<?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>Processor Expert Softwareのトピックlx2160 dpdk 17.11.2 dpaa2 rte_mbuf hash always zero</title>
    <link>https://community.nxp.com/t5/Processor-Expert-Software/lx2160-dpdk-17-11-2-dpaa2-rte-mbuf-hash-always-zero/m-p/803039#M4180</link>
    <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;P&gt;I have found while testing with dpdk 17.11.2 on the lx2160 that the hash field of the rte_mbuf is always zero.&lt;/P&gt;&lt;P&gt;It appears from the dpdk source that the value is not set in the dpaa2 driver but is in the dpaa driver.&lt;/P&gt;&lt;P&gt;The files I am referring to are&lt;BR /&gt;drivers/net/dpaa2/dpaa2_rxtx.c&lt;BR /&gt;and&lt;BR /&gt;drivers/net/dpaa/dpaa_rxtx.c&lt;BR /&gt;There is also this piece of code that I could use some clarification on in&lt;BR /&gt;drivers/net/dpaa2/dpaa2_ethdev.c&lt;BR /&gt;&lt;BR /&gt;&lt;/P&gt;&lt;P&gt;/* Enabling hash results in FD requires setting DPNI_FLCTYPE_HASH in&lt;BR /&gt; * dpni_set_offload API. Setting this FLCTYPE for DPNI sets the FD[SC]&lt;BR /&gt; * to 0 for LS2 in the hardware thus disabling data/annotation&lt;BR /&gt; * stashing. For LX2 this is fixed in hardware and thus hash result and&lt;BR /&gt; * parse results can be received in FD using this option.&lt;BR /&gt; */&lt;BR /&gt; if (dpaa2_svr_family == SVR_LX2160A) {&lt;BR /&gt; ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv-&amp;gt;token,&lt;BR /&gt; DPNI_FLCTYPE_HASH, true);&lt;BR /&gt; if (ret) {&lt;BR /&gt; DPAA2_PMD_ERR("Error setting FLCTYPE: Err = %d", ret);&lt;BR /&gt; return ret;&lt;BR /&gt; }&lt;BR /&gt; }&lt;/P&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
    <pubDate>Fri, 07 Sep 2018 13:06:07 GMT</pubDate>
    <dc:creator>chris_nardone</dc:creator>
    <dc:date>2018-09-07T13:06:07Z</dc:date>
    <item>
      <title>lx2160 dpdk 17.11.2 dpaa2 rte_mbuf hash always zero</title>
      <link>https://community.nxp.com/t5/Processor-Expert-Software/lx2160-dpdk-17-11-2-dpaa2-rte-mbuf-hash-always-zero/m-p/803039#M4180</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;P&gt;I have found while testing with dpdk 17.11.2 on the lx2160 that the hash field of the rte_mbuf is always zero.&lt;/P&gt;&lt;P&gt;It appears from the dpdk source that the value is not set in the dpaa2 driver but is in the dpaa driver.&lt;/P&gt;&lt;P&gt;The files I am referring to are&lt;BR /&gt;drivers/net/dpaa2/dpaa2_rxtx.c&lt;BR /&gt;and&lt;BR /&gt;drivers/net/dpaa/dpaa_rxtx.c&lt;BR /&gt;There is also this piece of code that I could use some clarification on in&lt;BR /&gt;drivers/net/dpaa2/dpaa2_ethdev.c&lt;BR /&gt;&lt;BR /&gt;&lt;/P&gt;&lt;P&gt;/* Enabling hash results in FD requires setting DPNI_FLCTYPE_HASH in&lt;BR /&gt; * dpni_set_offload API. Setting this FLCTYPE for DPNI sets the FD[SC]&lt;BR /&gt; * to 0 for LS2 in the hardware thus disabling data/annotation&lt;BR /&gt; * stashing. For LX2 this is fixed in hardware and thus hash result and&lt;BR /&gt; * parse results can be received in FD using this option.&lt;BR /&gt; */&lt;BR /&gt; if (dpaa2_svr_family == SVR_LX2160A) {&lt;BR /&gt; ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv-&amp;gt;token,&lt;BR /&gt; DPNI_FLCTYPE_HASH, true);&lt;BR /&gt; if (ret) {&lt;BR /&gt; DPAA2_PMD_ERR("Error setting FLCTYPE: Err = %d", ret);&lt;BR /&gt; return ret;&lt;BR /&gt; }&lt;BR /&gt; }&lt;/P&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Fri, 07 Sep 2018 13:06:07 GMT</pubDate>
      <guid>https://community.nxp.com/t5/Processor-Expert-Software/lx2160-dpdk-17-11-2-dpaa2-rte-mbuf-hash-always-zero/m-p/803039#M4180</guid>
      <dc:creator>chris_nardone</dc:creator>
      <dc:date>2018-09-07T13:06:07Z</dc:date>
    </item>
  </channel>
</rss>

