vmcore-dmesg ends with "No program header covering vaddr"

[Prasad Koya]

Thanks for replying Baoquan.

Our production kernel is 3.4 and it has legacy dmesg buffer. I was
able to dump it from oldmem. If this change is of any use for
community I could try submit the patch.

On Fri, Apr 4, 2014 at 4:14 AM, Baoquan He <bhe at redhat.com> wrote:
> On 03/14/14 at 10:39pm, Prasad Koya wrote:
>> Hi
>>
>> I'm increasing dmesg buffer size at runtime by passing log_buf_len=1M
>> (or 512K) at command line. At compile time, I have LOG_BUF_LEN as
>> 256K. However, when this kernel crashes (crashkernel set at 64M),
>> vmcore-dmesg fails with below error.
>>
>> No program header covering vaddr 0xffff88013ff00000found kexec bug?
>>
>> Am I missing something here?
>>
>> Thank you.
>>
>> Here are some debugs I added to kernel. This is with 3.4 kernel and
>> I've tried with kexec-tools-2.0.4.
>>
>> [    0.000000] log_buf_len_setup: log_buf_len 262144 new 1048576
>> ffffffff8170b140
>> [    0.000000] setup_log_buf: early 1 262144 new_log_buf_len 1048576
>> [    0.000000] log_buf_len: 1048576 ffff880037d58000
>> [    0.000000] early log buf free: 258671(98%)
>> [    0.000000] setup_log_buf: early 0 1048576 new_log_buf_len 0
>> [    0.503752] log_buf_kexec_setup: lbl 1048576 new 0 ffffffff816198c4
>> logbuf ffff880037d58000
>> In kdump, running vmcore_dmesg /proc/vmcore
>> No program header covering vaddr 0xffff88013ff00000found kexec bug?
>
> Hi Prasad,
>
> Did you try the latest kernel and vmcore-dmesg? Mine is OK. Surely you
> need apply below patch since it's not merged yet.
>
> Author: Lubomir Rintel <lkundrak at v3.sk>
> Date:   Wed Sep 18 14:21:43 2013 +0200
>
>     vmcore-dmesg: Understand >= v3.11-rc4 dmesg
>
>     Symbol name changed with the following commit:
>     62e32ac printk: rename struct log to struct printk_log
>
> Thanks
> Baoquan
>
>>
>> _______________________________________________
>> kexec mailing list
>> kexec at lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/kexec