From owner-svn-src-all@freebsd.org Mon Jul 15 22:28:45 2019 Return-Path: Delivered-To: svn-src-all@mailman.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.nyi.freebsd.org (Postfix) with ESMTP id 591D6C55C9; Mon, 15 Jul 2019 22:28:45 +0000 (UTC) (envelope-from jhb@FreeBSD.org) Received: from smtp.freebsd.org (smtp.freebsd.org [96.47.72.83]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) server-signature RSA-PSS (4096 bits) client-signature RSA-PSS (4096 bits) client-digest SHA256) (Client CN "smtp.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id D7AC382ADF; Mon, 15 Jul 2019 22:28:44 +0000 (UTC) (envelope-from jhb@FreeBSD.org) Received: from John-Baldwins-MacBook-Pro-4.local (ralph.baldwin.cx [66.234.199.215]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (Client did not present a certificate) (Authenticated sender: jhb) by smtp.freebsd.org (Postfix) with ESMTPSA id 11F336991; Mon, 15 Jul 2019 22:28:42 +0000 (UTC) (envelope-from jhb@FreeBSD.org) Subject: Re: svn commit: r349391 - head/sys/kern To: cem@freebsd.org, Alan Somers Cc: src-committers , svn-src-all , svn-src-head References: <201906251944.x5PJiNaA093352@repo.freebsd.org> From: John Baldwin Openpgp: preference=signencrypt Autocrypt: addr=jhb@FreeBSD.org; keydata= mQGiBETQ+XcRBADMFybiq69u+fJRy/0wzqTNS8jFfWaBTs5/OfcV7wWezVmf9sgwn8TW0Dk0 c9MBl0pz+H01dA2ZSGZ5fXlmFIsee1WEzqeJzpiwd/pejPgSzXB9ijbLHZ2/E0jhGBcVy5Yo /Tw5+U/+laeYKu2xb0XPvM0zMNls1ah5OnP9a6Ql6wCgupaoMySb7DXm2LHD1Z9jTsHcAQMD /1jzh2BoHriy/Q2s4KzzjVp/mQO5DSm2z14BvbQRcXU48oAosHA1u3Wrov6LfPY+0U1tG47X 1BGfnQH+rNAaH0livoSBQ0IPI/8WfIW7ub4qV6HYwWKVqkDkqwcpmGNDbz3gfaDht6nsie5Z pcuCcul4M9CW7Md6zzyvktjnbz61BADGDCopfZC4of0Z3Ka0u8Wik6UJOuqShBt1WcFS8ya1 oB4rc4tXfSHyMF63aPUBMxHR5DXeH+EO2edoSwViDMqWk1jTnYza51rbGY+pebLQOVOxAY7k do5Ordl3wklBPMVEPWoZ61SdbcjhHVwaC5zfiskcxj5wwXd2E9qYlBqRg7QeSm9obiBCYWxk d2luIDxqaGJARnJlZUJTRC5vcmc+iGAEExECACAFAkTQ+awCGwMGCwkIBwMCBBUCCAMEFgID AQIeAQIXgAAKCRBy3lIGd+N/BI6RAJ9S97fvbME+3hxzE3JUyUZ6vTewDACdE1stFuSfqMvM jomvZdYxIYyTUpC5Ag0ERND5ghAIAPwsO0B7BL+bz8sLlLoQktGxXwXQfS5cInvL17Dsgnr3 1AKa94j9EnXQyPEj7u0d+LmEe6CGEGDh1OcGFTMVrof2ZzkSy4+FkZwMKJpTiqeaShMh+Goj XlwIMDxyADYvBIg3eN5YdFKaPQpfgSqhT+7El7w+wSZZD8pPQuLAnie5iz9C8iKy4/cMSOrH YUK/tO+Nhw8Jjlw94Ik0T80iEhI2t+XBVjwdfjbq3HrJ0ehqdBwukyeJRYKmbn298KOFQVHO EVbHA4rF/37jzaMadK43FgJ0SAhPPF5l4l89z5oPu0b/+5e2inA3b8J3iGZxywjM+Csq1tqz hltEc7Q+E08AAwUIAL+15XH8bPbjNJdVyg2CMl10JNW2wWg2Q6qdljeaRqeR6zFus7EZTwtX sNzs5bP8y51PSUDJbeiy2RNCNKWFMndM22TZnk3GNG45nQd4OwYK0RZVrikalmJY5Q6m7Z16 4yrZgIXFdKj2t8F+x613/SJW1lIr9/bDp4U9tw0V1g3l2dFtD3p3ZrQ3hpoDtoK70ioIAjjH aIXIAcm3FGZFXy503DOA0KaTWwvOVdYCFLm3zWuSOmrX/GsEc7ovasOWwjPn878qVjbUKWwx Q4QkF4OhUV9zPtf9tDSAZ3x7QSwoKbCoRCZ/xbyTUPyQ1VvNy/mYrBcYlzHodsaqUDjHuW+I SQQYEQIACQUCRND5ggIbDAAKCRBy3lIGd+N/BCO8AJ9j1dWVQWxw/YdTbEyrRKOY8YZNwwCf afMAg8QvmOWnHx3wl8WslCaXaE8= Message-ID: <3b2ade4e-d19d-31e4-7ee2-c0085a00e53e@FreeBSD.org> Date: Mon, 15 Jul 2019 15:28:40 -0700 User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:60.0) Gecko/20100101 Thunderbird/60.7.0 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 8bit X-Rspamd-Queue-Id: D7AC382ADF X-Spamd-Bar: -- Authentication-Results: mx1.freebsd.org X-Spamd-Result: default: False [-2.84 / 15.00]; local_wl_from(0.00)[FreeBSD.org]; NEURAL_HAM_MEDIUM(-1.00)[-0.996,0]; NEURAL_HAM_SHORT(-0.85)[-0.848,0]; NEURAL_HAM_LONG(-1.00)[-1.000,0]; ASN(0.00)[asn:11403, ipnet:96.47.64.0/20, country:US] X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 15 Jul 2019 22:28:45 -0000 On 7/14/19 12:08 PM, Conrad Meyer wrote: > Hi Alan, > > This change restores the possible overflow beyond IO_SEQMAX that the > removed conditional prevented. > > On Tue, Jun 25, 2019 at 12:44 PM Alan Somers wrote: >> >> Author: asomers >> Date: Tue Jun 25 19:44:22 2019 >> New Revision: 349391 >> URL: https://svnweb.freebsd.org/changeset/base/349391 >> >> --- head/sys/kern/vfs_vnops.c Tue Jun 25 19:36:01 2019 (r349390) >> +++ head/sys/kern/vfs_vnops.c Tue Jun 25 19:44:22 2019 (r349391) >> @@ -499,10 +499,8 @@ sequential_heuristic(struct uio *uio, struct file *fp) >> * closely related to the best I/O size for real disks than >> * to any block size used by software. >> */ >> - fp->f_seqcount += MIN(IO_SEQMAX, >> + fp->f_seqcount += lmin(IO_SEQMAX, >> howmany(uio->uio_resid, 16384)); >> - if (fp->f_seqcount > IO_SEQMAX) >> - fp->f_seqcount = IO_SEQMAX; >> return (fp->f_seqcount << IO_SEQSHIFT); >> } > > Perhaps instead this should be: > > fp->f_seqcount = lmin(IO_SEQMAX, > fp->f_seqcount + howmany(...)); While I agree with your assessment of the removed conditional, I think the suggestion above can have a funky overflow where f_seqcount + howmany(...) ends up less than IO_SEQMAX and this expression would then be used instead of IO_SEQMAX? The first lmin seems designed to minimize what we add in the hope that since the existing value is bounded to '0...IO_SEQMAX' the result of the += can only be in the range '0...2*IO_SEQMAX'. I'm not sure which variants are most readable: A) fp->f_seqcount += lmin(IO_SEQMAX, howmany(resid, 16384)); if (fp->f_seqcount > IO_SEQMAX) fp->f_seqcount = IO_SEQMAX; B) fp->f_seqcount += lmin(IO_SEQMAX, howmany(resid, 16384)); fp->f_seqcount = lmin(fp->f_seqcount, IO_SEQMAX); C) if (fp->f_seqcount + howmany(resid, 16384) < fp->f_seqcount) fp->f_seqcount = IO_SEQMAX; else fp->f_seqcount = lmin(IO_SEQMAX, fp->f_seqcount + howmany(resid, 16384)); I'm probably not a fan of C). -- John Baldwin