Dave,
----- Original Message -----
You haven't mentioned your kernel version, but dump_timer_data_tvec_bases_v1()
is applicable for much older kernels, certainly older than to 2.6.18.
But it appears that, regardless of the kernel version, that the "timer"
command may not work with more recent kernels that are not configured
with CONFIG_SMP.
Though I have only vmlinux and vmcore, the crash uitlity told me
that my kernel is 2.6.35.7.
Now I understand UP kernel support is not maintained very well.
When I make some progress, I will try to post patches.
Takuo
----- Original Message -----
> Dave,
> There are a few commands that do not work for my ARM vmlinux/vmcore.
>
> One of those is "timer".
> I get the following messages;
>
> crash> timer
> timer: per_cpu_symbol_search(per_cpu__tvec_bases): NULL
> GETBUF(128 -> 0)
> GNU_GET_DATATYPE[tvec_root_s]: returned via gdb_error_hook (1 buffer
> in use)
> FREEBUF(0)
> GETBUF(128 -> 0)
> GNU_GET_DATATYPE[tvec_s]: returned via gdb_error_hook (1 buffer in
> use)
> FREEBUF(0)
> timer: zero-size memory allocation! (called from 81191e6)
> crash>
>
>
> Both vec_root_size and vec_size are 0 at
> dump_timer_data_tvec_bases_v1() in kernel.c
>
> vec_root_size = (i = ARRAY_LENGTH(tvec_root_s_vec)) ?
> i : get_array_length("tvec_root_s.vec", NULL, SIZE(list_head));
> vec_size = (i = ARRAY_LENGTH(tvec_s_vec)) ?
> i : get_array_length("tvec_s.vec", NULL, SIZE(list_head));
> vec = (ulong *)GETBUF(SIZE(list_head) * MAX(vec_root_size, vec_size));
>
> Do you have any suggestion?
You haven't mentioned your kernel version, but dump_timer_data_tvec_bases_v1()
is applicable for much older kernels, certainly older than to 2.6.18.
But it appears that, regardless of the kernel version, that the "timer"
command may not work with more recent kernels that are not configured
with CONFIG_SMP.
Dave
--
Crash-utility mailing list
Crash-utility(a)redhat.com
https://www.redhat.com/mailman/listinfo/crash-utility