<?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>topic Re: HP11.31 Kernel Memory Leak problem in Operating System - HP-UX</title>
    <link>https://community.hpe.com/t5/operating-system-hp-ux/hp11-31-kernel-memory-leak-problem/m-p/4120867#M738526</link>
    <description>Puhut,&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt; I checked 11.31 memory leak found The formal patch is not yet announcend&lt;BR /&gt;JAGag35944&lt;BR /&gt;&lt;BR /&gt;Please log case to HP support and give them the JAGag35944 to request for temporary fix.&lt;BR /&gt;&lt;BR /&gt;WK&lt;BR /&gt;</description>
    <pubDate>Wed, 26 Dec 2007 03:04:07 GMT</pubDate>
    <dc:creator>whiteknight</dc:creator>
    <dc:date>2007-12-26T03:04:07Z</dc:date>
    <item>
      <title>HP11.31 Kernel Memory Leak problem</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/hp11-31-kernel-memory-leak-problem/m-p/4120866#M738525</link>
      <description>Hi all,&lt;BR /&gt;I'm trying to get process credential in kernel module using proc_cred_hold() function. If I use proc_cred_hold() function, PM_CRED area of kernel memory is increased continuously and system hanging because of insufficient kernel memory. I think proc_cred_hold() leads to kernel memory leak.&lt;BR /&gt;&lt;BR /&gt;Physical memory       =  2088549    8.0g 100%&lt;BR /&gt;Free memory           =     3200   12.5m   0%&lt;BR /&gt;User processes        =   174884  683.1m   8%  details with -user&lt;BR /&gt;System                =  1852206    7.1g  89%&lt;BR /&gt;  Kernel              =  1852154    7.1g  89%  kernel text and data&lt;BR /&gt;    Dynamic Arenas    =  1671854    6.4g  80%  details with -arena&lt;BR /&gt;      PM_CRED         =  1568969    6.0g  75%&lt;BR /&gt;      vx_global_kmcac =    16237   63.4m   1%&lt;BR /&gt;      spinlock_arena  =    10034   39.2m   0%&lt;BR /&gt;      vm_pfn2v_arena  =     8518   33.3m   0%&lt;BR /&gt;      M_DYNAMIC       =     6125   23.9m   0%&lt;BR /&gt;&lt;BR /&gt;Any thoughts on what I'm missing to use proc_cred_hold() function? I have to do other things to use the function?&lt;BR /&gt;&lt;BR /&gt;Thanks,</description>
      <pubDate>Wed, 26 Dec 2007 01:57:03 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/hp11-31-kernel-memory-leak-problem/m-p/4120866#M738525</guid>
      <dc:creator>puhut</dc:creator>
      <dc:date>2007-12-26T01:57:03Z</dc:date>
    </item>
    <item>
      <title>Re: HP11.31 Kernel Memory Leak problem</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/hp11-31-kernel-memory-leak-problem/m-p/4120867#M738526</link>
      <description>Puhut,&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt; I checked 11.31 memory leak found The formal patch is not yet announcend&lt;BR /&gt;JAGag35944&lt;BR /&gt;&lt;BR /&gt;Please log case to HP support and give them the JAGag35944 to request for temporary fix.&lt;BR /&gt;&lt;BR /&gt;WK&lt;BR /&gt;</description>
      <pubDate>Wed, 26 Dec 2007 03:04:07 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/hp11-31-kernel-memory-leak-problem/m-p/4120867#M738526</guid>
      <dc:creator>whiteknight</dc:creator>
      <dc:date>2007-12-26T03:04:07Z</dc:date>
    </item>
    <item>
      <title>Re: HP11.31 Kernel Memory Leak problem</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/hp11-31-kernel-memory-leak-problem/m-p/4120868#M738527</link>
      <description>That suggests one of two things to me:&lt;BR /&gt;&lt;BR /&gt;1) You're putting holds on a massive amount of credentials at once (from what I can tell, these get held in a caching layer within this code and recycled as needed... but never freed back to the Arena level) resulting in a bunch of credentials in use at one time beyond the capacity of the system to handle.&lt;BR /&gt;&lt;BR /&gt;or &lt;BR /&gt;&lt;BR /&gt;2) You aren't calling crfree() with the credential you got from proc_cred_hold() when you're done with it (or you think you're never done with it). With the hold in place, the credential can't be returned to the cache, and hence new credentials are always allocated instead of the recycling.&lt;BR /&gt;&lt;BR /&gt;i.e. you should have:&lt;BR /&gt;&lt;BR /&gt;cred_t *temp_cred = NULL;&lt;BR /&gt;&lt;BR /&gt;temp_cred = proc_cred_hold(proc_ptr);&lt;BR /&gt;if(temp_cred != NULL ) {&lt;BR /&gt;// Do stuff&lt;BR /&gt;        crfree(temp_cred);&lt;BR /&gt;}&lt;BR /&gt;&lt;BR /&gt;If you are using crfree(), contact your support representative and get them your dump so they can file an issue and get folks involved... it isn't a known issue.&lt;BR /&gt;&lt;BR /&gt;Don&lt;BR /&gt;&lt;BR /&gt;PS -- whiteknight -- JAGag35944 does have a memory leak fix (for 11.11, 11.23 and 11.31) but this is a completely and totally unrelated issue [networking code instead of PM, completely different arena, nothing to do with credentials in any form --- just packet handling issues]. I seriously doubt acquiring this patch in an early form would be meaningful in addressing the problem.</description>
      <pubDate>Wed, 26 Dec 2007 04:41:40 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/hp11-31-kernel-memory-leak-problem/m-p/4120868#M738527</guid>
      <dc:creator>Don Morris_1</dc:creator>
      <dc:date>2007-12-26T04:41:40Z</dc:date>
    </item>
  </channel>
</rss>

