From nobody Mon Dec 1 21:29:23 2025 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 1777924A06D for ; Mon, 1 Dec 2025 19:19:47 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1764616789; cv=none; b=WW71BHlJsOnprRpKO3Spo9UUK0xacrwAD+WcUXhnTBRp7jgqlvJn0CDGt/zlP8OPqBKCaBYqnGnJoO5/1cNo/gG/qaio0VkXbRpErHnrCgbxE4WxN0mIaRIp8NsTe0xEig1RAKGIN7dt5x+SPcJew8l90QM1tdwF4T3C/b7xF2M= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1764616789; c=relaxed/simple; bh=SO/r1v05KSnDyaLwHJa1g/Jekz+lB9PhrclLWdtWnjo=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=qGsE0MC+2qWkYgL1loQ+4fCKsIIaXk1oUsdZTsr+Hy7TmMyxIpHutrWvSTrER/Y1B+6pucOJa4/lQIC/e2PCM1hW9G+VzKp8Of0o/ftDMAqX/S7dxZsFiIA7RaUIMmAe+ofAd6WJa/g+AV+SnOSNTkC/cAo2Exf5rzjwIuZgRNA= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=gDZwOpI3; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="gDZwOpI3" Received: by smtp.kernel.org (Postfix) with ESMTPSA id C8A98C4CEF1; Mon, 1 Dec 2025 19:19:46 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1764616786; bh=SO/r1v05KSnDyaLwHJa1g/Jekz+lB9PhrclLWdtWnjo=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=gDZwOpI3jV+iaqgiMSTCzGNGvPDFIMKwMr7104bIwDV4v12JskWj70R6yj/bAxqgk Kz3QD+OgiAzOqf8YdUL8uK0c+ZzgVcL4OWDG8EqAFjsrRXoVRNDDSMAQsUAWGVyWsM OpZETwlgObpofA8tf8HYgK5w7442PmkgGfxTvCQna1ZoVop0Ug2Rs7bU6jQRPBizSX sEpr3E9J/onaUGY1puDsnbfRfwNS63agNNcujnia3QmNmPWEjML5vZ8SQKx1n7afWB dlisWT7qJfarpaIzfDXgBi3elOTudXvBzHt2joiRb4JlY4HwXrM62NPTIsitPdSyB7 ox3D3nMcOfIQA== From: Jaegeuk Kim To: linux-kernel@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net Cc: Jaegeuk Kim Subject: [PATCH 1/4] mm/readahead: fix the broken readahead for POSIX_FADV_WILLNEED Date: Mon, 1 Dec 2025 19:16:13 +0000 Message-ID: <20251201191940.883657-2-jaegeuk@kernel.org> X-Mailer: git-send-email 2.52.0.107.ga0afd4fd5b-goog In-Reply-To: <20251201191940.883657-1-jaegeuk@kernel.org> References: <20251201191940.883657-1-jaegeuk@kernel.org> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" This patch fixes the broken readahead flow for POSIX_FADV_WILLNEED, where the problem is, in force_page_cache_ra(nr_to_read), nr_to_read is cut by the below code. max_pages =3D max_t(unsigned long, bdi->io_pages, ra->ra_pages); nr_to_read =3D min_t(unsigned long, nr_to_read, max_pages); IOWs, we are not able to read ahead larger than the above max_pages which is most likely the range of 2MB and 16MB. Note, it doesn't make sense to set ra->ra_pages to the entire file size. Instead, let's fix this logic. Before: f2fs_fadvise: dev =3D (252,16), ino =3D 14, i_size =3D 4294967296 offset:0,= len:4294967296, advise:3 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D0 nr_to_read=3D512 lo= okahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D512 nr_to_read=3D512 = lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D1024 nr_to_read=3D512= lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D1536 nr_to_read=3D512= lookahead_size=3D0 After: f2fs_fadvise: dev =3D (252,16), ino =3D 14, i_size =3D 4294967296 offset:0,= len:4294967296, advise:3 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D0 nr_to_read=3D2048 l= ookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D2048 nr_to_read=3D204= 8 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D4096 nr_to_read=3D204= 8 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D6144 nr_to_read=3D204= 8 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D8192 nr_to_read=3D204= 8 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D10240 nr_to_read=3D20= 48 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D12288 nr_to_read=3D20= 48 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D14336 nr_to_read=3D20= 48 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D16384 nr_to_read=3D20= 48 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D18432 nr_to_read=3D20= 48 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D20480 nr_to_read=3D20= 48 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D22528 nr_to_read=3D20= 48 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D24576 nr_to_read=3D20= 48 lookahead_size=3D0 ... page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D1042432 nr_to_read=3D= 2048 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D1044480 nr_to_read=3D= 2048 lookahead_size=3D0 page_cache_ra_unbounded: dev=3D252:16 ino=3De index=3D1046528 nr_to_read=3D= 2048 lookahead_size=3D0 Signed-off-by: Jaegeuk Kim Reviewed-by: Chao Yu --- mm/readahead.c | 27 ++++++++++++--------------- 1 file changed, 12 insertions(+), 15 deletions(-) diff --git a/mm/readahead.c b/mm/readahead.c index 3a4b5d58eeb6..c0db049a5b7b 100644 --- a/mm/readahead.c +++ b/mm/readahead.c @@ -311,7 +311,7 @@ EXPORT_SYMBOL_GPL(page_cache_ra_unbounded); * behaviour which would occur if page allocations are causing VM writebac= k. * We really don't want to intermingle reads and writes like that. */ -static void do_page_cache_ra(struct readahead_control *ractl, +static int do_page_cache_ra(struct readahead_control *ractl, unsigned long nr_to_read, unsigned long lookahead_size) { struct inode *inode =3D ractl->mapping->host; @@ -320,45 +320,42 @@ static void do_page_cache_ra(struct readahead_control= *ractl, pgoff_t end_index; /* The last page we want to read */ =20 if (isize =3D=3D 0) - return; + return -EINVAL; =20 end_index =3D (isize - 1) >> PAGE_SHIFT; if (index > end_index) - return; + return -EINVAL; /* Don't read past the page containing the last byte of the file */ if (nr_to_read > end_index - index) nr_to_read =3D end_index - index + 1; =20 page_cache_ra_unbounded(ractl, nr_to_read, lookahead_size); + return 0; } =20 /* - * Chunk the readahead into 2 megabyte units, so that we don't pin too much - * memory at once. + * Chunk the readahead per the block device capacity, and read all nr_to_r= ead. */ void force_page_cache_ra(struct readahead_control *ractl, unsigned long nr_to_read) { struct address_space *mapping =3D ractl->mapping; - struct file_ra_state *ra =3D ractl->ra; struct backing_dev_info *bdi =3D inode_to_bdi(mapping->host); - unsigned long max_pages; + unsigned long this_chunk; =20 if (unlikely(!mapping->a_ops->read_folio && !mapping->a_ops->readahead)) return; =20 /* - * If the request exceeds the readahead window, allow the read to - * be up to the optimal hardware IO size + * Consier the optimal hardware IO size for readahead chunk. */ - max_pages =3D max_t(unsigned long, bdi->io_pages, ra->ra_pages); - nr_to_read =3D min_t(unsigned long, nr_to_read, max_pages); + this_chunk =3D max_t(unsigned long, bdi->io_pages, ractl->ra->ra_pages); + while (nr_to_read) { - unsigned long this_chunk =3D (2 * 1024 * 1024) / PAGE_SIZE; + this_chunk =3D min_t(unsigned long, this_chunk, nr_to_read); =20 - if (this_chunk > nr_to_read) - this_chunk =3D nr_to_read; - do_page_cache_ra(ractl, this_chunk, 0); + if (do_page_cache_ra(ractl, this_chunk, 0)) + break; =20 nr_to_read -=3D this_chunk; } --=20 2.52.0.107.ga0afd4fd5b-goog