer to the previous VMA + * @mm: the mm struct + * @start: the start address of the range + * @end: the end address of the range + * @pprev: the pointer to the pointer of the previous VMA + * + * Returns: True if there is an overlapping VMA, false otherwise + */ +static inline +bool range_has_overlap(struct mm_struct *mm, unsigned long start, + unsigned long end, struct vm_area_struct **pprev) { - struct rb_node **__rb_link, *__rb_parent, *rb_prev; - - mmap_assert_locked(mm); - __rb_link =3D &mm->mm_rb.rb_node; - rb_prev =3D __rb_parent =3D NULL; + struct vm_area_struct *existing; =20 - while (*__rb_link) { - struct vm_area_struct *vma_tmp; - - __rb_parent =3D *__rb_link; - vma_tmp =3D rb_entry(__rb_parent, struct vm_area_struct, vm_rb); - - if (vma_tmp->vm_end > addr) { - /* Fail if an existing vma overlaps the area */ - if (vma_tmp->vm_start < end) - return -ENOMEM; - __rb_link =3D &__rb_parent->rb_left; - } else { - rb_prev =3D __rb_parent; - __rb_link =3D &__rb_parent->rb_right; - } - } - - *pprev =3D NULL; - if (rb_prev) - *pprev =3D rb_entry(rb_prev, struct vm_area_struct, vm_rb); - *rb_link =3D __rb_link; - *rb_parent =3D __rb_parent; - return 0; + MA_STATE(mas, &mm->mm_mt, start, start); + existing =3D mas_find(&mas, end - 1); + *pprev =3D mas_prev(&mas, 0); + return existing ? true : false; } =20 /* @@ -652,8 +479,6 @@ static inline struct vm_area_struct *vma_next(struct mm= _struct *mm, * @start: The start of the range. * @len: The length of the range. * @pprev: pointer to the pointer that will be set to previous vm_area_str= uct - * @rb_link: the rb_node - * @rb_parent: the parent rb_node * * Find all the vm_area_struct that overlap from @start to * @end and munmap them. Set @pprev to the previous vm_area_struct. @@ -662,76 +487,41 @@ static inline struct vm_area_struct *vma_next(struct = mm_struct *mm, */ static inline int munmap_vma_range(struct mm_struct *mm, unsigned long start, unsigned long = len, - struct vm_area_struct **pprev, struct rb_node ***link, - struct rb_node **parent, struct list_head *uf) + struct vm_area_struct **pprev, struct list_head *uf) { - - while (find_vma_links(mm, start, start + len, pprev, link, parent)) + // Needs optimization. + while (range_has_overlap(mm, start, start + len, pprev)) if (do_munmap(mm, start, len, uf)) return -ENOMEM; - return 0; } static unsigned long count_vma_pages_range(struct mm_struct *mm, unsigned long addr, unsigned long end) { unsigned long nr_pages =3D 0; - unsigned long nr_mt_pages =3D 0; struct vm_area_struct *vma; + unsigned long vm_start, vm_end; + MA_STATE(mas, &mm->mm_mt, addr, addr); =20 /* Find first overlapping mapping */ - vma =3D find_vma_intersection(mm, addr, end); + vma =3D mas_find(&mas, end - 1); if (!vma) return 0; =20 - nr_pages =3D (min(end, vma->vm_end) - - max(addr, vma->vm_start)) >> PAGE_SHIFT; + vm_start =3D vma->vm_start; + vm_end =3D vma->vm_end; + nr_pages =3D (min(end, vm_end) - max(addr, vm_start)) >> PAGE_SHIFT; =20 /* Iterate over the rest of the overlaps */ - for (vma =3D vma->vm_next; vma; vma =3D vma->vm_next) { - unsigned long overlap_len; - - if (vma->vm_start > end) - break; - - overlap_len =3D min(end, vma->vm_end) - vma->vm_start; - nr_pages +=3D overlap_len >> PAGE_SHIFT; + mas_for_each(&mas, vma, end) { + vm_start =3D vma->vm_start; + vm_end =3D vma->vm_end; + nr_pages +=3D (min(end, vm_end) - vm_start) >> PAGE_SHIFT; } =20 - mt_for_each(&mm->mm_mt, vma, addr, end) { - nr_mt_pages +=3D - (min(end, vma->vm_end) - vma->vm_start) >> PAGE_SHIFT; - } - - VM_BUG_ON_MM(nr_pages !=3D nr_mt_pages, mm); - return nr_pages; } =20 -void __vma_link_rb(struct mm_struct *mm, struct vm_area_struct *vma, - struct rb_node **rb_link, struct rb_node *rb_parent) -{ - /* Update tracking information for the gap following the new vma. */ - if (vma->vm_next) - vma_gap_update(vma->vm_next); - else - mm->highest_vm_end =3D vm_end_gap(vma); - - /* - * vma->vm_prev wasn't known when we followed the rbtree to find the - * correct insertion point for that vma. As a result, we could not - * update the vma vm_rb parents rb_subtree_gap values on the way down. - * So, we first insert the vma with a zero rb_subtree_gap value - * (to be consistent with what we did on the way down), and then - * immediately update the gap to the correct value. Finally we - * rebalance the rbtree after all augmented values have been set. - */ - rb_link_node(&vma->vm_rb, rb_parent, rb_link); - vma->rb_subtree_gap =3D 0; - vma_gap_update(vma); - vma_rb_insert(vma, &mm->mm_rb); -} - static void __vma_link_file(struct vm_area_struct *vma) { struct file *file; @@ -782,19 +572,8 @@ void vma_mt_store(struct mm_struct *mm, struct vm_area= _struct *vma) GFP_KERNEL); } =20 -static void -__vma_link(struct mm_struct *mm, struct vm_area_struct *vma, - struct vm_area_struct *prev, struct rb_node **rb_link, - struct rb_node *rb_parent) -{ - vma_mt_store(mm, vma); - __vma_link_list(mm, vma, prev); - __vma_link_rb(mm, vma, rb_link, rb_parent); -} - static void vma_link(struct mm_struct *mm, struct vm_area_struct *vma, - struct vm_area_struct *prev, struct rb_node **rb_link, - struct rb_node *rb_parent) + struct vm_area_struct *prev) { struct address_space *mapping =3D NULL; =20 @@ -803,7 +582,8 @@ static void vma_link(struct mm_struct *mm, struct vm_ar= ea_struct *vma, i_mmap_lock_write(mapping); } =20 - __vma_link(mm, vma, prev, rb_link, rb_parent); + vma_mt_store(mm, vma); + __vma_link_list(mm, vma, prev); __vma_link_file(vma); =20 if (mapping) @@ -815,30 +595,17 @@ static void vma_link(struct mm_struct *mm, struct vm_= area_struct *vma, =20 /* * Helper for vma_adjust() in the split_vma insert case: insert a vma into= the - * mm's list and rbtree. It has already been inserted into the interval t= ree. + * mm's list and the mm tree. It has already been inserted into the inter= val tree. */ static void __insert_vm_struct(struct mm_struct *mm, struct vm_area_struct= *vma) { struct vm_area_struct *prev; - struct rb_node **rb_link, *rb_parent; =20 - if (find_vma_links(mm, vma->vm_start, vma->vm_end, - &prev, &rb_link, &rb_parent)) - BUG(); - __vma_link(mm, vma, prev, rb_link, rb_parent); + vma_mt_store(mm, vma); + __vma_link_list(mm, vma, prev); mm->map_count++; } =20 -static __always_inline void __vma_unlink(struct mm_struct *mm, - struct vm_area_struct *vma, - struct vm_area_struct *ignore) -{ - vma_rb_erase_ignore(vma, &mm->mm_rb, ignore); - __vma_unlink_list(mm, vma); - /* Kill the cache */ - vmacache_invalidate(mm); -} - /* * We cannot adjust vm_start, vm_end, vm_pgoff fields of a vma that * is already present in an i_mmap tree without adjusting the tree. @@ -856,13 +623,10 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned= long start, struct rb_root_cached *root =3D NULL; struct anon_vma *anon_vma =3D NULL; struct file *file =3D vma->vm_file; - bool start_changed =3D false, end_changed =3D false; + bool vma_changed =3D false; long adjust_next =3D 0; int remove_next =3D 0; =20 - validate_mm(mm); - validate_mm_mt(mm); - if (next && !insert) { struct vm_area_struct *exporter =3D NULL, *importer =3D NULL; =20 @@ -988,21 +752,23 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned= long start, } =20 if (start !=3D vma->vm_start) { - unsigned long old_start =3D vma->vm_start; + if (vma->vm_start < start) + vma_mt_szero(mm, vma->vm_start, start); + else + vma_changed =3D true; vma->vm_start =3D start; - if (old_start < start) - vma_mt_szero(mm, old_start, start); - start_changed =3D true; } if (end !=3D vma->vm_end) { - unsigned long old_end =3D vma->vm_end; + if (vma->vm_end > end) + vma_mt_szero(mm, end, vma->vm_end); + else + vma_changed =3D true; vma->vm_end =3D end; - if (old_end > end) - vma_mt_szero(mm, end, old_end); - end_changed =3D true; + if (!next) + mm->highest_vm_end =3D vm_end_gap(vma); } =20 - if (end_changed || start_changed) + if (vma_changed) vma_mt_store(mm, vma); =20 vma->vm_pgoff =3D pgoff; @@ -1020,25 +786,9 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned= long start, } =20 if (remove_next) { - /* - * vma_merge has merged next into vma, and needs - * us to remove next before dropping the locks. - * Since we have expanded over this vma, the maple tree will - * have overwritten by storing the value - */ - if (remove_next !=3D 3) - __vma_unlink(mm, next, next); - else - /* - * vma is not before next if they've been - * swapped. - * - * pre-swap() next->vm_start was reduced so - * tell validate_mm_rb to ignore pre-swap() - * "next" (which is stored in post-swap() - * "vma"). - */ - __vma_unlink(mm, next, vma); + __vma_unlink_list(mm, next); + /* Kill the cache */ + vmacache_invalidate(mm); if (file) __remove_shared_vm_struct(next, file, mapping); } else if (insert) { @@ -1048,15 +798,6 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned= long start, * (it may either follow vma or precede it). */ __insert_vm_struct(mm, insert); - } else { - if (start_changed) - vma_gap_update(vma); - if (end_changed) { - if (!next) - mm->highest_vm_end =3D vm_end_gap(vma); - else if (!adjust_next) - vma_gap_update(next); - } } =20 if (anon_vma) { @@ -1114,10 +855,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned= long start, remove_next =3D 1; end =3D next->vm_end; goto again; - } - else if (next) - vma_gap_update(next); - else { + } else if (!next) { /* * If remove_next =3D=3D 2 we obviously can't * reach this path. @@ -1144,8 +882,6 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned = long start, uprobe_mmap(insert); =20 validate_mm(mm); - validate_mm_mt(mm); - return 0; } =20 @@ -1292,7 +1028,6 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm= , struct vm_area_struct *area, *next; int err; =20 - validate_mm_mt(mm); /* * We later require that vma->vm_flags =3D=3D vm_flags, * so this tests vma->vm_flags & VM_SPECIAL, too. @@ -1368,7 +1103,6 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm= , khugepaged_enter_vma_merge(area, vm_flags); return area; } - validate_mm_mt(mm); =20 return NULL; } @@ -1538,6 +1272,7 @@ unsigned long do_mmap(struct file *file, unsigned lon= g addr, vm_flags_t vm_flags; int pkey =3D 0; =20 + validate_mm(mm); *populate =3D 0; =20 if (!len) @@ -1848,10 +1583,8 @@ unsigned long mmap_region(struct file *file, unsigne= d long addr, struct mm_struct *mm =3D current->mm; struct vm_area_struct *vma, *prev, *merge; int error; - struct rb_node **rb_link, *rb_parent; unsigned long charged =3D 0; =20 - validate_mm_mt(mm); /* Check against address space limit. */ if (!may_expand_vm(mm, vm_flags, len >> PAGE_SHIFT)) { unsigned long nr_pages; @@ -1867,8 +1600,8 @@ unsigned long mmap_region(struct file *file, unsigned= long addr, return -ENOMEM; } =20 - /* Clear old maps, set up prev, rb_link, rb_parent, and uf */ - if (munmap_vma_range(mm, addr, len, &prev, &rb_link, &rb_parent, uf)) + /* Clear old maps, set up prev and uf */ + if (munmap_vma_range(mm, addr, len, &prev, uf)) return -ENOMEM; /* * Private writable mapping: check memory availability @@ -1966,7 +1699,7 @@ unsigned long mmap_region(struct file *file, unsigned= long addr, goto free_vma; } =20 - vma_link(mm, vma, prev, rb_link, rb_parent); + vma_link(mm, vma, prev); /* Once vma denies write, undo our temporary denial count */ unmap_writable: if (file && vm_flags & VM_SHARED) @@ -1999,7 +1732,6 @@ unsigned long mmap_region(struct file *file, unsigned= long addr, =20 vma_set_page_prot(vma); =20 - validate_mm_mt(mm); return addr; =20 unmap_and_free_vma: @@ -2016,7 +1748,6 @@ unsigned long mmap_region(struct file *file, unsigned= long addr, unacct_error: if (charged) vm_unacct_memory(charged); - validate_mm_mt(mm); return error; } =20 @@ -2307,9 +2038,6 @@ find_vma_prev(struct mm_struct *mm, unsigned long add= r, =20 rcu_read_lock(); vma =3D mas_find(&mas, ULONG_MAX); - if (!vma) - mas_reset(&mas); - *pprev =3D mas_prev(&mas, 0); rcu_read_unlock(); return vma; @@ -2373,7 +2101,6 @@ int expand_upwards(struct vm_area_struct *vma, unsign= ed long address) unsigned long gap_addr; int error =3D 0; =20 - validate_mm_mt(mm); if (!(vma->vm_flags & VM_GROWSUP)) return -EFAULT; =20 @@ -2420,15 +2147,13 @@ int expand_upwards(struct vm_area_struct *vma, unsi= gned long address) error =3D acct_stack_growth(vma, size, grow); if (!error) { /* - * vma_gap_update() doesn't support concurrent - * updates, but we only hold a shared mmap_lock - * lock here, so we need to protect against - * concurrent vma expansions. - * anon_vma_lock_write() doesn't help here, as - * we don't guarantee that all growable vmas - * in a mm share the same root anon vma. - * So, we reuse mm->page_table_lock to guard - * against concurrent vma expansions. + * We only hold a shared mmap_lock lock here, so + * we need to protect against concurrent vma + * expansions. anon_vma_lock_write() doesn't + * help here, as we don't guarantee that all + * growable vmas in a mm share the same root + * anon vma. So, we reuse mm->page_table_lock + * to guard against concurrent vma expansions. */ spin_lock(&mm->page_table_lock); if (vma->vm_flags & VM_LOCKED) @@ -2436,10 +2161,9 @@ int expand_upwards(struct vm_area_struct *vma, unsig= ned long address) vm_stat_account(mm, vma->vm_flags, grow); anon_vma_interval_tree_pre_update_vma(vma); vma->vm_end =3D address; + vma_mt_store(mm, vma); anon_vma_interval_tree_post_update_vma(vma); - if (vma->vm_next) - vma_gap_update(vma->vm_next); - else + if (!vma->vm_next) mm->highest_vm_end =3D vm_end_gap(vma); spin_unlock(&mm->page_table_lock); =20 @@ -2449,8 +2173,6 @@ int expand_upwards(struct vm_area_struct *vma, unsign= ed long address) } anon_vma_unlock_write(vma->anon_vma); khugepaged_enter_vma_merge(vma, vma->vm_flags); - validate_mm(mm); - validate_mm_mt(mm); return error; } #endif /* CONFIG_STACK_GROWSUP || CONFIG_IA64 */ @@ -2458,14 +2180,12 @@ int expand_upwards(struct vm_area_struct *vma, unsi= gned long address) /* * vma is the first one with address < vma->vm_start. Have to extend vma. */ -int expand_downwards(struct vm_area_struct *vma, - unsigned long address) +int expand_downwards(struct vm_area_struct *vma, unsigned long address) { struct mm_struct *mm =3D vma->vm_mm; struct vm_area_struct *prev; int error =3D 0; =20 - validate_mm(mm); address &=3D PAGE_MASK; if (address < mmap_min_addr) return -EPERM; @@ -2502,15 +2222,13 @@ int expand_downwards(struct vm_area_struct *vma, error =3D acct_stack_growth(vma, size, grow); if (!error) { /* - * vma_gap_update() doesn't support concurrent - * updates, but we only hold a shared mmap_lock - * lock here, so we need to protect against - * concurrent vma expansions. - * anon_vma_lock_write() doesn't help here, as - * we don't guarantee that all growable vmas - * in a mm share the same root anon vma. - * So, we reuse mm->page_table_lock to guard - * against concurrent vma expansions. + * We only hold a shared mmap_lock lock here, so + * we need to protect against concurrent vma + * expansions. anon_vma_lock_write() doesn't + * help here, as we don't guarantee that all + * growable vmas in a mm share the same root + * anon vma. So, we reuse mm->page_table_lock + * to guard against concurrent vma expansions. */ spin_lock(&mm->page_table_lock); if (vma->vm_flags & VM_LOCKED) @@ -2522,7 +2240,6 @@ int expand_downwards(struct vm_area_struct *vma, /* Overwrite old entry in mtree. */ vma_mt_store(mm, vma); anon_vma_interval_tree_post_update_vma(vma); - vma_gap_update(vma); spin_unlock(&mm->page_table_lock); =20 perf_event_mmap(vma); @@ -2531,7 +2248,6 @@ int expand_downwards(struct vm_area_struct *vma, } anon_vma_unlock_write(vma->anon_vma); khugepaged_enter_vma_merge(vma, vma->vm_flags); - validate_mm(mm); return error; } =20 @@ -2664,16 +2380,14 @@ detach_vmas_to_be_unmapped(struct mm_struct *mm, st= ruct vm_area_struct *vma, vma->vm_prev =3D NULL; vma_mt_szero(mm, vma->vm_start, end); do { - vma_rb_erase(vma, &mm->mm_rb); mm->map_count--; tail_vma =3D vma; vma =3D vma->vm_next; } while (vma && vma->vm_start < end); *insertion_point =3D vma; - if (vma) { + if (vma) vma->vm_prev =3D prev; - vma_gap_update(vma); - } else + else mm->highest_vm_end =3D prev ? vm_end_gap(prev) : 0; tail_vma->vm_next =3D NULL; =20 @@ -2806,11 +2520,7 @@ int __do_munmap(struct mm_struct *mm, unsigned long = start, size_t len, if (len =3D=3D 0) return -EINVAL; =20 - /* - * arch_unmap() might do unmaps itself. It must be called - * and finish any rbtree manipulation before this code - * runs and also starts to manipulate the rbtree. - */ + /* arch_unmap() might do unmaps itself. */ arch_unmap(mm, start, end); =20 /* Find the first overlapping VMA where start < vma->vm_end */ @@ -2818,6 +2528,11 @@ int __do_munmap(struct mm_struct *mm, unsigned long = start, size_t len, if (!vma) return 0; prev =3D vma->vm_prev; + /* we have start < vma->vm_end */ + + /* if it doesn't overlap, we have nothing.. */ + if (vma->vm_start >=3D end) + return 0; =20 /* * If we need to split any vma, do it now to save pain later. @@ -2873,7 +2588,7 @@ int __do_munmap(struct mm_struct *mm, unsigned long s= tart, size_t len, if (mm->locked_vm) unlock_range(vma, end); =20 - /* Detach vmas from rbtree */ + /* Detach vmas from the MM linked list and remove from the mm tree*/ if (!detach_vmas_to_be_unmapped(mm, vma, prev, end)) downgrade =3D false; =20 @@ -3018,11 +2733,11 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, st= art, unsigned long, size, * anonymous maps. eventually we may be able to do some * brk-specific accounting here. */ -static int do_brk_flags(unsigned long addr, unsigned long len, unsigned lo= ng flags, struct list_head *uf) +static int do_brk_flags(unsigned long addr, unsigned long len, + unsigned long flags, struct list_head *uf) { struct mm_struct *mm =3D current->mm; struct vm_area_struct *vma, *prev; - struct rb_node **rb_link, *rb_parent; pgoff_t pgoff =3D addr >> PAGE_SHIFT; int error; unsigned long mapped_addr; @@ -3041,8 +2756,8 @@ static int do_brk_flags(unsigned long addr, unsigned = long len, unsigned long fla if (error) return error; =20 - /* Clear old maps, set up prev, rb_link, rb_parent, and uf */ - if (munmap_vma_range(mm, addr, len, &prev, &rb_link, &rb_parent, uf)) + /* Clear old maps, set up prev and uf */ + if (munmap_vma_range(mm, addr, len, &prev, uf)) return -ENOMEM; =20 /* Check against address space limits *after* clearing old maps... */ @@ -3076,7 +2791,7 @@ static int do_brk_flags(unsigned long addr, unsigned = long len, unsigned long fla vma->vm_pgoff =3D pgoff; vma->vm_flags =3D flags; vma->vm_page_prot =3D vm_get_page_prot(flags); - vma_link(mm, vma, prev, rb_link, rb_parent); + vma_link(mm, vma, prev); out: perf_event_mmap(vma); mm->total_vm +=3D len >> PAGE_SHIFT; @@ -3196,26 +2911,10 @@ void exit_mmap(struct mm_struct *mm) int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma) { struct vm_area_struct *prev; - struct rb_node **rb_link, *rb_parent; - unsigned long start =3D vma->vm_start; - struct vm_area_struct *overlap =3D NULL; =20 - if (find_vma_links(mm, vma->vm_start, vma->vm_end, - &prev, &rb_link, &rb_parent)) + if (range_has_overlap(mm, vma->vm_start, vma->vm_end, &prev)) return -ENOMEM; =20 - overlap =3D mt_find(&mm->mm_mt, &start, vma->vm_end - 1); - if (overlap) { - - pr_err("Found vma ending at %lu\n", start - 1); - pr_err("vma : %lu =3D> %lu-%lu\n", (unsigned long)overlap, - overlap->vm_start, overlap->vm_end - 1); -#if defined(CONFIG_DEBUG_MAPLE_TREE) - mt_dump(&mm->mm_mt); -#endif - BUG(); - } - if ((vma->vm_flags & VM_ACCOUNT) && security_vm_enough_memory_mm(mm, vma_pages(vma))) return -ENOMEM; @@ -3237,7 +2936,7 @@ int insert_vm_struct(struct mm_struct *mm, struct vm_= area_struct *vma) vma->vm_pgoff =3D vma->vm_start >> PAGE_SHIFT; } =20 - vma_link(mm, vma, prev, rb_link, rb_parent); + vma_link(mm, vma, prev); return 0; } =20 @@ -3253,9 +2952,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct= **vmap, unsigned long vma_start =3D vma->vm_start; struct mm_struct *mm =3D vma->vm_mm; struct vm_area_struct *new_vma, *prev; - struct rb_node **rb_link, *rb_parent; bool faulted_in_anon_vma =3D true; - unsigned long index =3D addr; =20 validate_mm_mt(mm); /* @@ -3267,10 +2964,9 @@ struct vm_area_struct *copy_vma(struct vm_area_struc= t **vmap, faulted_in_anon_vma =3D false; } =20 - if (find_vma_links(mm, addr, addr + len, &prev, &rb_link, &rb_parent)) + if (range_has_overlap(mm, addr, addr + len, &prev)) return NULL; /* should never get here */ - if (mt_find(&mm->mm_mt, &index, addr+len - 1)) - BUG(); + new_vma =3D vma_merge(mm, prev, addr, addr + len, vma->vm_flags, vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma), vma->vm_userfaultfd_ctx); @@ -3311,7 +3007,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct= **vmap, get_file(new_vma->vm_file); if (new_vma->vm_ops && new_vma->vm_ops->open) new_vma->vm_ops->open(new_vma); - vma_link(mm, new_vma, prev, rb_link, rb_parent); + vma_link(mm, new_vma, prev); *need_rmap_locks =3D false; } validate_mm_mt(mm); diff --git a/mm/nommu.c b/mm/nommu.c index 02d2427b8f9e..86b162011f76 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -552,9 +552,9 @@ static void put_nommu_region(struct vm_region *region) */ static void add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma= ) { - struct vm_area_struct *pvma, *prev; struct address_space *mapping; - struct rb_node **p, *parent, *rb_prev; + struct vm_area_struct *prev; + MA_STATE(mas, &mm->mm_mt, vma->vm_start, vma->vm_end); =20 BUG_ON(!vma->vm_region); =20 @@ -572,42 +572,12 @@ static void add_vma_to_mm(struct mm_struct *mm, struc= t vm_area_struct *vma) i_mmap_unlock_write(mapping); } =20 + rcu_read_lock(); + prev =3D mas_prev(&mas, 0); + rcu_read_unlock(); + mas_reset(&mas); /* add the VMA to the tree */ - parent =3D rb_prev =3D NULL; - p =3D &mm->mm_rb.rb_node; - while (*p) { - parent =3D *p; - pvma =3D rb_entry(parent, struct vm_area_struct, vm_rb); - - /* sort by: start addr, end addr, VMA struct addr in that order - * (the latter is necessary as we may get identical VMAs) */ - if (vma->vm_start < pvma->vm_start) - p =3D &(*p)->rb_left; - else if (vma->vm_start > pvma->vm_start) { - rb_prev =3D parent; - p =3D &(*p)->rb_right; - } else if (vma->vm_end < pvma->vm_end) - p =3D &(*p)->rb_left; - else if (vma->vm_end > pvma->vm_end) { - rb_prev =3D parent; - p =3D &(*p)->rb_right; - } else if (vma < pvma) - p =3D &(*p)->rb_left; - else if (vma > pvma) { - rb_prev =3D parent; - p =3D &(*p)->rb_right; - } else - BUG(); - } - - rb_link_node(&vma->vm_rb, parent, p); - rb_insert_color(&vma->vm_rb, &mm->mm_rb); - - /* add VMA to the VMA list also */ - prev =3D NULL; - if (rb_prev) - prev =3D rb_entry(rb_prev, struct vm_area_struct, vm_rb); - + vma_mas_store(vma, &mas); __vma_link_list(mm, vma, prev); } =20 @@ -620,6 +590,7 @@ static void delete_vma_from_mm(struct vm_area_struct *v= ma) struct address_space *mapping; struct mm_struct *mm =3D vma->vm_mm; struct task_struct *curr =3D current; + MA_STATE(mas, &vma->vm_mm->mm_mt, 0, 0); =20 mm->map_count--; for (i =3D 0; i < VMACACHE_SIZE; i++) { @@ -629,7 +600,6 @@ static void delete_vma_from_mm(struct vm_area_struct *v= ma) break; } } - /* remove the VMA from the mapping */ if (vma->vm_file) { mapping =3D vma->vm_file->f_mapping; @@ -642,8 +612,7 @@ static void delete_vma_from_mm(struct vm_area_struct *v= ma) } =20 /* remove from the MM's tree and list */ - rb_erase(&vma->vm_rb, &mm->mm_rb); - + vma_mas_remove(vma, &mas); __vma_unlink_list(mm, vma); } =20 @@ -667,24 +636,21 @@ static void delete_vma(struct mm_struct *mm, struct v= m_area_struct *vma) struct vm_area_struct *find_vma(struct mm_struct *mm, unsigned long addr) { struct vm_area_struct *vma; + MA_STATE(mas, &mm->mm_mt, addr, addr); =20 /* check the cache first */ vma =3D vmacache_find(mm, addr); if (likely(vma)) return vma; =20 - /* trawl the list (there may be multiple mappings in which addr - * resides) */ - for (vma =3D mm->mmap; vma; vma =3D vma->vm_next) { - if (vma->vm_start > addr) - return NULL; - if (vma->vm_end > addr) { - vmacache_update(addr, vma); - return vma; - } - } + rcu_read_lock(); + vma =3D mas_walk(&mas); + rcu_read_unlock(); =20 - return NULL; + if (vma) + vmacache_update(addr, vma); + + return vma; } EXPORT_SYMBOL(find_vma); =20 @@ -716,26 +682,25 @@ static struct vm_area_struct *find_vma_exact(struct m= m_struct *mm, { struct vm_area_struct *vma; unsigned long end =3D addr + len; + MA_STATE(mas, &mm->mm_mt, addr, addr); =20 /* check the cache first */ vma =3D vmacache_find_exact(mm, addr, end); if (vma) return vma; =20 - /* trawl the list (there may be multiple mappings in which addr - * resides) */ - for (vma =3D mm->mmap; vma; vma =3D vma->vm_next) { - if (vma->vm_start < addr) - continue; - if (vma->vm_start > addr) - return NULL; - if (vma->vm_end =3D=3D end) { - vmacache_update(addr, vma); - return vma; - } - } - - return NULL; + rcu_read_lock(); + vma =3D mas_walk(&mas); + rcu_read_unlock(); + if (!vma) + return NULL; + if (vma->vm_start !=3D addr) + return NULL; + if (vma->vm_end !=3D end) + return NULL; + + vmacache_update(addr, vma); + return vma; } =20 /* diff --git a/mm/util.c b/mm/util.c index bacabe446906..7179a25f7f2b 100644 --- a/mm/util.c +++ b/mm/util.c @@ -287,6 +287,8 @@ void __vma_link_list(struct mm_struct *mm, struct vm_ar= ea_struct *vma, vma->vm_next =3D next; if (next) next->vm_prev =3D vma; + else + mm->highest_vm_end =3D vm_end_gap(vma); } =20 void __vma_unlink_list(struct mm_struct *mm, struct vm_area_struct *vma) @@ -301,6 +303,12 @@ void __vma_unlink_list(struct mm_struct *mm, struct vm= _area_struct *vma) mm->mmap =3D next; if (next) next->vm_prev =3D prev; + else { + if (prev) + mm->highest_vm_end =3D vm_end_gap(prev); + else + mm->highest_vm_end =3D 0; + } } =20 /* Check if the vma is being used as a stack by this task */ --=20 2.30.2