Jay Lan wrote:
Dave Anderson wrote:
> Ok, so what happens if you comment out these two lines
> in memory.c at line 573:
>
> if (symbol_exists("pgdat_list") && !IS_SPARSEMEM())
> vt->flags |= NODES;
By commenting out these two lines (and previous patch), the crash
did come up.
KERNEL: vmlinuz-2.6.18-kdump
DUMPFILE: vmcore-nmi
CPUS: 2
DATE: Fri Oct 20 11:20:13 2006
UPTIME: 00:11:05
LOAD AVERAGE: 0.12, 0.04, 0.01
TASKS: 67
NODENAME: pogo1
RELEASE: 2.6.18-kdump
VERSION: #40 SMP Wed Oct 18 17:53:19 PDT 2006
MACHINE: ia64 (1400 Mhz)
MEMORY: 1.6 GB
PANIC: ""
PID: 0
COMMAND: "swapper"
TASK: a0000001008f4000 (1 of 4) [THREAD_INFO: a0000001008f5060]
CPU: 0
STATE: TASK_RUNNING
WARNING: panic task not found
crash>
>
> Also, can you do this:
>
> $ gdb vmlinuz-2.6.18-kdump
> ...
> (gdb) ptype struct pglist_data
Here it is:
(gdb) ptype struct pglist_data
type = struct pglist_data {
struct zone node_zones[4];
struct zonelist node_zonelists[5];
int nr_zones;
struct page *node_mem_map;
struct bootmem_data *bdata;
long unsigned int node_start_pfn;
long unsigned int node_present_pages;
long unsigned int node_spanned_pages;
int node_id;
wait_queue_head_t kswapd_wait;
struct task_struct *kswapd;
int kswapd_max_order;
}
(gdb)
Thanks,
- jay
Hi Jay,
I'm still searching for the proper fix without breaking the
other arches and kernel configurations.
What is the output of "help -v"?
Thanks,
Dave