[Crash-utility] the kmem get some errors

Dave Anderson anderson at redhat.com
Tue Apr 29 13:05:33 UTC 2008


Yang Zhiguo wrote:
> ----- Original Message ----- 
> From: "Dave Anderson" <anderson at redhat.com>
> To: "Discussion list for crash utility usage,maintenance and development" <crash-utility at redhat.com>
> Sent: Monday, April 28, 2008 8:36 PM
> Subject: Re: [Crash-utility] the kmem get some errors
> 
> 
>> Yang Zhiguo wrote:
>>> hi,
>>>
>>> When i use kmem command on a IA64 with RHEL5.1 as following, there are 
>>> some errors.
>>> Anyone can explain it?
>>>
>>> crash> kmem -c
>>> kmem: cannot determine page cache size
>>> crash> kmem -C
>>> kmem: address_space page cache radix tree not supported
>>> crash> kmem -la
>>> kmem: active_list does not exist in this kernel
>>> crash> kmem -La
>>> kmem: active_list does not exist in this kernel
>>> crash> kmem -li
>>> kmem: inactive_list does not exist in this kernel
>>> crash> kmem -Li
>>> kmem: inactive_list does not exist in this kernel
>> None of the above are applicable to the RHEL5.1 (2.6.18-based) kernel,
>> because the symbols they reference no longer exist.
> 
> If so, maybe we should explain it in man pages.

Well, this is not uncommon -- because the crash utility maintains
backwards compatibility, there are numerous command options that
are no longer applicable, and when they fail, the idea is to
display a concise error message explaining why they fail.

> 
>>> crash> kmem -P e0000001843760c0
>>>   <segmentation violation>
>> Shouldn't happen, but the command makes no sense (e0000001843760c0
>> cannot be a physical address.
> 
> An error message may be more better.
> 

That's true.  On the other arches I've got available, the command
returns: "<address>: physical address not found in mem map".
I don't have an ia64 readily available, so I'm not sure exactly
where it takes the SIGSEGV.

Dave




More information about the Crash-utility mailing list