[Crash-utility] Problem with "xm save" x86-64 cores -crash.4.0-3.5

[Dave Anderson]

Dave Anderson wrote:

> Tejasvi Aswathanarayana wrote:
>
>>
>>
>> Yes,  the PIDTYPE_MAX enum is equal  to 4.  Setting 'plen' to 4 seems to work.
>> So, it seems like a problem with get_array_length.
>>
>> I will get you the dump/vmlinux files as soon as I can upload it
>> someplace (yousendit is acting up).
>>
>>
>
> Thanks -- the download is underway (*very* slowly)...
>
> In the meantime, can you remove the "plen = 4" hack,
> and then do the following:
>
>   $ crash -d7 vmlinux-2.6.16.13-xen test.core
>
> At the very end, just prior to the crash session ending, you should
> see the gdb request for pid_hash, which should show something
> like this:
>
> GDB IN: command: 15 (GNU_GET_SYMBOL_TYPE)
> buf: 0
> fp: 31d072d8c0 flags: 8  (GNU_RETURN_ON_ERROR)
> addr: 0 addr2: 0 count: 0
> name: "pid_hash" length: 0 typecode: 0
> typename: (null)
> target_typename: (null)
> target_length: 0 target_typecode: 0 is_typedef: 0 member: "(null)" member_offset: 0
> member_length: 0
> member_typecode: 0
> value: 0 tagname: "(null)" pc: 0
> sp: 0 ra: 0 frame: 0 prevsp: 0
> prevpc: 0 lastsp: 0 task: 0 debug: 0
>
> GDB OUT: command: 15 (GNU_GET_SYMBOL_TYPE)
> buf: 0
> fp: 31d072d8c0 flags: 8  (GNU_RETURN_ON_ERROR)
> addr: 0 addr2: 0 count: 0
> name: "pid_hash" length: 32 typecode: 2
> typename: (null)
> target_typename: (null)
> target_length: 8 target_typecode: 1 is_typedef: 0 member: "(null)" member_offset: 0
> member_length: 0
> member_typecode: 0
> value: 0 tagname: "(null)" pc: 0
> sp: 0 ra: 0 frame: 0 prevsp: 0
> prevpc: 0 lastsp: 0 task: 0 debug: 0
>
> What do you see when you do that?
>
> Thanks,
>   Dave
>

Never mind -- I've got the dump.

I'll let you know what I find out...

Thanks again,
  Dave

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/crash-utility/attachments/20061004/9b9a8670/attachment.htm>