From patchwork Mon Mar 21 17:18:18 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Sebastian Andrzej Siewior X-Patchwork-Id: 12787597 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 kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by smtp.lore.kernel.org (Postfix) with ESMTP id 2AB32C433EF for ; Mon, 21 Mar 2022 17:18:24 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id B095A8D0001; Mon, 21 Mar 2022 13:18:23 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id AB89C6B0074; Mon, 21 Mar 2022 13:18:23 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 980208D0001; Mon, 21 Mar 2022 13:18:23 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (relay.hostedemail.com [64.99.140.25]) by kanga.kvack.org (Postfix) with ESMTP id 8ABB56B0073 for ; Mon, 21 Mar 2022 13:18:23 -0400 (EDT) Received: from smtpin01.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay12.hostedemail.com (Postfix) with ESMTP id 4086F1213A3 for ; Mon, 21 Mar 2022 17:18:23 +0000 (UTC) X-FDA: 79269052086.01.AF3D57E Received: from galois.linutronix.de (Galois.linutronix.de [193.142.43.55]) by imf18.hostedemail.com (Postfix) with ESMTP id 5245B1C0028 for ; Mon, 21 Mar 2022 17:18:22 +0000 (UTC) Date: Mon, 21 Mar 2022 18:18:18 +0100 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linutronix.de; s=2020; t=1647883099; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type; bh=J6UAQZOvm1BYR4CPqZLJq59eHAVyZtGNYU9xerQOHnc=; b=T6ykB7ohnkIYnLbDjPiVd0FtQl1eEiOCQ+iqNt+kDPhxSQo8ueJ4mRfHGiZ7WsR5mZKHra VzxEH/cLUrkhpCvSDOkssRsOJGa66O00Ck9E3zQeVrFtL7GSz69k4gjdmQ1sQacexfO59n EdVresoVsmS/WFbr3d6uNnmYvy5YXo84n6EVkwZapKPBxS7wVdHoNrAw9FrumcN4yjJe0d PsRsZtqOND5SNkiFtjgVPvxdE6P+fOVdTFdqe1Q/xSegGKkCYl7Eyo9KCg8ZqIh1H+s+j3 vAMF68OBypqBxSycwQB6BfFVYbj+PHm5lb/1tGaX17MFxkbExrjw3jxkebZOgQ== DKIM-Signature: v=1; a=ed25519-sha256; c=relaxed/relaxed; d=linutronix.de; s=2020e; t=1647883099; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type; bh=J6UAQZOvm1BYR4CPqZLJq59eHAVyZtGNYU9xerQOHnc=; b=ILOccq/YE3BqszHRGT4+Cocotsc9S5QfdPSTQq1nWiGZA0ZLmPy2pWE/UE3pM191+OhGrg Ji+lfm5aLUq5A+CQ== From: Sebastian Andrzej Siewior To: linux-mm@kvack.org Cc: Hugh Dickins , Vlastimil Babka , Matthew Wilcox , Andrew Morton , Thomas Gleixner Subject: [PATCH] mm/munlock: Protect the per-CPU pagevec by a local_lock_t. Message-ID: MIME-Version: 1.0 Content-Disposition: inline X-Rspam-User: X-Rspamd-Queue-Id: 5245B1C0028 X-Stat-Signature: 8nxkfti1a95a4bxpjiuk8ywt7tpt9xbs Authentication-Results: imf18.hostedemail.com; dkim=pass header.d=linutronix.de header.s=2020 header.b=T6ykB7oh; dkim=pass header.d=linutronix.de header.s=2020e header.b="ILOccq/Y"; dmarc=pass (policy=none) header.from=linutronix.de; spf=pass (imf18.hostedemail.com: domain of bigeasy@linutronix.de designates 193.142.43.55 as permitted sender) smtp.mailfrom=bigeasy@linutronix.de X-Rspamd-Server: rspam03 X-HE-Tag: 1647883102-462696 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: The access to mlock_pvec is protected by disabling preemption via get_cpu_var() or implicit by having preemption disabled by the caller (in mlock_page_drain() case). This breaks on PREEMPT_RT since folio_lruvec_lock_irq() acquires a sleeping lock in this section. Create struct mlock_pvec which consits of the local_lock_t and the pagevec. Acquire the local_lock() before accessing the per-CPU pagevec. Replace mlock_page_drain() with a _local() version which is invoked on the local CPU and acquires the local_lock_t and a _remote() version which uses the pagevec from a remote CPU which offline. Signed-off-by: Sebastian Andrzej Siewior Acked-by: Hugh Dickins --- mm/internal.h | 6 ++++-- mm/migrate.c | 2 +- mm/mlock.c | 46 ++++++++++++++++++++++++++++++++++++---------- mm/page_alloc.c | 1 + mm/rmap.c | 4 ++-- mm/swap.c | 4 +++- 6 files changed, 47 insertions(+), 16 deletions(-) diff --git a/mm/internal.h b/mm/internal.h index 9be0227ccc942..50c3fd71d7ddd 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -456,7 +456,8 @@ static inline void munlock_vma_page(struct page *page, } void mlock_new_page(struct page *page); bool need_mlock_page_drain(int cpu); -void mlock_page_drain(int cpu); +void mlock_page_drain_local(void); +void mlock_page_drain_remote(int cpu); extern pmd_t maybe_pmd_mkwrite(pmd_t pmd, struct vm_area_struct *vma); @@ -547,7 +548,8 @@ static inline void munlock_vma_page(struct page *page, struct vm_area_struct *vma, bool compound) { } static inline void mlock_new_page(struct page *page) { } static inline bool need_mlock_page_drain(int cpu) { return false; } -static inline void mlock_page_drain(int cpu) { } +static inline void mlock_page_drain_local(void) { } +static inline void mlock_page_drain_remote(int cpu) { } static inline void vunmap_range_noflush(unsigned long start, unsigned long end) { } diff --git a/mm/migrate.c b/mm/migrate.c index 3d60823afd2d3..de175e2fdba5d 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -246,7 +246,7 @@ static bool remove_migration_pte(struct folio *folio, set_pte_at(vma->vm_mm, pvmw.address, pvmw.pte, pte); } if (vma->vm_flags & VM_LOCKED) - mlock_page_drain(smp_processor_id()); + mlock_page_drain_local(); trace_remove_migration_pte(pvmw.address, pte_val(pte), compound_order(new)); diff --git a/mm/mlock.c b/mm/mlock.c index efd2dd2943dec..6e503b8690f52 100644 --- a/mm/mlock.c +++ b/mm/mlock.c @@ -28,7 +28,14 @@ #include "internal.h" -static DEFINE_PER_CPU(struct pagevec, mlock_pvec); +struct mlock_pvec { + local_lock_t lock; + struct pagevec vec; +}; + +static DEFINE_PER_CPU(struct mlock_pvec, mlock_pvec) = { + .lock = INIT_LOCAL_LOCK(lock), +}; bool can_do_mlock(void) { @@ -203,18 +210,30 @@ static void mlock_pagevec(struct pagevec *pvec) pagevec_reinit(pvec); } -void mlock_page_drain(int cpu) +void mlock_page_drain_local(void) { struct pagevec *pvec; - pvec = &per_cpu(mlock_pvec, cpu); + local_lock(&mlock_pvec.lock); + pvec = this_cpu_ptr(&mlock_pvec.vec); + if (pagevec_count(pvec)) + mlock_pagevec(pvec); + local_unlock(&mlock_pvec.lock); +} + +void mlock_page_drain_remote(int cpu) +{ + struct pagevec *pvec; + + WARN_ON_ONCE(cpu_online(cpu)); + pvec = &per_cpu(mlock_pvec.vec, cpu); if (pagevec_count(pvec)) mlock_pagevec(pvec); } bool need_mlock_page_drain(int cpu) { - return pagevec_count(&per_cpu(mlock_pvec, cpu)); + return pagevec_count(&per_cpu(mlock_pvec.vec, cpu)); } /** @@ -223,7 +242,10 @@ bool need_mlock_page_drain(int cpu) */ void mlock_folio(struct folio *folio) { - struct pagevec *pvec = &get_cpu_var(mlock_pvec); + struct pagevec *pvec; + + local_lock(&mlock_pvec.lock); + pvec = this_cpu_ptr(&mlock_pvec.vec); if (!folio_test_set_mlocked(folio)) { int nr_pages = folio_nr_pages(folio); @@ -236,7 +258,7 @@ void mlock_folio(struct folio *folio) if (!pagevec_add(pvec, mlock_lru(&folio->page)) || folio_test_large(folio) || lru_cache_disabled()) mlock_pagevec(pvec); - put_cpu_var(mlock_pvec); + local_unlock(&mlock_pvec.lock); } /** @@ -245,9 +267,11 @@ void mlock_folio(struct folio *folio) */ void mlock_new_page(struct page *page) { - struct pagevec *pvec = &get_cpu_var(mlock_pvec); + struct pagevec *pvec; int nr_pages = thp_nr_pages(page); + local_lock(&mlock_pvec.lock); + pvec = this_cpu_ptr(&mlock_pvec.vec); SetPageMlocked(page); mod_zone_page_state(page_zone(page), NR_MLOCK, nr_pages); __count_vm_events(UNEVICTABLE_PGMLOCKED, nr_pages); @@ -256,7 +280,7 @@ void mlock_new_page(struct page *page) if (!pagevec_add(pvec, mlock_new(page)) || PageHead(page) || lru_cache_disabled()) mlock_pagevec(pvec); - put_cpu_var(mlock_pvec); + local_unlock(&mlock_pvec.lock); } /** @@ -265,8 +289,10 @@ void mlock_new_page(struct page *page) */ void munlock_page(struct page *page) { - struct pagevec *pvec = &get_cpu_var(mlock_pvec); + struct pagevec *pvec; + local_lock(&mlock_pvec.lock); + pvec = this_cpu_ptr(&mlock_pvec.vec); /* * TestClearPageMlocked(page) must be left to __munlock_page(), * which will check whether the page is multiply mlocked. @@ -276,7 +302,7 @@ void munlock_page(struct page *page) if (!pagevec_add(pvec, page) || PageHead(page) || lru_cache_disabled()) mlock_pagevec(pvec); - put_cpu_var(mlock_pvec); + local_unlock(&mlock_pvec.lock); } static int mlock_pte_range(pmd_t *pmd, unsigned long addr, diff --git a/mm/page_alloc.c b/mm/page_alloc.c index f648decfe39d1..94f515845d53a 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -8355,6 +8355,7 @@ static int page_alloc_cpu_dead(unsigned int cpu) struct zone *zone; lru_add_drain_cpu(cpu); + mlock_page_drain_remote(cpu); drain_pages(cpu); /* diff --git a/mm/rmap.c b/mm/rmap.c index 3eb95fcde7000..90b4956096ef3 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -1727,7 +1727,7 @@ static bool try_to_unmap_one(struct folio *folio, struct vm_area_struct *vma, */ page_remove_rmap(subpage, vma, folio_test_hugetlb(folio)); if (vma->vm_flags & VM_LOCKED) - mlock_page_drain(smp_processor_id()); + mlock_page_drain_local(); folio_put(folio); } @@ -2005,7 +2005,7 @@ static bool try_to_migrate_one(struct folio *folio, struct vm_area_struct *vma, */ page_remove_rmap(subpage, vma, folio_test_hugetlb(folio)); if (vma->vm_flags & VM_LOCKED) - mlock_page_drain(smp_processor_id()); + mlock_page_drain_local(); folio_put(folio); } diff --git a/mm/swap.c b/mm/swap.c index bceff0cb559c9..7e320ec08c6ae 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -624,7 +624,6 @@ void lru_add_drain_cpu(int cpu) pagevec_lru_move_fn(pvec, lru_lazyfree_fn); activate_page_drain(cpu); - mlock_page_drain(cpu); } /** @@ -706,6 +705,7 @@ void lru_add_drain(void) local_lock(&lru_pvecs.lock); lru_add_drain_cpu(smp_processor_id()); local_unlock(&lru_pvecs.lock); + mlock_page_drain_local(); } /* @@ -720,6 +720,7 @@ static void lru_add_and_bh_lrus_drain(void) lru_add_drain_cpu(smp_processor_id()); local_unlock(&lru_pvecs.lock); invalidate_bh_lrus_cpu(); + mlock_page_drain_local(); } void lru_add_drain_cpu_zone(struct zone *zone) @@ -728,6 +729,7 @@ void lru_add_drain_cpu_zone(struct zone *zone) lru_add_drain_cpu(smp_processor_id()); drain_local_pages(zone); local_unlock(&lru_pvecs.lock); + mlock_page_drain_local(); } #ifdef CONFIG_SMP