<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Nvme on vmlinuz</title>
    <link>https://vmlinuz.pl/en/tags/nvme/</link>
    <description>Recent content in Nvme on vmlinuz</description>
    <generator>Hugo</generator>
    <language>en</language>
    <copyright>2025-2026 vmlinuz.pl</copyright>
    <lastBuildDate>Tue, 23 Jun 2026 10:00:00 +0100</lastBuildDate>
    <atom:link href="https://vmlinuz.pl/en/tags/nvme/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Two Lines of Code, &#43;5% IOPS: How Moving a memset Sped Up SSDs in Linux</title>
      <link>https://vmlinuz.pl/en/posts/iomap-memset-iops/</link>
      <pubDate>Tue, 23 Jun 2026 10:00:00 +0100</pubDate>
      <guid>https://vmlinuz.pl/en/posts/iomap-memset-iops/</guid>
      <description>&lt;p&gt;Sometimes you do not need a new driver or a rewritten subsystem to speed up the kernel. Sometimes it is enough to move a couple of lines that were doing pointless work for years. That is exactly what Fengnan Chang from ByteDance did: he moved a &lt;code&gt;memset()&lt;/code&gt; call in the iomap layer and squeezed out about 5% more IOPS on ext4 and XFS. The change lands in Linux 7.2.&lt;/p&gt;&#xA;&lt;h2 id=&#34;what-iomap-is&#34;&gt;What iomap Is&lt;/h2&gt;&#xA;&lt;p&gt;iomap is the kernel&amp;rsquo;s shared I/O mapping layer. Filesystems like ext4, XFS and exFAT use it to translate a file offset into a concrete location on the block device. The heart of the layer is the &lt;code&gt;iomap_iter()&lt;/code&gt; loop: for a given file range it asks the filesystem to fill in a &lt;code&gt;struct iomap&lt;/code&gt; describing the next contiguous chunk, performs the I/O, then asks for the next chunk until the range is done.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
