<?xml version='1.0' encoding='UTF-8'?>
<?xml-stylesheet href="/static/style.xsl" type="text/xsl"?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" version="2.0">
  <channel>
    <title>Most recent entries from all</title>
    <link>https://vulnerability.circl.lu</link>
    <description>Contains only the most 10 recent entries.</description>
    <docs>http://www.rssboard.org/rss-specification</docs>
    <generator>python-feedgen</generator>
    <language>en</language>
    <lastBuildDate>Thu, 01 Oct 2026 20:28:20 +0000</lastBuildDate>
    <item>
      <title>CVE-2026-74346 — RDMA/irdma: Fix OOB read during CQ MR registration</title>
      <link>https://vulnerability.circl.lu/vuln/cve-2026-74346</link>
      <description>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; Linux&lt;/p&gt;
&lt;p&gt;In the Linux kernel, the following vulnerability has been resolved:&lt;/p&gt;
&lt;p&gt;RDMA/irdma: Fix OOB read during CQ MR registration&lt;/p&gt;
&lt;p&gt;Sashiko pointed out an unrelated bug during a previous patch:
https://sashiko.dev/#/patchset/20260512183852.614045-1-jmoroni%40google.com&lt;/p&gt;
&lt;p&gt;This change fixes the bug by eliminating the cqmr-&amp;gt;split field which
was not being set properly and instead just checks the CQ resize
feature flag directly.&lt;/p&gt;
&lt;p&gt;The cqmr-&amp;gt;split field essentially tracks whether IRDMA_FEATURE_CQ_RESIZE
is set, but it was not being set until CQ creation time, which is _after_
CQ memory registration (the only other place where it is referenced).&lt;/p&gt;
&lt;p&gt;As a result, it would always be false during MR registration and would
therefore cause irdma_handle_q_mem to populate cqmr-&amp;gt;shadow even for GEN_2
HW and beyond:&lt;/p&gt;
&lt;p&gt;cqmr-&amp;gt;shadow = (dma_addr_t)arr[req-&amp;gt;cq_pages];&lt;/p&gt;
&lt;p&gt;The issue is that for GEN_2 and beyond, req-&amp;gt;cq_pages may be exactly equal
to iwmr-&amp;gt;page_cnt and therefore equal to the size of arr, which would cause
an OOB read by one.&lt;/p&gt;</description>
      <content:encoded>&lt;p&gt;&lt;strong&gt;Affected:&lt;/strong&gt; Linux&lt;/p&gt;
&lt;p&gt;In the Linux kernel, the following vulnerability has been resolved:&lt;/p&gt;
&lt;p&gt;RDMA/irdma: Fix OOB read during CQ MR registration&lt;/p&gt;
&lt;p&gt;Sashiko pointed out an unrelated bug during a previous patch:
https://sashiko.dev/#/patchset/20260512183852.614045-1-jmoroni%40google.com&lt;/p&gt;
&lt;p&gt;This change fixes the bug by eliminating the cqmr-&amp;gt;split field which
was not being set properly and instead just checks the CQ resize
feature flag directly.&lt;/p&gt;
&lt;p&gt;The cqmr-&amp;gt;split field essentially tracks whether IRDMA_FEATURE_CQ_RESIZE
is set, but it was not being set until CQ creation time, which is _after_
CQ memory registration (the only other place where it is referenced).&lt;/p&gt;
&lt;p&gt;As a result, it would always be false during MR registration and would
therefore cause irdma_handle_q_mem to populate cqmr-&amp;gt;shadow even for GEN_2
HW and beyond:&lt;/p&gt;
&lt;p&gt;cqmr-&amp;gt;shadow = (dma_addr_t)arr[req-&amp;gt;cq_pages];&lt;/p&gt;
&lt;p&gt;The issue is that for GEN_2 and beyond, req-&amp;gt;cq_pages may be exactly equal
to iwmr-&amp;gt;page_cnt and therefore equal to the size of arr, which would cause
an OOB read by one.&lt;/p&gt;</content:encoded>
      <guid isPermaLink="false">https://vulnerability.circl.lu/vuln/cve-2026-74346</guid>
    </item>
  </channel>
</rss>
