<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Dpdk on Under the Model</title><link>https://oguzpastirmaci.com/tags/dpdk/</link><description>Recent content in Dpdk on Under the Model</description><generator>Hugo</generator><language>en-US</language><lastBuildDate>Tue, 20 Oct 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://oguzpastirmaci.com/tags/dpdk/index.xml" rel="self" type="application/rss+xml"/><item><title>Why Large RoCE Frames Disappeared on OVS-DOCA 3.5</title><link>https://oguzpastirmaci.com/posts/why-large-roce-frames-disappeared-on-ovs-doca-3-5/</link><pubDate>Tue, 20 Oct 2026 00:00:00 +0000</pubDate><guid>https://oguzpastirmaci.com/posts/why-large-roce-frames-disappeared-on-ovs-doca-3-5/</guid><description>&lt;p&gt;I looked into GB300 hosts where RDMA broke after an image update. &lt;code&gt;ib_write_bw&lt;/code&gt; between two hosts worked with 512-byte messages. With 64 KiB messages, it showed 0 Gb/s. The links were up on the host and on the switch. Hosts on the older image worked.&lt;/p&gt;&#10;&lt;p&gt;The cause was a small change in how OVS-DOCA reuses its receive buffer pools. On the new version, the pools kept the size for a 1500-byte MTU, although every port said 9216. RoCE frames of about 4.2 KB did not fit, and no counter called them dropped.&lt;/p&gt;</description></item></channel></rss>