From patchwork Tue Nov 13 18:20:00 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Timofey Titovets X-Patchwork-Id: 10681207 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 17CD2109C for ; Tue, 13 Nov 2018 18:20:15 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 0DB782AFDE for ; Tue, 13 Nov 2018 18:20:13 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 01D062B1F0; Tue, 13 Nov 2018 18:20:12 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-3.0 required=2.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,MAILING_LIST_MULTI,RCVD_IN_DNSWL_NONE autolearn=ham version=3.3.1 Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id AA48A2AFDE for ; Tue, 13 Nov 2018 18:20:11 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id BD3C66B000D; Tue, 13 Nov 2018 13:20:10 -0500 (EST) Delivered-To: linux-mm-outgoing@kvack.org Received: by kanga.kvack.org (Postfix, from userid 40) id B842A6B000E; Tue, 13 Nov 2018 13:20:10 -0500 (EST) X-Original-To: int-list-linux-mm@kvack.org X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id A4AB16B0010; Tue, 13 Nov 2018 13:20:10 -0500 (EST) X-Original-To: linux-mm@kvack.org X-Delivered-To: linux-mm@kvack.org Received: from mail-ed1-f69.google.com (mail-ed1-f69.google.com [209.85.208.69]) by kanga.kvack.org (Postfix) with ESMTP id 454BD6B000D for ; Tue, 13 Nov 2018 13:20:10 -0500 (EST) Received: by mail-ed1-f69.google.com with SMTP id c12-v6so6905013ede.6 for ; Tue, 13 Nov 2018 10:20:10 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:dkim-signature:from:to:cc:subject:date :message-id:mime-version:content-transfer-encoding; bh=r/vg+ZMPkcTiL6a4x5hGstC/N4Asetq2HyHOv6vvEx4=; b=mRw+pg3lQmBl8qMZdgDlgAvWxzmYVglWkfA2Ja2yjcd+1nDkf2lheIi+x5Hcr4rYhT gphOTeED2wwMXWqd0lxQhGSmImx3LRvSNf/WYyj8wP0Z5Ga8YMvWTc8MLzImDOznOsOt TQUikknl8mtMBYDbpL4RHOs0hDehh6D36Z+7rGG3HX0V7SeTUxWbFZ9/p+QiZ1I5wHJw x+eZxnqm4SaAysTxnNv8RF+nq6WlCXr2RarOSiKWxJLhSu0OfW7XuAu4Y8fsk5s1RQu0 9/NFghXeYEM8ofDxUPHTCcIf8UBgHmxb/T9S7zLeESwj5AQWThi6HcAxMEJpBoGNuHOM iBAQ== X-Gm-Message-State: AGRZ1gLfy8UYne7l/qd09z5bWOdJ3N7PBtgAFroSfblPJtkfp2ltdyjg DaJUQgFXtX8tpYT/eV1vYAQUneIpW8l8ZBsmA4y/9mhCqAUIHW1W5kfbhjJjbi9AotqMxvlbgyI YdGp3wz/grlXtZZ2SC8ZiQribB1+vGYC73+14H5OhN5wsN1oP7595S+1ccX9kgmcZpd4qifj6/p h72ZtByR7IBYxupIY3Jbp/ncCt6WyS21A9QQJJ/38Ra8EUXK72iicQO6sP65hYjli8pjhhXQ0D8 13tJhGzqZoRJ4QUr1vbl2eaqXoJpmg2XQQQpSfqF6hRmvlvBEUbTbwvcX2gype1psf+53cAq1nt TP2i2tjZzFOahL+6FHaOkaRq/xW1G9iv4bS6tDTAh+y+vl6vw0gB1UMlIjqYGdXkSAwl7F0jOzm 4 X-Received: by 2002:a17:906:5a84:: with SMTP id l4-v6mr13402344ejq.229.1542133209692; Tue, 13 Nov 2018 10:20:09 -0800 (PST) X-Received: by 2002:a17:906:5a84:: with SMTP id l4-v6mr13402303ejq.229.1542133208426; Tue, 13 Nov 2018 10:20:08 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1542133208; cv=none; d=google.com; s=arc-20160816; b=CseG4gbgrF9/6p8jNaWWb1NK90uxyLwUKY3uF//pkIedy36S3np8+Ea0q7m512PN1I QpJVgQ/UMtj5X0TKr3ezRdAYQfqzizTYmOgti7O2mOYik8Wg9PXBSyyza09ri3yMx9ea p/XKIQSfbyBcdXf807GVnNBN6BFEKWxZdTUskaaU+c7hz3Pjj7yhNVXPliVbVEBrXHVd 2ch6BO38yaQrRYegVqa7nX+jhy5fcbRUqTwhMLeUFhyxgODBXIFM0zhPgejsFkL04tIb nkT+eOHfA5D6glIsxzeOpODDCW8zpj5F/XB7/hpZbEfRP/sx2Z4HPfUJ3e9RUnl8GXFX swyQ== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=content-transfer-encoding:mime-version:message-id:date:subject:cc :to:from:dkim-signature; bh=r/vg+ZMPkcTiL6a4x5hGstC/N4Asetq2HyHOv6vvEx4=; b=K9Ue7QtuH5Q8tFC3PQFdAzVoEEszqNtEdpZp/FFHWCe00Hdk+TlRI1q6g2Tla+FI34 bXXz9HDPkugoBp8QD5XmLfteEZ7H3Hdfc+QFo+T81XjD0ExHCnBwwfLZ97owOPwidaiq lUgcQFQk75TDnFQrIAOvCN0tsDwn9jz+5obArB3L5/AOsmCMTXbz5kD+aCF+GXSK/6g3 jY7pKguSpuy7nm0DCqWNSJcbQ+hHzmNHqSXEWnu4ycRZoaCQDAsZ3lWJ/xM8MZr+Uyga IpXH99jVzLseEu4g9WnA34O44nxPUTydLP5qGObTR1Gnc3nifGWE3acs55hPW/0Zwd1R r9fA== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@gmail.com header.s=20161025 header.b=UdvklMMS; spf=pass (google.com: domain of nefelim4ag@gmail.com designates 209.85.220.65 as permitted sender) smtp.mailfrom=nefelim4ag@gmail.com; dmarc=pass (p=NONE sp=QUARANTINE dis=NONE) header.from=gmail.com Received: from mail-sor-f65.google.com (mail-sor-f65.google.com. [209.85.220.65]) by mx.google.com with SMTPS id n25sor5417618edd.1.2018.11.13.10.20.08 for (Google Transport Security); Tue, 13 Nov 2018 10:20:08 -0800 (PST) Received-SPF: pass (google.com: domain of nefelim4ag@gmail.com designates 209.85.220.65 as permitted sender) client-ip=209.85.220.65; Authentication-Results: mx.google.com; dkim=pass header.i=@gmail.com header.s=20161025 header.b=UdvklMMS; spf=pass (google.com: domain of nefelim4ag@gmail.com designates 209.85.220.65 as permitted sender) smtp.mailfrom=nefelim4ag@gmail.com; dmarc=pass (p=NONE sp=QUARANTINE dis=NONE) header.from=gmail.com DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:mime-version :content-transfer-encoding; bh=r/vg+ZMPkcTiL6a4x5hGstC/N4Asetq2HyHOv6vvEx4=; b=UdvklMMSsA5AAhvJRKzZRIDjyEFAab1y5n6vsiYZOjstLEzLrJNCpZ2+7/SCnPtaaY P6Vy6oRC/f52nGEFCt3KTPxr2IOtu6YnMoUiASjIrJPA/oHNx7gMyiXrA9JCqvuhQClR 44nSUamWEejjXPviMYHq1k6utNyburC5eu8xLlH6OcO7ORZrtXQqBHl0mm3pyJTQyoYZ GZRJqpLKlr7CBJ9Wl3B95XHMOXLF4v/rL1zbPow9L+QwYUSD13Pu/nWZTBmDMzJq6dKM 9QHQEPyuHQuXR6gf0aaqBHuJes9L6SM52FGadAnRiBGjfXn9xD2wlQggN6bsiqZ3a1ky +W3g== X-Google-Smtp-Source: AJdET5dEPz+i0qi8ttDzY+PsDWS+uR1qnBsFFt2Inkl2D8uQTYkAyQvF1QhDVHr5ZzgOspmn+sec7A== X-Received: by 2002:aa7:dd16:: with SMTP id i22mr1558822edv.182.1542133207759; Tue, 13 Nov 2018 10:20:07 -0800 (PST) Received: from TitovetsT.synesis.local ([86.57.155.118]) by smtp.gmail.com with ESMTPSA id b37-v6sm5475633eda.92.2018.11.13.10.20.06 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Tue, 13 Nov 2018 10:20:07 -0800 (PST) From: Timofey Titovets To: linux-kernel@vger.kernel.org Cc: Timofey Titovets , Matthew Wilcox , linux-mm@kvack.org, linux-doc@vger.kernel.org Subject: [PATCH V3] KSM: allow dedup all tasks memory Date: Tue, 13 Nov 2018 21:20:00 +0300 Message-Id: <20181113182000.20055-1-nefelim4ag@gmail.com> X-Mailer: git-send-email 2.19.1 MIME-Version: 1.0 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: X-Virus-Scanned: ClamAV using ClamSMTP ksm by default working only on memory that added by madvise(). And only way get that work on other applications: * Use LD_PRELOAD and libraries * Patch kernel Lets use kernel task list and add logic to import VMAs from tasks. That behaviour controlled by new attributes: * mode: I try mimic hugepages attribute, so mode have two states: * madvise - old default behaviour * always [new] - allow ksm to get tasks vma and try working on that. * seeker_sleep_millisecs: Add pauses between imports tasks VMA For rate limiting proporses and tasklist locking time, ksm seeker thread only import VMAs from one task per loop. Some numbers from different not madvised workloads. Formulas: Percentage ratio = (pages_sharing - pages_shared)/pages_unshared Memory saved = (pages_sharing - pages_shared)*4/1024 MiB Memory used = free -h * Name: My working laptop Description: Many different chrome/electron apps + KDE Ratio: 5% Saved: ~100 MiB Used: ~2000 MiB * Name: K8s test VM Description: Some small random running docker images Ratio: 40% Saved: ~160 MiB Used: ~920 MiB * Name: Ceph test VM Description: Ceph Mon/OSD, some containers Ratio: 20% Saved: ~60 MiB Used: ~600 MiB * Name: BareMetal K8s backend server Description: Different server apps in containers C, Java, GO & etc Ratio: 72% Saved: ~5800 MiB Used: ~35.7 GiB * Name: BareMetal K8s processing server Description: Many instance of one CPU intensive application Ratio: 55% Saved: ~2600 MiB Used: ~28.0 GiB * Name: BareMetal Ceph node Description: Only OSD storage daemons running Raio: 2% Saved: ~190 MiB Used: ~11.7 GiB Changes: v1 -> v2: * Rebase on v4.19.1 (must also apply on 4.20-rc2+) v2 -> v3: * Reformat patch description * Rename mode normal to madvise * Add some memory numbers * Fix checkpatch.pl warnings * Separate ksm vma seeker to another kthread * Fix: "BUG: scheduling while atomic: ksmd" by move seeker to another thread Signed-off-by: Timofey Titovets CC: Matthew Wilcox CC: linux-mm@kvack.org CC: linux-doc@vger.kernel.org --- Documentation/admin-guide/mm/ksm.rst | 15 ++ mm/ksm.c | 215 +++++++++++++++++++++++---- 2 files changed, 198 insertions(+), 32 deletions(-) diff --git a/Documentation/admin-guide/mm/ksm.rst b/Documentation/admin-guide/mm/ksm.rst index 9303786632d1..7cffd47f9b38 100644 --- a/Documentation/admin-guide/mm/ksm.rst +++ b/Documentation/admin-guide/mm/ksm.rst @@ -116,6 +116,21 @@ run Default: 0 (must be changed to 1 to activate KSM, except if CONFIG_SYSFS is disabled) +mode + * set always to allow ksm deduplicate memory of every process + * set madvise to use only madvised memory + + Default: madvise (dedupulicate only madvised memory as in + earlier releases) + +seeker_sleep_millisecs + how many milliseconds ksmd task seeker should sleep try another + task. + e.g. ``echo 1000 > /sys/kernel/mm/ksm/seeker_sleep_millisecs`` + + Default: 1000 (chosen for rate limit purposes) + + use_zero_pages specifies whether empty pages (i.e. allocated pages that only contain zeroes) should be treated specially. When set to 1, diff --git a/mm/ksm.c b/mm/ksm.c index 5b0894b45ee5..1a03b28b6288 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -273,6 +273,9 @@ static unsigned int ksm_thread_pages_to_scan = 100; /* Milliseconds ksmd should sleep between batches */ static unsigned int ksm_thread_sleep_millisecs = 20; +/* Milliseconds ksmd seeker should sleep between runs */ +static unsigned int ksm_thread_seeker_sleep_millisecs = 1000; + /* Checksum of an empty (zeroed) page */ static unsigned int zero_checksum __read_mostly; @@ -295,7 +298,12 @@ static int ksm_nr_node_ids = 1; static unsigned long ksm_run = KSM_RUN_STOP; static void wait_while_offlining(void); +#define KSM_MODE_MADVISE 0 +#define KSM_MODE_ALWAYS 1 +static unsigned long ksm_mode = KSM_MODE_MADVISE; + static DECLARE_WAIT_QUEUE_HEAD(ksm_thread_wait); +static DECLARE_WAIT_QUEUE_HEAD(ksm_seeker_thread_wait); static DEFINE_MUTEX(ksm_thread_mutex); static DEFINE_SPINLOCK(ksm_mmlist_lock); @@ -303,6 +311,11 @@ static DEFINE_SPINLOCK(ksm_mmlist_lock); sizeof(struct __struct), __alignof__(struct __struct),\ (__flags), NULL) +static inline int ksm_mode_always(void) +{ + return (ksm_mode == KSM_MODE_ALWAYS); +} + static int __init ksm_slab_init(void) { rmap_item_cache = KSM_KMEM_CACHE(rmap_item, 0); @@ -2389,6 +2402,106 @@ static int ksmd_should_run(void) return (ksm_run & KSM_RUN_MERGE) && !list_empty(&ksm_mm_head.mm_list); } + +static int ksm_enter(struct mm_struct *mm, unsigned long *vm_flags) +{ + int err; + + if (*vm_flags & (VM_MERGEABLE | VM_SHARED | VM_MAYSHARE | + VM_PFNMAP | VM_IO | VM_DONTEXPAND | + VM_HUGETLB | VM_MIXEDMAP)) + return 0; + +#ifdef VM_SAO + if (*vm_flags & VM_SAO) + return 0; +#endif +#ifdef VM_SPARC_ADI + if (*vm_flags & VM_SPARC_ADI) + return 0; +#endif + if (!test_bit(MMF_VM_MERGEABLE, &mm->flags)) { + err = __ksm_enter(mm); + if (err) + return err; + } + + *vm_flags |= VM_MERGEABLE; + + return 0; +} + +/* + * Register all vmas for all processes in the system with KSM. + * Note that every call to ksm_, for a given vma, after the first + * does nothing but set flags. + */ +void ksm_import_task_vma(struct task_struct *task) +{ + struct vm_area_struct *vma; + struct mm_struct *mm; + int error; + + mm = get_task_mm(task); + if (!mm) + return; + down_write(&mm->mmap_sem); + vma = mm->mmap; + while (vma) { + error = ksm_enter(vma->vm_mm, &vma->vm_flags); + vma = vma->vm_next; + } + up_write(&mm->mmap_sem); + mmput(mm); +} + +static int ksm_seeker_thread(void *nothing) +{ + pid_t last_pid = 1; + pid_t curr_pid; + struct task_struct *task; + + set_freezable(); + set_user_nice(current, 5); + + while (!kthread_should_stop()) { + wait_while_offlining(); + + try_to_freeze(); + + if (!ksm_mode_always()) { + wait_event_freezable(ksm_seeker_thread_wait, + ksm_mode_always() || kthread_should_stop()); + continue; + } + + /* + * import one task's vma per run + */ + read_lock(&tasklist_lock); + + /* Try always get next task */ + for_each_process(task) { + curr_pid = task_pid_nr(task); + if (curr_pid == last_pid) { + task = next_task(task); + break; + } + + if (curr_pid > last_pid) + break; + } + + last_pid = task_pid_nr(task); + ksm_import_task_vma(task); + read_unlock(&tasklist_lock); + + schedule_timeout_interruptible( + msecs_to_jiffies(ksm_thread_seeker_sleep_millisecs)); + } + return 0; +} + static int ksm_scan_thread(void *nothing) { set_freezable(); @@ -2422,33 +2535,9 @@ int ksm_madvise(struct vm_area_struct *vma, unsigned long start, switch (advice) { case MADV_MERGEABLE: - /* - * Be somewhat over-protective for now! - */ - if (*vm_flags & (VM_MERGEABLE | VM_SHARED | VM_MAYSHARE | - VM_PFNMAP | VM_IO | VM_DONTEXPAND | - VM_HUGETLB | VM_MIXEDMAP)) - return 0; /* just ignore the advice */ - - if (vma_is_dax(vma)) - return 0; - -#ifdef VM_SAO - if (*vm_flags & VM_SAO) - return 0; -#endif -#ifdef VM_SPARC_ADI - if (*vm_flags & VM_SPARC_ADI) - return 0; -#endif - - if (!test_bit(MMF_VM_MERGEABLE, &mm->flags)) { - err = __ksm_enter(mm); - if (err) - return err; - } - - *vm_flags |= VM_MERGEABLE; + err = ksm_enter(mm, vm_flags); + if (err) + return err; break; case MADV_UNMERGEABLE: @@ -2829,6 +2918,29 @@ static ssize_t sleep_millisecs_store(struct kobject *kobj, } KSM_ATTR(sleep_millisecs); +static ssize_t seeker_sleep_millisecs_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%u\n", ksm_thread_seeker_sleep_millisecs); +} + +static ssize_t seeker_sleep_millisecs_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + unsigned long msecs; + int err; + + err = kstrtoul(buf, 10, &msecs); + if (err || msecs > UINT_MAX) + return -EINVAL; + + ksm_thread_seeker_sleep_millisecs = msecs; + + return count; +} +KSM_ATTR(seeker_sleep_millisecs); + static ssize_t pages_to_scan_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { @@ -2852,6 +2964,34 @@ static ssize_t pages_to_scan_store(struct kobject *kobj, } KSM_ATTR(pages_to_scan); +static ssize_t mode_show(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + switch (ksm_mode) { + case KSM_MODE_ALWAYS: + return sprintf(buf, "[always] madvise\n"); + case KSM_MODE_MADVISE: + return sprintf(buf, "always [madvise]\n"); + } + + return sprintf(buf, "always [madvise]\n"); +} + +static ssize_t mode_store(struct kobject *kobj, struct kobj_attribute *attr, + const char *buf, size_t count) +{ + if (!memcmp("always", buf, min(sizeof("always")-1, count))) { + ksm_mode = KSM_MODE_ALWAYS; + wake_up_interruptible(&ksm_seeker_thread_wait); + } else if (!memcmp("madvise", buf, min(sizeof("madvise")-1, count))) { + ksm_mode = KSM_MODE_MADVISE; + } else + return -EINVAL; + + return count; +} +KSM_ATTR(mode); + static ssize_t run_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { @@ -3108,7 +3248,9 @@ KSM_ATTR_RO(full_scans); static struct attribute *ksm_attrs[] = { &sleep_millisecs_attr.attr, + &seeker_sleep_millisecs_attr.attr, &pages_to_scan_attr.attr, + &mode_attr.attr, &run_attr.attr, &pages_shared_attr.attr, &pages_sharing_attr.attr, @@ -3134,7 +3276,7 @@ static const struct attribute_group ksm_attr_group = { static int __init ksm_init(void) { - struct task_struct *ksm_thread; + struct task_struct *ksm_thread[2]; int err; /* The correct value depends on page size and endianness */ @@ -3146,10 +3288,18 @@ static int __init ksm_init(void) if (err) goto out; - ksm_thread = kthread_run(ksm_scan_thread, NULL, "ksmd"); - if (IS_ERR(ksm_thread)) { + ksm_thread[0] = kthread_run(ksm_scan_thread, NULL, "ksmd"); + if (IS_ERR(ksm_thread[0])) { pr_err("ksm: creating kthread failed\n"); - err = PTR_ERR(ksm_thread); + err = PTR_ERR(ksm_thread[0]); + goto out_free; + } + + ksm_thread[1] = kthread_run(ksm_seeker_thread, NULL, "ksmd_seeker"); + if (IS_ERR(ksm_thread[1])) { + pr_err("ksm: creating seeker kthread failed\n"); + err = PTR_ERR(ksm_thread[1]); + kthread_stop(ksm_thread[0]); goto out_free; } @@ -3157,7 +3307,8 @@ static int __init ksm_init(void) err = sysfs_create_group(mm_kobj, &ksm_attr_group); if (err) { pr_err("ksm: register sysfs failed\n"); - kthread_stop(ksm_thread); + kthread_stop(ksm_thread[0]); + kthread_stop(ksm_thread[1]); goto out_free; } #else