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 kanga.kvack.org (kanga.kvack.org [205.233.56.17]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 03E34CCD1AB for ; Wed, 22 Oct 2025 17:28:38 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id 416AB8E0006; Wed, 22 Oct 2025 13:28:38 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 3C6A88E0003; Wed, 22 Oct 2025 13:28:38 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 28F3F8E0006; Wed, 22 Oct 2025 13:28:38 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0017.hostedemail.com [216.40.44.17]) by kanga.kvack.org (Postfix) with ESMTP id 108558E0003 for ; Wed, 22 Oct 2025 13:28:38 -0400 (EDT) Received: from smtpin23.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay03.hostedemail.com (Postfix) with ESMTP id C98F3BBF0F for ; Wed, 22 Oct 2025 17:28:37 +0000 (UTC) X-FDA: 84026434674.23.97F5EA1 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) by imf20.hostedemail.com (Postfix) with ESMTP id 5815E1C0007 for ; Wed, 22 Oct 2025 17:28:34 +0000 (UTC) Authentication-Results: imf20.hostedemail.com; dkim=pass header.d=redhat.com header.s=mimecast20190719 header.b=HfWzuNtz; spf=pass (imf20.hostedemail.com: domain of david@redhat.com designates 170.10.133.124 as permitted sender) smtp.mailfrom=david@redhat.com; dmarc=pass (policy=quarantine) header.from=redhat.com ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1761154115; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=ersH+A2YnVEG4u4wWpk7ksHDpnwZb29+THsgHlY+RNw=; b=WcThasoM+qjJKq1XEptt1CSVx2bPEunfiZu5V94Hndrej2QNTZC+9hJTt2bHQB89ehngIh jZr1+O3kEPEmQrDQc149HmZhmwwUFQlRCrElIKzr38B4DRg3V5g0fKeAIEGtIo2rsGsINJ sPxrF3xFObNCxzB5f1YIk4s51HdLrgo= ARC-Authentication-Results: i=1; imf20.hostedemail.com; dkim=pass header.d=redhat.com header.s=mimecast20190719 header.b=HfWzuNtz; spf=pass (imf20.hostedemail.com: domain of david@redhat.com designates 170.10.133.124 as permitted sender) smtp.mailfrom=david@redhat.com; dmarc=pass (policy=quarantine) header.from=redhat.com ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1761154115; a=rsa-sha256; cv=none; b=GFFjSXLz+lsOrvTNfrcntc5N+dlWxpEJqjqaVOMSIuPe0jRsmhXqQn8Qc++tfvr4eqKcUh q641yOQk65OqQk9TlRCFOOTfXPt9RQrlTxKt4FU3ItXBGIWIyCAko03KpCfR3pcIqsYz6t ZeqAC/LsMWHu3CnBNW9c5gyTkMhV4ow= DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1761154113; 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: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:autocrypt:autocrypt; bh=ersH+A2YnVEG4u4wWpk7ksHDpnwZb29+THsgHlY+RNw=; b=HfWzuNtzbLalH+oGp2Kq3H7jZoILlqscLS8WUPa8gnnj1K8F02jzU51RXW7dIVa7Jw9wqJ yCdl6Iyz8bT+Bcg1w4Dg7PxqEiCjJcrlTd5OEQOYLA6oSDqLubppe5WalNKnFA/oSDCrXj W/tvVZVSN59jVH4PKV1wbNW64i/sQfs= Received: from mail-wm1-f69.google.com (mail-wm1-f69.google.com [209.85.128.69]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-596-c8jjjWtSNECikfG5tbOPrg-1; Wed, 22 Oct 2025 13:28:31 -0400 X-MC-Unique: c8jjjWtSNECikfG5tbOPrg-1 X-Mimecast-MFC-AGG-ID: c8jjjWtSNECikfG5tbOPrg_1761154110 Received: by mail-wm1-f69.google.com with SMTP id 5b1f17b1804b1-4710c04a403so70962105e9.3 for ; Wed, 22 Oct 2025 10:28:31 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1761154110; x=1761758910; h=content-transfer-encoding:in-reply-to:autocrypt:content-language :from:references:cc:to:subject:user-agent:mime-version:date :message-id:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=ersH+A2YnVEG4u4wWpk7ksHDpnwZb29+THsgHlY+RNw=; b=r6R4lb1DQ7rtBoNOekrBlQxSdCuYEfNPQnXnmC6TL4S334V1LioBVPGD6q1p/+s9Ek PB0lRId9nvWAmeEhHRJaNQ18MYzqRdpRoqte+LluVhkVcNpHxU3kjHoxuguun0TMl1qy EryoWdv1gpia4WJAuHjvDsEwwS4K/OfmvxKd68tyV9c5saCSuImXwOzpitYp01T43iSw MAVzxTpYrrCKJvFJDqWIZsgDtsdBz3/8NnZNl804qZDjCqlgA5gpZ05umGjJxmIAAcjP fbTH4GMo/OSpMrLZ+nl+0ZlISTvQiMYtMIphcnuLoAtDqJ4FwadUqxnIbq7EHhQoAkH1 pHMg== X-Gm-Message-State: AOJu0YwLl0VQO3dx92Pt7of0pb1w8sShR+mRXwcY49k18b9U0JslNINO 4oPRWNZUqYCLIvp9bdTHtlXFPUdt3KypHdLndWbwObkTR+NQ/Cu3HKATtocVMQarUn5FAsV3G41 2/6aULFmJ5GTu7UZ4PLRVf1vsNj7nNGm5skw+9NRAl+nQDMSmQVJs X-Gm-Gg: ASbGnctAJpTiZj5LjIxyGu1Uyj5Hb5w4CxDMH0MOCeSXVPZqxyOM81bvhCqRusOmqQa +2BjSZmIgGCHhoIlmNv+qmq2+EyZg2j/f7Ev6baEqjxbeKpCRF3GFzzxBN5MxuAlsTYH/Z201ez 111Edj78Pz8wng32cT8no35HJ3TCuu6qv4hwyUqKjaeiFwfoRl5VDZ+kSnYJY1yAPfM+jlMgJCY WfnepJWHdMCYu/ts5USNwU8wlPv0N/Q+sp8F7tibzbqKsCZSxX9O3oJDonYOpME/lG23vdDRdPv 7YvkPHo9KZSWGQH1wdJCi4qA+1YF3eOmMkFsibctY94KvG6IKSm6slPVMydWrT/txEfdKcPP7H+ jxpnXQZN4Owng7UUxF1wVjf+ZQAowrU0Rmloy+RKsVtRcL3hgNXDGQHo+F46KhCCzjBzSnoD9ld 6JPMKbk/Hi3YoNKtGgJU+XvkRwvng= X-Received: by 2002:a05:600c:6291:b0:471:a43:123f with SMTP id 5b1f17b1804b1-471178a6882mr170490355e9.9.1761154109943; Wed, 22 Oct 2025 10:28:29 -0700 (PDT) X-Google-Smtp-Source: AGHT+IHlRHaZovyGNJpfxC/RlqXuNv/7Z+a5XDpsqOCnLx0yEjUAljMFKF3SfTdCasSCmNwsHcnW1A== X-Received: by 2002:a05:600c:6291:b0:471:a43:123f with SMTP id 5b1f17b1804b1-471178a6882mr170490155e9.9.1761154109456; Wed, 22 Oct 2025 10:28:29 -0700 (PDT) Received: from ?IPV6:2003:d8:2f4e:3200:c99d:a38b:3f3a:d4b3? (p200300d82f4e3200c99da38b3f3ad4b3.dip0.t-ipconnect.de. [2003:d8:2f4e:3200:c99d:a38b:3f3a:d4b3]) by smtp.gmail.com with ESMTPSA id 5b1f17b1804b1-475c428c62fsm56170405e9.8.2025.10.22.10.28.28 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Wed, 22 Oct 2025 10:28:28 -0700 (PDT) Message-ID: Date: Wed, 22 Oct 2025 19:28:27 +0200 MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH] mm/filemap: Implement fast short reads To: Kiryl Shutsemau , Andrew Morton , Matthew Wilcox , Linus Torvalds , Alexander Viro , Christian Brauner , Jan Kara Cc: linux-mm@kvack.org, linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, Kiryl Shutsemau References: <20251017141536.577466-1-kirill@shutemov.name> From: David Hildenbrand Autocrypt: addr=david@redhat.com; keydata= xsFNBFXLn5EBEAC+zYvAFJxCBY9Tr1xZgcESmxVNI/0ffzE/ZQOiHJl6mGkmA1R7/uUpiCjJ dBrn+lhhOYjjNefFQou6478faXE6o2AhmebqT4KiQoUQFV4R7y1KMEKoSyy8hQaK1umALTdL QZLQMzNE74ap+GDK0wnacPQFpcG1AE9RMq3aeErY5tujekBS32jfC/7AnH7I0v1v1TbbK3Gp XNeiN4QroO+5qaSr0ID2sz5jtBLRb15RMre27E1ImpaIv2Jw8NJgW0k/D1RyKCwaTsgRdwuK Kx/Y91XuSBdz0uOyU/S8kM1+ag0wvsGlpBVxRR/xw/E8M7TEwuCZQArqqTCmkG6HGcXFT0V9 PXFNNgV5jXMQRwU0O/ztJIQqsE5LsUomE//bLwzj9IVsaQpKDqW6TAPjcdBDPLHvriq7kGjt WhVhdl0qEYB8lkBEU7V2Yb+SYhmhpDrti9Fq1EsmhiHSkxJcGREoMK/63r9WLZYI3+4W2rAc UucZa4OT27U5ZISjNg3Ev0rxU5UH2/pT4wJCfxwocmqaRr6UYmrtZmND89X0KigoFD/XSeVv jwBRNjPAubK9/k5NoRrYqztM9W6sJqrH8+UWZ1Idd/DdmogJh0gNC0+N42Za9yBRURfIdKSb B3JfpUqcWwE7vUaYrHG1nw54pLUoPG6sAA7Mehl3nd4pZUALHwARAQABzSREYXZpZCBIaWxk ZW5icmFuZCA8ZGF2aWRAcmVkaGF0LmNvbT7CwZoEEwEIAEQCGwMCF4ACGQEFCwkIBwICIgIG FQoJCAsCBBYCAwECHgcWIQQb2cqtc1xMOkYN/MpN3hD3AP+DWgUCaJzangUJJlgIpAAKCRBN 3hD3AP+DWhAxD/9wcL0A+2rtaAmutaKTfxhTP0b4AAp1r/eLxjrbfbCCmh4pqzBhmSX/4z11 opn2KqcOsueRF1t2ENLOWzQu3Roiny2HOU7DajqB4dm1BVMaXQya5ae2ghzlJN9SIoopTWlR 0Af3hPj5E2PYvQhlcqeoehKlBo9rROJv/rjmr2x0yOM8qeTroH/ZzNlCtJ56AsE6Tvl+r7cW 3x7/Jq5WvWeudKrhFh7/yQ7eRvHCjd9bBrZTlgAfiHmX9AnCCPRPpNGNedV9Yty2Jnxhfmbv Pw37LA/jef8zlCDyUh2KCU1xVEOWqg15o1RtTyGV1nXV2O/mfuQJud5vIgzBvHhypc3p6VZJ lEf8YmT+Ol5P7SfCs5/uGdWUYQEMqOlg6w9R4Pe8d+mk8KGvfE9/zTwGg0nRgKqlQXrWRERv cuEwQbridlPAoQHrFWtwpgYMXx2TaZ3sihcIPo9uU5eBs0rf4mOERY75SK+Ekayv2ucTfjxr Kf014py2aoRJHuvy85ee/zIyLmve5hngZTTe3Wg3TInT9UTFzTPhItam6dZ1xqdTGHZYGU0O otRHcwLGt470grdiob6PfVTXoHlBvkWRadMhSuG4RORCDpq89vu5QralFNIf3EysNohoFy2A LYg2/D53xbU/aa4DDzBb5b1Rkg/udO1gZocVQWrDh6I2K3+cCs7BTQRVy5+RARAA59fefSDR 9nMGCb9LbMX+TFAoIQo/wgP5XPyzLYakO+94GrgfZjfhdaxPXMsl2+o8jhp/hlIzG56taNdt VZtPp3ih1AgbR8rHgXw1xwOpuAd5lE1qNd54ndHuADO9a9A0vPimIes78Hi1/yy+ZEEvRkHk /kDa6F3AtTc1m4rbbOk2fiKzzsE9YXweFjQvl9p+AMw6qd/iC4lUk9g0+FQXNdRs+o4o6Qvy iOQJfGQ4UcBuOy1IrkJrd8qq5jet1fcM2j4QvsW8CLDWZS1L7kZ5gT5EycMKxUWb8LuRjxzZ 3QY1aQH2kkzn6acigU3HLtgFyV1gBNV44ehjgvJpRY2cC8VhanTx0dZ9mj1YKIky5N+C0f21 zvntBqcxV0+3p8MrxRRcgEtDZNav+xAoT3G0W4SahAaUTWXpsZoOecwtxi74CyneQNPTDjNg azHmvpdBVEfj7k3p4dmJp5i0U66Onmf6mMFpArvBRSMOKU9DlAzMi4IvhiNWjKVaIE2Se9BY FdKVAJaZq85P2y20ZBd08ILnKcj7XKZkLU5FkoA0udEBvQ0f9QLNyyy3DZMCQWcwRuj1m73D sq8DEFBdZ5eEkj1dCyx+t/ga6x2rHyc8Sl86oK1tvAkwBNsfKou3v+jP/l14a7DGBvrmlYjO 59o3t6inu6H7pt7OL6u6BQj7DoMAEQEAAcLBfAQYAQgAJgIbDBYhBBvZyq1zXEw6Rg38yk3e EPcA/4NaBQJonNqrBQkmWAihAAoJEE3eEPcA/4NaKtMQALAJ8PzprBEXbXcEXwDKQu+P/vts IfUb1UNMfMV76BicGa5NCZnJNQASDP/+bFg6O3gx5NbhHHPeaWz/VxlOmYHokHodOvtL0WCC 8A5PEP8tOk6029Z+J+xUcMrJClNVFpzVvOpb1lCbhjwAV465Hy+NUSbbUiRxdzNQtLtgZzOV Zw7jxUCs4UUZLQTCuBpFgb15bBxYZ/BL9MbzxPxvfUQIPbnzQMcqtpUs21CMK2PdfCh5c4gS sDci6D5/ZIBw94UQWmGpM/O1ilGXde2ZzzGYl64glmccD8e87OnEgKnH3FbnJnT4iJchtSvx yJNi1+t0+qDti4m88+/9IuPqCKb6Stl+s2dnLtJNrjXBGJtsQG/sRpqsJz5x1/2nPJSRMsx9 5YfqbdrJSOFXDzZ8/r82HgQEtUvlSXNaXCa95ez0UkOG7+bDm2b3s0XahBQeLVCH0mw3RAQg r7xDAYKIrAwfHHmMTnBQDPJwVqxJjVNr7yBic4yfzVWGCGNE4DnOW0vcIeoyhy9vnIa3w1uZ 3iyY2Nsd7JxfKu1PRhCGwXzRw5TlfEsoRI7V9A8isUCoqE2Dzh3FvYHVeX4Us+bRL/oqareJ CIFqgYMyvHj7Q06kTKmauOe4Nf0l0qEkIuIzfoLJ3qr5UyXc2hLtWyT9Ir+lYlX9efqh7mOY qIws/H2t In-Reply-To: <20251017141536.577466-1-kirill@shutemov.name> X-Mimecast-Spam-Score: 0 X-Mimecast-MFC-PROC-ID: ksY9Zc-Ifn89ZIHTfsW0D3INktpiH8ao127HBGdQtf8_1761154110 X-Mimecast-Originator: redhat.com Content-Language: en-US Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit X-Stat-Signature: qqbz7axjoq3m9pa7b139nak1w57xxqkb X-Rspam-User: X-Rspamd-Server: rspam07 X-Rspamd-Queue-Id: 5815E1C0007 X-HE-Tag: 1761154114-335887 X-HE-Meta: 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 XZwJieTf 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 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: List-Subscribe: List-Unsubscribe: On 17.10.25 16:15, Kiryl Shutsemau wrote: > From: Kiryl Shutsemau > > The protocol for page cache lookup is as follows: > > 1. Locate a folio in XArray. > 2. Obtain a reference on the folio using folio_try_get(). > 3. If successful, verify that the folio still belongs to > the mapping and has not been truncated or reclaimed. > 4. Perform operations on the folio, such as copying data > to userspace. > 5. Release the reference. > > For short reads, the overhead of atomic operations on reference > manipulation can be significant, particularly when multiple tasks access > the same folio, leading to cache line bouncing. > > To address this issue, introduce i_pages_delete_seqcnt, which increments > each time a folio is deleted from the page cache and implement a modified > page cache lookup protocol for short reads: > > 1. Locate a folio in XArray. > 2. Take note of the i_pages_delete_seqcnt. > 3. Copy the data to a local buffer on the stack. > 4. Verify that the i_pages_delete_seqcnt has not changed. > 5. Copy the data from the local buffer to the iterator. > > If any issues arise in the fast path, fallback to the slow path that > relies on the refcount to stabilize the folio. > > The new approach requires a local buffer in the stack. The size of the > buffer determines which read requests are served by the fast path. Set > the buffer to 1k. This seems to be a reasonable amount of stack usage > for the function at the bottom of the call stack. > > The fast read approach demonstrates significant performance > improvements, particularly in contended cases. > > 16 threads, reads from 4k file(s), mean MiB/s (StdDev) > > ------------------------------------------------------------- > | Block | Baseline | Baseline | Patched | Patched | > | size | same file | diff files | same file | diff files | > ------------------------------------------------------------- > | 1 | 10.96 | 27.56 | 30.42 | 30.4 | > | | (0.497) | (0.114) | (0.130) | (0.158) | > | 32 | 350.8 | 886.2 | 980.6 | 981.8 | > | | (13.64) | (2.863) | (3.361) | (1.303) | > | 256 | 2798 | 7009.6 | 7641.4 | 7653.6 | > | | (103.9) | (28.00) | (33.26) | (25.50) | > | 1024 | 10780 | 27040 | 29280 | 29320 | > | | (389.8) | (89.44) | (130.3) | (83.66) | > | 4096 | 43700 | 103800 | 48420 | 102000 | > | | (1953) | (447.2) | (2012) | (0) | > ------------------------------------------------------------- > > 16 threads, reads from 1M file(s), mean MiB/s (StdDev) > > -------------------------------------------------------------- > | Block | Baseline | Baseline | Patched | Patched | > | size | same file | diff files | same file | diff files | > --------------------------------------------------------- > | 1 | 26.38 | 27.34 | 30.38 | 30.36 | > | | (0.998) | (0.114) | (0.083) | (0.089) | > | 32 | 824.4 | 877.2 | 977.8 | 975.8 | > | | (15.78) | (3.271) | (2.683) | (1.095) | > | 256 | 6494 | 6992.8 | 7619.8 | 7625 | > | | (116.0) | (32.39) | (10.66) | (28.19) | > | 1024 | 24960 | 26840 | 29100 | 29180 | > | | (606.6) | (151.6) | (122.4) | (83.66) | > | 4096 | 94420 | 100520 | 95260 | 99760 | > | | (3144) | (672.3) | (2874) | (134.1) | > | 32768 | 386000 | 402400 | 368600 | 397400 | > | | (36599) | (10526) | (47188) | (6107) | > -------------------------------------------------------------- > > There's also improvement on kernel build: > > Base line: 61.3462 +- 0.0597 seconds time elapsed ( +- 0.10% ) > Patched: 60.6106 +- 0.0759 seconds time elapsed ( +- 0.13% ) > > Co-developed-by: Linus Torvalds > Signed-off-by: Linus Torvalds > Signed-off-by: Kiryl Shutsemau > --- > fs/inode.c | 2 + > include/linux/fs.h | 1 + > mm/filemap.c | 150 ++++++++++++++++++++++++++++++++++++++------- > 3 files changed, 130 insertions(+), 23 deletions(-) > > diff --git a/fs/inode.c b/fs/inode.c > index ec9339024ac3..52163d28d630 100644 > --- a/fs/inode.c > +++ b/fs/inode.c > @@ -482,6 +482,8 @@ EXPORT_SYMBOL(inc_nlink); > static void __address_space_init_once(struct address_space *mapping) > { > xa_init_flags(&mapping->i_pages, XA_FLAGS_LOCK_IRQ | XA_FLAGS_ACCOUNT); > + seqcount_spinlock_init(&mapping->i_pages_delete_seqcnt, > + &mapping->i_pages->xa_lock); > init_rwsem(&mapping->i_mmap_rwsem); > INIT_LIST_HEAD(&mapping->i_private_list); > spin_lock_init(&mapping->i_private_lock); > diff --git a/include/linux/fs.h b/include/linux/fs.h > index c895146c1444..c9588d555f73 100644 > --- a/include/linux/fs.h > +++ b/include/linux/fs.h > @@ -523,6 +523,7 @@ struct address_space { > struct list_head i_private_list; > struct rw_semaphore i_mmap_rwsem; > void * i_private_data; > + seqcount_spinlock_t i_pages_delete_seqcnt; > } __attribute__((aligned(sizeof(long)))) __randomize_layout; > /* > * On most architectures that alignment is already the case; but > diff --git a/mm/filemap.c b/mm/filemap.c > index 13f0259d993c..51689c4f3773 100644 > --- a/mm/filemap.c > +++ b/mm/filemap.c > @@ -138,8 +138,10 @@ static void page_cache_delete(struct address_space *mapping, > > VM_BUG_ON_FOLIO(!folio_test_locked(folio), folio); > > + write_seqcount_begin(&mapping->i_pages_delete_seqcnt); > xas_store(&xas, shadow); > xas_init_marks(&xas); > + write_seqcount_end(&mapping->i_pages_delete_seqcnt); > > folio->mapping = NULL; > /* Leave folio->index set: truncation lookup relies upon it */ > @@ -2695,21 +2697,98 @@ static void filemap_end_dropbehind_read(struct folio *folio) > } > } > > -/** > - * filemap_read - Read data from the page cache. > - * @iocb: The iocb to read. > - * @iter: Destination for the data. > - * @already_read: Number of bytes already read by the caller. > - * > - * Copies data from the page cache. If the data is not currently present, > - * uses the readahead and read_folio address_space operations to fetch it. > - * > - * Return: Total number of bytes copied, including those already read by > - * the caller. If an error happens before any bytes are copied, returns > - * a negative error number. > - */ > -ssize_t filemap_read(struct kiocb *iocb, struct iov_iter *iter, > - ssize_t already_read) > +static inline unsigned long filemap_read_fast_rcu(struct address_space *mapping, > + loff_t pos, char *buffer, > + size_t size) > +{ > + XA_STATE(xas, &mapping->i_pages, pos >> PAGE_SHIFT); > + struct folio *folio; > + loff_t file_size; > + unsigned int seq; > + > + lockdep_assert_in_rcu_read_lock(); > + > + /* Give up and go to slow path if raced with page_cache_delete() */ > + if (!raw_seqcount_try_begin(&mapping->i_pages_delete_seqcnt, seq)) > + return false; > + > + folio = xas_load(&xas); > + if (xas_retry(&xas, folio)) > + return 0; > + > + if (!folio || xa_is_value(folio)) > + return 0; > + > + if (!folio_test_uptodate(folio)) > + return 0; > + > + /* No fast-case if readahead is supposed to started */ > + if (folio_test_readahead(folio)) > + return 0; > + /* .. or mark it accessed */ > + if (!folio_test_referenced(folio)) > + return 0; > + > + /* i_size check must be after folio_test_uptodate() */ > + file_size = i_size_read(mapping->host); > + if (unlikely(pos >= file_size)) > + return 0; > + if (size > file_size - pos) > + size = file_size - pos; > + > + /* Do the data copy */ > + size = memcpy_from_file_folio(buffer, folio, pos, size); We can be racing with folio splitting or freeing+reallocation, right? So our folio might actually be suddenly a tail page I assume? Or change from small to large, or change from large to small. Or a large folio can change its size? In that case things like folio_test_uptodate() won't be happy, because they will bail out if called on a tail page (see PF_NO_TAIL). But also, are we sure memcpy_from_file_folio() is save to be used in that racy context? offset_in_folio() which depends on folio_size(). When racing with concurrent folio reuse, folio_size()->folio_order() can be tricked into returning a wrong number I guess and making the result of offset_in_folio() rather bogus. At least that's my understanding after taking a quick look. The concern I would have in that case (A) the page pointer we calculate in kmap_local_folio() could be garbage (B) the "from" pointer we return could be garbage "garbage" as in pointing at something without a direct map, something that's protected differently (MTE? weird CoCo protection?) or even worse MMIO with undesired read-effects. I'll note that some of these concerns would be gone once folios are allocated separately: we would not suddenly see a tail page. But concurrent folio splitting or size changes could still be an issue IIUC Willy today once I discussed the plans about "struct folio" freeing. Maybe I'm wrong, just wanted to raise that these functions are not prepared to be called when we are not holding a refcount or cannot otherwise guarantee that the folio cannot concurrently be freed/reallocated/merged/split. -- Cheers David / dhildenb