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=-16.8 required=3.0 tests=BAYES_00,DKIM_INVALID, DKIM_SIGNED,INCLUDES_CR_TRAILER,INCLUDES_PATCH,MAILING_LIST_MULTI, SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT 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 F24CAC433E6 for ; Thu, 14 Jan 2021 18:00:51 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id AD58923B7A for ; Thu, 14 Jan 2021 18:00:51 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org AD58923B7A Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 0BAE48D00F9; Thu, 14 Jan 2021 13:00:51 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 06B808D00F0; Thu, 14 Jan 2021 13:00:51 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id E9AC58D00F9; Thu, 14 Jan 2021 13:00:50 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0098.hostedemail.com [216.40.44.98]) by kanga.kvack.org (Postfix) with ESMTP id D08198D00F0 for ; Thu, 14 Jan 2021 13:00:50 -0500 (EST) Received: from smtpin15.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay03.hostedemail.com (Postfix) with ESMTP id 9317A824556B for ; Thu, 14 Jan 2021 18:00:50 +0000 (UTC) X-FDA: 77705146260.15.mice13_2303e5827528 Received: from filter.hostedemail.com (10.5.16.251.rfc1918.com [10.5.16.251]) by smtpin15.hostedemail.com (Postfix) with ESMTP id 610251809EA1A for ; Thu, 14 Jan 2021 18:00:50 +0000 (UTC) X-HE-Tag: mice13_2303e5827528 X-Filterd-Recvd-Size: 7397 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by imf20.hostedemail.com (Postfix) with ESMTP for ; Thu, 14 Jan 2021 18:00:49 +0000 (UTC) Received: by mail.kernel.org (Postfix) with ESMTPSA id CD9AC23B75; Thu, 14 Jan 2021 18:00:46 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1610647248; bh=IeHJxTh2dLGEMk6hqY4LyIwvV7pcdSMi8pKKcvjOfAM=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=FGU10sysYAVoZc/G/+OQKyDy4BWFiRF2X3KThHeiB8s5OYZ5csvIjMLRBj4ZuZTB6 Oqhp9TmvTmrQCsXbSZpjcrUANJIfYs1Az0oF0w2Q6JTUTnqQaCI/EI3SAakXPoSV70 o4ub6RZ48joAsjCHc2AnoLrnXr43U0BeemlWDvYl1dtxAWbjtPuoCB7ozQs7egCSjK gTKD4ekc4uCFZSOStKzA3ThZn/e2W8mnLClqmmVbSCYtqDbVM8tceCsvP53I0bVc5E FpsNwRFdhHukVWQdL/3+av/w7Xm34ZDbsuEOS8Tu0Q+sUh1n65UA5mJbahfTKDMCVw bbgb5RjQMtEKA== From: Will Deacon To: linux-kernel@vger.kernel.org Cc: linux-mm@kvack.org, linux-arm-kernel@lists.infradead.org, Will Deacon , Catalin Marinas , Jan Kara , Minchan Kim , Andrew Morton , "Kirill A . Shutemov" , Linus Torvalds , Vinayak Menon , Hugh Dickins , kernel-team@android.com Subject: [PATCH v3 2/8] mm: Allow architectures to request 'old' entries when prefaulting Date: Thu, 14 Jan 2021 17:59:28 +0000 Message-Id: <20210114175934.13070-3-will@kernel.org> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20210114175934.13070-1-will@kernel.org> References: <20210114175934.13070-1-will@kernel.org> MIME-Version: 1.0 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: Commit 5c0a85fad949 ("mm: make faultaround produce old ptes") changed the "faultaround" behaviour to initialise prefaulted PTEs as 'old', since this avoids vmscan wrongly assuming that they are hot, despite having never been explicitly accessed by userspace. The change has been shown to benefit numerous arm64 micro-architectures (with hardware access flag) running Android, where both application launch latency and direct reclaim time are significantly reduced (by 10%+ and ~80% respectively). Unfortunately, commit 315d09bf30c2 ("Revert "mm: make faultaround produce old ptes"") reverted the change due to it being identified as the cause of a ~6% regression in unixbench on x86. Experiments on a variety of recent arm64 micro-architectures indicate that unixbench is not affected by the original commit, which appears to yield a 0-1% performance improvement. Since one size does not fit all for the initial state of prefaulted PTEs, introduce arch_wants_old_prefaulted_pte(), which allows an architecture to opt-in to 'old' prefaulted PTEs at runtime based on whatever criteria it may have. Cc: Jan Kara Cc: Minchan Kim Cc: Andrew Morton Cc: Kirill A. Shutemov Cc: Linus Torvalds Reported-by: Vinayak Menon Signed-off-by: Will Deacon --- include/linux/mm.h | 5 ++++- mm/filemap.c | 14 ++++++++++---- mm/memory.c | 20 +++++++++++++++++++- 3 files changed, 33 insertions(+), 6 deletions(-) diff --git a/include/linux/mm.h b/include/linux/mm.h index 4572a9bc5862..251a2339befb 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -434,6 +434,7 @@ extern pgprot_t protection_map[16]; * @FAULT_FLAG_REMOTE: The fault is not for current task/mm. * @FAULT_FLAG_INSTRUCTION: The fault was during an instruction fetch. * @FAULT_FLAG_INTERRUPTIBLE: The fault can be interrupted by non-fatal = signals. + * @FAULT_FLAG_PREFAULT: Fault was a prefault. * * About @FAULT_FLAG_ALLOW_RETRY and @FAULT_FLAG_TRIED: we can specify * whether we would allow page faults to retry by specifying these two @@ -464,6 +465,7 @@ extern pgprot_t protection_map[16]; #define FAULT_FLAG_REMOTE 0x80 #define FAULT_FLAG_INSTRUCTION 0x100 #define FAULT_FLAG_INTERRUPTIBLE 0x200 +#define FAULT_FLAG_PREFAULT 0x400 =20 /* * The default fault flags that should be used by most of the @@ -501,7 +503,8 @@ static inline bool fault_flag_allow_retry_first(unsig= ned int flags) { FAULT_FLAG_USER, "USER" }, \ { FAULT_FLAG_REMOTE, "REMOTE" }, \ { FAULT_FLAG_INSTRUCTION, "INSTRUCTION" }, \ - { FAULT_FLAG_INTERRUPTIBLE, "INTERRUPTIBLE" } + { FAULT_FLAG_INTERRUPTIBLE, "INTERRUPTIBLE" }, \ + { FAULT_FLAG_PREFAULT, "PREFAULT" } =20 /* * vm_fault is filled by the pagefault handler and passed to the vma's diff --git a/mm/filemap.c b/mm/filemap.c index c1f2dc89b8a7..a6dc97906c8e 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -3019,6 +3019,7 @@ vm_fault_t filemap_map_pages(struct vm_fault *vmf, struct address_space *mapping =3D file->f_mapping; pgoff_t last_pgoff =3D start_pgoff; unsigned long address =3D vmf->address; + unsigned long flags =3D vmf->flags; XA_STATE(xas, &mapping->i_pages, start_pgoff); struct page *head, *page; unsigned int mmap_miss =3D READ_ONCE(file->f_ra.mmap_miss); @@ -3051,14 +3052,18 @@ vm_fault_t filemap_map_pages(struct vm_fault *vmf= , if (!pte_none(*vmf->pte)) goto unlock; =20 + /* We're about to handle the fault */ + if (vmf->address =3D=3D address) { + vmf->flags &=3D ~FAULT_FLAG_PREFAULT; + ret =3D VM_FAULT_NOPAGE; + } else { + vmf->flags |=3D FAULT_FLAG_PREFAULT; + } + do_set_pte(vmf, page); /* no need to invalidate: a not-present page won't be cached */ update_mmu_cache(vma, vmf->address, vmf->pte); unlock_page(head); - - /* The fault is handled */ - if (vmf->address =3D=3D address) - ret =3D VM_FAULT_NOPAGE; continue; unlock: unlock_page(head); @@ -3067,6 +3072,7 @@ vm_fault_t filemap_map_pages(struct vm_fault *vmf, pte_unmap_unlock(vmf->pte, vmf->ptl); out: rcu_read_unlock(); + vmf->flags =3D flags; vmf->address =3D address; WRITE_ONCE(file->f_ra.mmap_miss, mmap_miss); return ret; diff --git a/mm/memory.c b/mm/memory.c index 3e2fc2950ad7..f0e7c589ca9d 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -134,6 +134,18 @@ static inline bool arch_faults_on_old_pte(void) } #endif =20 +#ifndef arch_wants_old_prefaulted_pte +static inline bool arch_wants_old_prefaulted_pte(void) +{ + /* + * Transitioning a PTE from 'old' to 'young' can be expensive on + * some architectures, even if it's performed in hardware. By + * default, "false" means prefaulted entries will be 'young'. + */ + return false; +} +#endif + static int __init disable_randmaps(char *s) { randomize_va_space =3D 0; @@ -3725,11 +3737,17 @@ void do_set_pte(struct vm_fault *vmf, struct page= *page) { struct vm_area_struct *vma =3D vmf->vma; bool write =3D vmf->flags & FAULT_FLAG_WRITE; + bool prefault =3D vmf->flags & FAULT_FLAG_PREFAULT; pte_t entry; =20 flush_icache_page(vma, page); entry =3D mk_pte(page, vma->vm_page_prot); - entry =3D pte_sw_mkyoung(entry); + + if (prefault && arch_wants_old_prefaulted_pte()) + entry =3D pte_mkold(entry); + else + entry =3D pte_sw_mkyoung(entry); + if (write) entry =3D maybe_mkwrite(pte_mkdirty(entry), vma); /* copy-on-write page */ --=20 2.30.0.284.gd98b1dd5eaa7-goog