<html>
    <head>
      <base href="https://bugs.dpdk.org/">
    </head>
    <body><span class="vcard"><a class="email" href="mailto:baciumariuscristian@yahoo.com" title="Marius-Cristian Baciu (baciumariuscristian@yahoo.com)"> <span class="fn">Marius-Cristian Baciu</span></a>
</span> changed
          <a class="bz_bug_link 
          bz_status_UNCONFIRMED "
   title="UNCONFIRMED - drivers/common/mlx5: mlx5_malloc() called on invalid socket ID when global MR cache is full and rte_extmem_* API is used"
   href="https://bugs.dpdk.org/show_bug.cgi?id=1256">bug 1256</a>
          <br>
             <table border="1" cellspacing="0" cellpadding="8" class="bz_diff_table">
          <tr>
            <th>What</th>
            <th>Removed</th>
            <th>Added</th>
          </tr>

         <tr>
           <td style="text-align:right;">Resolution</td>
           <td>FIXED
           </td>
           <td>---
           </td>
         </tr>

         <tr>
           <td style="text-align:right;">Status</td>
           <td>RESOLVED
           </td>
           <td>UNCONFIRMED
           </td>
         </tr></table>
      <p>
        <div class="bz_comment_block">
            <b class="bz_comment_head"><a class="bz_bug_link 
          bz_status_UNCONFIRMED "
   title="UNCONFIRMED - drivers/common/mlx5: mlx5_malloc() called on invalid socket ID when global MR cache is full and rte_extmem_* API is used"
   href="https://bugs.dpdk.org/show_bug.cgi?id=1256#c3">Comment # 3</a>
              on <a class="bz_bug_link 
          bz_status_UNCONFIRMED "
   title="UNCONFIRMED - drivers/common/mlx5: mlx5_malloc() called on invalid socket ID when global MR cache is full and rte_extmem_* API is used"
   href="https://bugs.dpdk.org/show_bug.cgi?id=1256">bug 1256</a>
              from <span class="vcard"><a class="email" href="mailto:baciumariuscristian@yahoo.com" title="Marius-Cristian Baciu (baciumariuscristian@yahoo.com)"> <span class="fn">Marius-Cristian Baciu</span></a>
</span>
            </b>
          <pre class="bz_comment_text">Hi,

Unfortunately that patch only targets a memory socket issue with the ASO
mechanism. However, in my setup ASO is never an issue - I actually do not
believe it is enabled.

To give a little more insight, the problem I am describing manifests on the
data path:
- rte_eth_tx_burst();
- mlx5_tx_burst_*() is called;
- at some later point, in mr_lookup_caches(), mr_btree_lookup() returns
UINT32_MAX because all 256 entries in the cache have been occupied and last
memory registration did not catch an empty slot;
- when mr_lookup_caches() fails, mlx5_mr_create() -> mlx5_mr_create_primary()
is called;
- mlx5_malloc() at line 723 fails because it is called with an inappropriate
socket ID (the socket ID of the memseg list associated with an external buffer
(prior with rte_extmem_register()), EXTERNAL_HEAP_MIN_SOCKET_ID, which does not
actually have a valid heap associated, from which memory could be allocated.
          </pre>
        </div>
      </p>


      <hr>
      <span>You are receiving this mail because:</span>

      <ul>
          <li>You are the assignee for the bug.</li>
      </ul>
      <div itemscope itemtype="http://schema.org/EmailMessage">
        <div itemprop="action" itemscope itemtype="http://schema.org/ViewAction">
          
          <link itemprop="url" href="https://bugs.dpdk.org/show_bug.cgi?id=1256#c3">
          <meta itemprop="name" content="View bug">
        </div>
        <meta itemprop="description" content="Bugzilla bug update notification">
      </div>
    </body>
</html>