CINXE.COM

<?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: Hector Martin: 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 Hector Martin" href="/groupie.php?aid=" /><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></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/2025/2/8"> [8]</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/2025/2/8/196" onclick="this.href='/lkml/headers'+'/2025/2/8/196';">[headers]</a>聽 <a href="/lkml/bounce/2025/2/8/196">[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/2025/2/7/755">"Dr. Greg"</a><ul><li><a href="/lkml/2025/2/7/1955">Steven Rostedt</a><ul><li><a href="/lkml/2025/2/7/1957">Steven Rostedt</a></li><li class="origin"><a href="/lkml/2025/2/10/443">Hector Martin</a><ul><li><a href="/lkml/2025/2/10/443">Icenowy Zheng</a><ul><li><a href="/lkml/2025/2/10/554">Danilo Krummrich</a></li></ul></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><ul><li><a href="/lkml/2025/2/20/1417">Jason Gunthorpe</a></li></ul></li></ul></li><li><a href="/lkml/2025/2/13/1665">Ronja Meyer</a></li></ul></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">Sat, 8 Feb 2025 17:31:04 +0900</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><tr><td class="lp">From</td><td class="rp" itemprop="author">Hector Martin <></td></tr></table></td><td></td></tr></table><pre itemprop="articleBody"><br /><br />On 2025/02/08 13:26, Steven Rostedt wrote:<br />> 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 />By kernel and mesa policy, while a drm/ kernel driver is not merged with<br />a stable UAPI, the mesa side cannot be merged/enabled either. That means<br />that every day the driver isn't upstream, our users suffer due to<br />various container technologies not shipping the driver in userspace<br />since they ship upstream mesa.<br /><br />The process is *designed* to encourage upstreaming early and *punish*<br />downstream trees.<br /><br />It's not just distros shipping, or even typical container technologies<br />like Flatpak (we have a workaround for that specific one). We'd need to<br />get stuff like Android/AOSP to ship the patched userspace driver so it<br />works in Waydroid, which some people want to use. It is an impossible task.<br /><br />Never mind that distros aren't, in fact, inclined to ship out-of-tree<br />kernels. PREEMPT_RT was an exception due to its general usefulness, and<br />even then was not available everywhere, and had a reasonable excuse for<br />being out-of-tree so long due to its intrusiveness in core code. Rust<br />and the Rust drivers aren't even remotely as intrusive, they just stay<br />off to the side in their own directories. Nobody wants to ship<br />downstream kernels just for random hardware support.<br /><br />Out of tree doesn't work. People, both in the kernel and out of the<br />kernel, want things in tree.<br /><br />- Hector<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 09:31 聽聽 [W:3.405 / U:1.375 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>