CINXE.COM
LKML: Steven Rostedt: Re: On community influencing (was Re: [PATCH v8 2/2] rust: add dma coherent allocator abstraction.)
<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>LKML: Steven Rostedt: Re: On community influencing (was Re: [PATCH v8 2/2] rust: add dma coherent allocator abstraction.)</title><link href="/css/message.css" rel="stylesheet" type="text/css" /><link href="/css/wrap.css" rel="alternate stylesheet" type="text/css" title="wrap" /><link href="/css/nowrap.css" rel="stylesheet" type="text/css" title="nowrap" /><link href="/favicon.ico" rel="shortcut icon" /><script src="/js/simple-calendar.js" type="text/javascript"></script><script src="/js/styleswitcher.js" type="text/javascript"></script><link rel="alternate" type="application/rss+xml" title="lkml.org : last 100 messages" href="/rss.php" /><link rel="alternate" type="application/rss+xml" title="lkml.org : last messages by Steven Rostedt" href="/groupie.php?aid=" /><!--Matomo--><script> var _paq = window._paq = window._paq || []; /* tracker methods like "setCustomDimension" should be called before "trackPageView" */ _paq.push(["setDoNotTrack", true]); _paq.push(["disableCookies"]); _paq.push(['trackPageView']); _paq.push(['enableLinkTracking']); (function() { var u="//m.lkml.org/"; _paq.push(['setTrackerUrl', u+'matomo.php']); _paq.push(['setSiteId', '1']); var d=document, g=d.createElement('script'), s=d.getElementsByTagName('script')[0]; g.async=true; g.src=u+'matomo.js'; s.parentNode.insertBefore(g,s); })(); </script><!--End Matomo Code--></head><body onload="es.jasper.simpleCalendar.init();" itemscope="itemscope" itemtype="http://schema.org/BlogPosting"><table border="0" cellpadding="0" cellspacing="0"><tr><td width="180" align="center"><a href="/"><img style="border:0;width:135px;height:32px" src="/images/toprowlk.gif" alt="lkml.org" /></a></td><td width="32">聽</td><td class="nb"><div><a class="nb" href="/lkml"> [lkml]</a> 聽 <a class="nb" href="/lkml/2025"> [2025]</a> 聽 <a class="nb" href="/lkml/2025/2"> [Feb]</a> 聽 <a class="nb" href="/lkml/2025/2/7"> [7]</a> 聽 <a class="nb" href="/lkml/last100"> [last100]</a> 聽 <a href="/rss.php"><img src="/images/rss-or.gif" border="0" alt="RSS Feed" /></a></div><div>Views: <a href="#" class="nowrap" onclick="setActiveStyleSheet('wrap');return false;">[wrap]</a><a href="#" class="wrap" onclick="setActiveStyleSheet('nowrap');return false;">[no wrap]</a> 聽 <a class="nb" href="/lkml/mheaders/2025/2/7/1955" onclick="this.href='/lkml/headers'+'/2025/2/7/1955';">[headers]</a>聽 <a href="/lkml/bounce/2025/2/7/1955">[forward]</a>聽 </div></td><td width="32">聽</td></tr><tr><td valign="top"><div class="es-jasper-simpleCalendar" baseurl="/lkml/"></div><div class="threadlist">Messages in this thread</div><ul class="threadlist"><li class="root"><a href="/lkml/2025/1/8/801">First message in thread</a></li><li><a href="/lkml/2025/2/6/1292">Linus Torvalds</a><ul><li><a href="/lkml/2025/2/7/755">"Dr. Greg"</a><ul><li class="origin"><a href="/lkml/2025/2/7/1957">Steven Rostedt</a><ul><li><a href="/lkml/2025/2/7/1957">Steven Rostedt</a></li><li><a href="/lkml/2025/2/8/196">Hector Martin</a><ul><li><a href="/lkml/2025/2/10/443">Icenowy Zheng</a></li></ul></li></ul></li><li><a href="/lkml/2025/2/8/494">"Theodore Ts'o"</a><ul><li><a href="/lkml/2025/2/9/43">Danilo Krummrich</a></li><li><a href="/lkml/2025/2/8/556">comex</a></li><li><a href="/lkml/2025/2/13/571">David Airlie</a><ul><li><a href="/lkml/2025/2/20/1391">Simona Vetter</a></li></ul></li><li><a href="/lkml/2025/2/13/1665">Ronja Meyer</a></li></ul></li></ul></li><li><a href="/lkml/2025/2/13/1645">33KK</a></li></ul></li></ul></td><td width="32" rowspan="2" class="c" valign="top"><img src="/images/icornerl.gif" width="32" height="32" alt="/" /></td><td class="c" rowspan="2" valign="top" style="padding-top: 1em"><table><tr><td><table><tr><td class="lp">Date</td><td class="rp" itemprop="datePublished">Fri, 7 Feb 2025 23:26:50 -0500</td></tr><tr><td class="lp">From</td><td class="rp" itemprop="author">Steven Rostedt <></td></tr><tr><td class="lp">Subject</td><td class="rp" itemprop="name">Re: On community influencing (was Re: [PATCH v8 2/2] rust: add dma coherent allocator abstraction.)</td></tr></table></td><td></td></tr></table><pre itemprop="articleBody">On Fri, Feb 07, 2025 at 06:16:38AM -0600, Dr. Greg wrote:<br />> <br />> Not sure what the fix is, from a project management perspective the<br />> technology industry has never faced a challenge like this. The fork<br />> model, which was the classic protection in open-source, doesn't work<br />> at this scale.<br /><br />Maybe not quite a fork, but I wonder if the Rust project did something similar<br />to what PREEMPT_RT did. That was to keep an out of tree patch. The full patch<br />was just merged last September after being out of tree for a good 20 years.<br /><br />In the beginning there was a few things in that patch that Christoph was<br />against, but over time he became accepting of it.<br /><br />Yes, being out of tree is very difficult because you have to constantly rebase<br />(we actually found quilt being a better tool than git in this case!). But it<br />also gives you full flexibility to try new approaches. Just because something<br />is out of tree doesn't mean it can't be published and used. Red Hat and SUSE,<br />as well as many others shipped PREEMPT_RT while it was out of tree.<br /><br />Note, even though PREEMPT_RT started in 2004 and wasn't fully merged until<br />2024, it slowly did creep in bit by bit. For example, here's a few things that<br />came from the RT patch, and each was rewritten at least 3 times to become<br />acceptable by the upstream maintainers:<br /><br /> - NOHZ<br /> - High res timers<br /> - threaded interrupts<br /> - mutex code (yes, before RT everything used a semaphore)<br /> - lockdep<br /> - ftrace<br /> - generic interrupt code<br /> - generic timer code<br /> - priority inheritance<br /> - SCHED_DEADLINE<br /> - RT push/pull scheduling<br /><br />and more.<br /><br />Point being, if you are having issues with getting code upstream, don't be<br />afraid to make it out of tree. Just make it easy for others to use. And slowly<br />move your code into mainline. This is also a way to prove beyond a doubt how<br />useful your code is. The number of users is a technical argument. Linus does<br />pull things in when there is a large user base, even over other maintainer's<br />objections (see sched_ext).<br /><br />It worked for us, it can work for you ;-)<br /><br />-- Steve<br /><br /><br /></pre></td><td width="32" rowspan="2" class="c" valign="top"><img src="/images/icornerr.gif" width="32" height="32" alt="\" /></td></tr><tr><td align="right" valign="bottom"> 聽 </td></tr><tr><td align="right" valign="bottom">聽</td><td class="c" valign="bottom" style="padding-bottom: 0px"><img src="/images/bcornerl.gif" width="32" height="32" alt="\" /></td><td class="c">聽</td><td class="c" valign="bottom" style="padding-bottom: 0px"><img src="/images/bcornerr.gif" width="32" height="32" alt="/" /></td></tr><tr><td align="right" valign="top" colspan="2"> 聽 </td><td class="lm">Last update: 2025-02-08 05:27 聽聽 [W:4.702 / U:0.910 seconds]<br />漏2003-2020 <a href="http://blog.jasper.es/"><span itemprop="editor">Jasper Spaans</span></a>|hosted at <a href="https://www.digitalocean.com/?refcode=9a8e99d24cf9">Digital Ocean</a> and my Meterkast|<a href="http://blog.jasper.es/categories.html#lkml-ref">Read the blog</a></td><td>聽</td></tr></table><script language="javascript" src="/js/styleswitcher.js" type="text/javascript"></script></body></html>