LWN: Comments on "Last 3.18.x stable release: 3.18.140" https://lwn.net/Articles/788689/ This is a special feed containing comments posted to the individual LWN article titled "Last 3.18.x stable release: 3.18.140". en-us Wed, 08 Oct 2025 23:18:52 +0000 Wed, 08 Oct 2025 23:18:52 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net Last 3.18.x stable release: 3.18.140 https://lwn.net/Articles/788813/ https://lwn.net/Articles/788813/ hailfinger <div class="FormattedComment"> Oh, that SoC vendor is not the only bad one, and they've exited the smartphone market anyway. Specifically, if we're talking about SoCs for Android smartphones released since 2018, the following SoC vendors have some market presence: HiSilicon, Mediatek, Qualcomm, Samsung, Spreadtrum.<br> Neither of them is really good at upstreaming support for older SoCs, but there are those with an abysmal track record and those slowly getting the hang of it for select SoCs. Updates for the development kits of older SoCs to recent kernels only happen if the target market niche requires a current Android version.<br> <p> There are some positive exceptions in upstreaming (see <a href="https://lwn.net/Articles/771974/">https://lwn.net/Articles/771974/</a> and associated comments for a perspective from Q4/2018 and <a href="https://lwn.net/Articles/738225/">https://lwn.net/Articles/738225/</a> for a perspective from Q4/2017). Upstream support for older SoCs is essentially a community-only project, whereas upstream support for newer SoCs is mostly driven by OEMs (Google, Sony) making SoC selection for some products dependent on reasonable upstream policies of the SoC vendor. Most notable on the upstreaming front is the Qualcomm SDM845, but this does not extend to newer SoCs.<br> <p> Google enforcing minimum kernel versions for newer Android versions (Linux 3.18 for Android Oreo, Linux 4.4.107 for Android Pie, and Linux 4.9.recent for Android Q) helped a lot in getting the SoC development kits updated.<br> </div> Sat, 18 May 2019 01:27:52 +0000 Last 3.18.x stable release: 3.18.140 https://lwn.net/Articles/788711/ https://lwn.net/Articles/788711/ adobriyan <div class="FormattedComment"> If you think .140 was bad there is 4.4!<br> <p> That 20 year über stable kernel may break .500 barrier at ~2.5 releases per month.<br> </div> Thu, 16 May 2019 18:40:16 +0000 Last 3.18.x stable release: 3.18.140 https://lwn.net/Articles/788707/ https://lwn.net/Articles/788707/ NightMonkey <div class="FormattedComment"> This is one of the many reasons I love LWN. OftTImes the commentators offer better coverage for non-Linux topics than sites that presume to be the TItular authorities in their space. ;)<br> </div> Thu, 16 May 2019 18:06:37 +0000 Last 3.18.x stable release: 3.18.140 https://lwn.net/Articles/788702/ https://lwn.net/Articles/788702/ nix <div class="FormattedComment"> Read gscmecher's post more carefully. Pay particular attention to any capiTal letters. They are sIgnificant.<br> </div> Thu, 16 May 2019 16:54:00 +0000 Last 3.18.x stable release: 3.18.140 https://lwn.net/Articles/788700/ https://lwn.net/Articles/788700/ jfebrer <div class="FormattedComment"> I really would appreciate to know which are those problematic vendors, and I think that I'm not the only one which will avoid them if we know who they are.<br> </div> Thu, 16 May 2019 16:50:55 +0000 Last 3.18.x stable release: 3.18.140 https://lwn.net/Articles/788697/ https://lwn.net/Articles/788697/ gsmecher <div class="FormattedComment"> It's frustrating that kernel developers have to carry water for SOC vendors who either don't Get It, or who've realized it's cheaper that way.<br> <p> There's a whisper network among hardware engineers about which vendors have better or worse support. It's not just idle grumbling; reputation really does affect component selection and there are cerTaIn vendors I've learned to avoid.<br> <p> I've often wondered if it's worth publicly calling out problematic vendors, rather than implying as Greg does here. Reputation matters and the kernel community has a fairly powerful bullhorn at its disposal.<br> </div> Thu, 16 May 2019 16:33:37 +0000