<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>LayerscapeのトピックLS1028A / Cortex-A72: Is software L1/L2 cache ECC error injection supported?</title>
    <link>https://community.nxp.com/t5/Layerscape/LS1028A-Cortex-A72-Is-software-L1-L2-cache-ECC-error-injection/m-p/2411119#M16860</link>
    <description>&lt;DIV class=""&gt;&lt;SPAN&gt;I need to confirm whether L1/L2 cache ECC can be&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN&gt;tested by injection.&lt;/SPAN&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&amp;nbsp;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;From the Cortex-A72 TRM I understand:&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;- L1D and L2 provide ECC&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;- Errors are reported via CPUMERRSR_EL1 / L2MERRSR_EL1&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;- I have not found any software-accessible way to inject ECC errors&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN&gt;into the A72 L1D or L2 cache arrays&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&amp;nbsp;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;Questions:&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;1) Does LS1028A (or the A72 implementation on this part) provide any&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN&gt;software-accessible way to inject ECC errors into L1D/L2?&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;2) If injection is not supported, what does NXP recommend for verifying&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN&gt;L1/L2 ECC detect/correct behavior in a power-on BIT context?&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;</description>
    <pubDate>Thu, 03 Sep 2026 21:23:13 GMT</pubDate>
    <dc:creator>philz12</dc:creator>
    <dc:date>2026-09-03T21:23:13Z</dc:date>
    <item>
      <title>LS1028A / Cortex-A72: Is software L1/L2 cache ECC error injection supported?</title>
      <link>https://community.nxp.com/t5/Layerscape/LS1028A-Cortex-A72-Is-software-L1-L2-cache-ECC-error-injection/m-p/2411119#M16860</link>
      <description>&lt;DIV class=""&gt;&lt;SPAN&gt;I need to confirm whether L1/L2 cache ECC can be&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN&gt;tested by injection.&lt;/SPAN&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&amp;nbsp;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;From the Cortex-A72 TRM I understand:&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;- L1D and L2 provide ECC&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;- Errors are reported via CPUMERRSR_EL1 / L2MERRSR_EL1&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;- I have not found any software-accessible way to inject ECC errors&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN&gt;into the A72 L1D or L2 cache arrays&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&amp;nbsp;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;Questions:&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;1) Does LS1028A (or the A72 implementation on this part) provide any&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN&gt;software-accessible way to inject ECC errors into L1D/L2?&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;&lt;DIV class=""&gt;&lt;DIV class=""&gt;&lt;SPAN&gt;2) If injection is not supported, what does NXP recommend for verifying&amp;nbsp;&lt;/SPAN&gt;&lt;SPAN&gt;L1/L2 ECC detect/correct behavior in a power-on BIT context?&lt;/SPAN&gt;&lt;/DIV&gt;&lt;/DIV&gt;</description>
      <pubDate>Thu, 03 Sep 2026 21:23:13 GMT</pubDate>
      <guid>https://community.nxp.com/t5/Layerscape/LS1028A-Cortex-A72-Is-software-L1-L2-cache-ECC-error-injection/m-p/2411119#M16860</guid>
      <dc:creator>philz12</dc:creator>
      <dc:date>2026-09-03T21:23:13Z</dc:date>
    </item>
    <item>
      <title>Re: LS1028A / Cortex-A72: Is software L1/L2 cache ECC error injection supported?</title>
      <link>https://community.nxp.com/t5/Layerscape/LS1028A-Cortex-A72-Is-software-L1-L2-cache-ECC-error-injection/m-p/2411367#M16866</link>
      <description>&lt;P&gt;Hello,&lt;/P&gt;
&lt;H2 class="_9k2iva0 p8i6j0c _1ibi0s314 heading2 _9k2iva1"&gt;Q1: Does LS1028A / Cortex-A72 Support Software ECC Error Injection into L1D/L2?&lt;/H2&gt;
&lt;P class="p8i6j01 paragraph"&gt;No — the Cortex-A72 does not support software-accessible ECC error injection for L1 or L2 cache.&lt;/P&gt;
&lt;P&gt;&amp;nbsp;&lt;/P&gt;
&lt;H2 class="_9k2iva0 p8i6j0c _1ibi0s314 heading2 _9k2iva1"&gt;Q2: NXP-Recommended Alternatives for Power-On BIT Verification of L1/L2 ECC&lt;/H2&gt;
&lt;P class="p8i6j01 paragraph"&gt;Since injection is not available on the A72, NXP's position is that direct ECC error injection testing of L1/L2 is not feasible in software on this core, and customers requiring this level of BIT coverage should consider the following approaches:&lt;/P&gt;
&lt;H3 class="_9k2iva0 p8i6j0c _1ibi0s314 heading3 _9k2iva1"&gt;1. Verify ECC Reporting Path via CPUMERRSR_EL1 / L2MERRSR_EL1&lt;/H3&gt;
&lt;H3 class="_9k2iva0 p8i6j0c _1ibi0s314 heading3 _9k2iva1"&gt;2. Consult ARM Directly for Any Undocumented Debug Hooks&lt;/H3&gt;
&lt;H3 class="_9k2iva0 p8i6j0c _1ibi0s314 heading3 _9k2iva1"&gt;3. Rely on ARM's Architectural Validation (Design-Level Assurance)&lt;/H3&gt;
&lt;P class="p8i6j01 paragraph"&gt;For safety-critical applications, the standard industry approach when injection is not available is to rely on ARM's own silicon validation and the architectural guarantee that SECDED ECC is implemented. The LS1028A datasheet confirms &lt;EM&gt;"Parity and ECC protected 32 KB L1 instruction and 32 KB L1 data cache"&lt;/EM&gt; and &lt;EM&gt;"1 MB L2 cache with ECC protection."&lt;/EM&gt;&lt;A id="base-ui-:r18s:" class="wdbi343 wdbi341 _1ibi0s3ec _1ibi0s376" tabindex="0" role="button" href="https://nxp1.sharepoint.com/teams/210_48/_layouts/15/Doc.aspx?sourcedoc=%7BBA7CEC26-17DC-49B5-9DB1-0A4A112115D0%7D&amp;amp;file=260729-Layerscape_roadmap_to_iMX9.pptx&amp;amp;action=edit&amp;amp;mobileredirect=true&amp;amp;web=1" rel="noreferrer" aria-expanded="false" aria-haspopup="dialog" data-base-ui-click-trigger="" data-label="5" aria-label="Citation 5" target="_blank"&gt;&lt;/A&gt;ARM's silicon qualification process covers ECC correctness at the IP level.&lt;/P&gt;
&lt;H3 class="_9k2iva0 p8i6j0c _1ibi0s314 heading3 _9k2iva1"&gt;4. Consider Migration to a Newer Core for Full BIT Coverage&lt;/H3&gt;
&lt;P class="p8i6j01 paragraph"&gt;If software-injectable ECC BIT is a hard requirement (e.g., for IEC 61508 / DO-254 compliance), NXP's newer Cortex-A55–based SoCs (e.g., i.MX 93) provide dedicated ECC Error Injection registers (e.g., &lt;CODE class="p8i6j0f"&gt;CODE_CACHE_TAG0_ECC_ERROR_INJEC&lt;/CODE&gt;, &lt;CODE class="p8i6j0f"&gt;SYSTEM_CACHE_DATA0_ECC_ERROR_INJEC&lt;/CODE&gt;) that are explicitly designed for verification and debugging. &lt;A id="base-ui-:r18v:" class="wdbi343 wdbi341 _1ibi0s3ec _1ibi0s376" tabindex="0" role="button" href="https://www.nxp.com/docs/en/reference-manual/IMX93RM.pdf" rel="noreferrer" aria-expanded="false" aria-haspopup="dialog" data-base-ui-click-trigger="" data-label="6" aria-label="Citation 6" target="_blank"&gt;&lt;/A&gt;The A72 → A55/A78AE migration path is something NXP actively supports.&lt;/P&gt;
&lt;P class="p8i6j01 paragraph"&gt;&amp;nbsp;&lt;/P&gt;
&lt;P class="p8i6j01 paragraph"&gt;Regards&lt;/P&gt;
&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Fri, 04 Sep 2026 13:25:55 GMT</pubDate>
      <guid>https://community.nxp.com/t5/Layerscape/LS1028A-Cortex-A72-Is-software-L1-L2-cache-ECC-error-injection/m-p/2411367#M16866</guid>
      <dc:creator>Bio_TICFSL</dc:creator>
      <dc:date>2026-09-04T13:25:55Z</dc:date>
    </item>
  </channel>
</rss>

