[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] xm dump-core and an XP HVM


  • To: "Akio Takebe" <takebe_akio@xxxxxxxxxxxxxx>
  • From: "Ryan Persaud" <ryan.persaud@xxxxxxxxx>
  • Date: Fri, 29 Feb 2008 15:42:03 -0500
  • Cc: xen-users@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Fri, 29 Feb 2008 12:42:40 -0800
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:references; b=av2iXp10pgGft4vE/S0rPOokIa2OEWFA4JowfMc7bJikfgpmdqU+S7cIXGfMFxzr8LUhe7mxzfArY/LGSLnNJ25T4PAoMRlI6wXg8utMTiQLMhEKYzSn57zDKHRv/1GCemvDcp2NmTQJM5DLeqN5w1syiX1KJAR8oUh3Ekn+O8c=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

Akio,

That worked.  What causes the discrepancy between the value for nr_pages and max_memkb/4?  Also, if I wanted to add the code as a function named xc_dump_physical() in xc_core.c and be able to call it from main.py, what intermediate files are involved?  I've tried to trace a call to dump-core to figure out the files that are involved, but I can only trace it to the call to server.xend.dump.core in main.py (I assume at this point it becomes an xml-rpc call?).  Basically, I need the chain of files from main.py to xc_core.c involved in an invocation of dump-core.

Thanks again for your time,
-Ryan


On Thu, Feb 28, 2008 at 12:14 AM, Akio Takebe <takebe_akio@xxxxxxxxxxxxxx> wrote:
Hi,

You'd better use max_memkb if you don't make p2m table(like the .xen_pfn).
e.g.
"info.max_memkb >> (PAGE_SHIFT - 10)"

>max_pfn = nr_pages, or would they be different for some reason?  I'm
>planning on adding a dump-physical tool, and I'll post the code to xen-devel
>once I get it working.
Great. It will be useful. Please CC to me at posting it into xen-devel ML. ;)

Best Regards,

Akio Takebe

>Akio,
>
>Thanks for the reply.  How would I obtain max_pfn?  In
>xc_domain_dumpcore_via_callback, the code pulls nr_pages from the xc_dominfo
>struct and uses this value as the number of pfns for the domain.  Would
>max_pfn = nr_pages, or would they be different for some reason?  I'm
>planning on adding a dump-physical tool, and I'll post the code to xen-devel
>once I get it working.
>
>-Ryan
>
>
>
>On Wed, Feb 27, 2008 at 5:37 AM, Akio Takebe <takebe_akio@xxxxxxxxxxxxxx>
>wrote:
>
>> Hi, Ryan
>>
>> Will you try to make a convert tools (from raw memory dump to
>> Windows fomat dump)? If so, please post it to xen-devel ML.
>>
>> You can see docs/misc/dump-core-format.txt.
>> ".xen_pages" section is raw memory,
>> and .xen_p2m or .xen_pfn sections are used
>> for looking for a memory address we want to see.
>>
>> You can see them with readelf -x option.
>>
>> If you want to make a new tools to get a straight dump of physical memory,
>> it would just do xc_map_foreign_range() from 0 to max_pfn.
>>
>> Best Regards,
>>
>> Akio Takebe
>>
>> >Hello,
>> >
>> >This may be more of a developer question.  If so, please let me know, and
>> I
>> >will ask it there instead.  I am running Xen 3.1 with a Windows XP HVM,
>> and
>> >I want to get a straight dump of physical memory.  The dump-core command
>> >adds a lot of ELF cruft, so I commented out the ELF related portions of
>> >xc_domain_dumpcore_via_callback.  However, when I run dump-core, I am
>> still
>> >getting quite a few more pages in my dump file than I should (nearly 10MB
>> >worth).  Has anyone else attempted to do this, or have any idea of what
>> >might be the issue?  I can understand the utility of having the dump file
>> >conform to ELF format, but it's also useful to get just the raw memory
>> since
>> >there are tools out there that can take a Windows physical dump and
>> analyze
>> >it.
>> >
>> >Thanks for the help,
>> >-Ryan
>> >
>> >-------------------------------text/plain-------------------------------
>> >_______________________________________________
>> >Xen-users mailing list
>> >Xen-users@xxxxxxxxxxxxxxxxxxx
>> >http://lists.xensource.com/xen-users
>>
>>
>
>
>---html-part included links-------
>mailto:takebe_akio@xxxxxxxxxxxxxx
>mailto:Xen-users@xxxxxxxxxxxxxxxxxxx


_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.