TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Block-layer I/O polling merged into Linux kernel

109 点作者 chilledheart大约 9 年前

3 条评论

xlayn大约 9 年前
Interesting resume from a comment on the article<p>&gt;There are basically two types of polling on the block side. One takes care of interrupt mitigation, so that we can reduce the IRQ load in high IOPS scenarios. That is governed by block&#x2F;blk-iopoll.c, and is very much like NAPI on the networking side, we&#x27;ve had that since 2009 roughly. It still relies on an initial IRQ trigger, and from there we poll for more completions, and finally re-enable interrupts once we think it&#x27;s sane to do so. This is driver managed, and opt-in.<p>&gt;The new block poll support is a bit different. We don&#x27;t rely on an initial IRQ to trigger, since we never put the application to sleep. We can poll for a specific piece of IO, not just for &quot;any IO&quot;. It&#x27;s all about reducing latencies, as opposed to just reducing the overhead of an IRQ storm.
评论 #11208720 未加载
mangeletti大约 9 年前
Would somebody mind explaining what this means in layman terms? I understand what this change is, based on the article, but I don&#x27;t know what implications it has, from a high level viewpoint, because I don&#x27;t know much about an I&#x2F;O system in the first place. For instance, does this mean that a program now be able to detect block level disk changes via this mechanism, or something completely different?
评论 #11204740 未加载
评论 #11205076 未加载
评论 #11204728 未加载
评论 #11204600 未加载
评论 #11204652 未加载
atomic77大约 9 年前
Where this could be interesting is in cloud&#x2F;VM environments where the block device may actually be mounted over the network.<p>The performance improvement for fast devices cited in a link on the article [1] are pretty dramatic, but I wonder about how slow the device needs to be before polling becomes a problem. That same link mentions that slow devices benefit, but, speculate that it may be due to the CPU not being able to go into a deeper sleep state.<p>[1] <a href="http:&#x2F;&#x2F;lwn.net&#x2F;Articles&#x2F;663543&#x2F;" rel="nofollow">http:&#x2F;&#x2F;lwn.net&#x2F;Articles&#x2F;663543&#x2F;</a>
评论 #11213892 未加载