Olaf,
        I was looking in 2.6.16-rc5-git8, but did not find the patch that you mentioned.
 

arch/powerpc/kernel/setup_64.c:

#ifdef CONFIG_SMP
void __init setup_per_cpu_areas(void)
{
        int i;
        unsigned long size;
        char *ptr;

        /* Copy section for each CPU (we discard the original) */
        size = ALIGN(__per_cpu_end - __per_cpu_start, SMP_CACHE_BYTES);
#ifdef CONFIG_MODULES
        if (size < PERCPU_ENOUGH_ROOM)
                size = PERCPU_ENOUGH_ROOM;
#endif

        for_each_cpu(i) {
                ptr = alloc_bootmem_node(NODE_DATA(cpu_to_node(i)), size);
                if (!ptr)
                        panic("Cannot allocate cpu data for CPU %d\n", i);

                paca[i].data_offset = ptr - __per_cpu_start;
                memcpy(ptr, __per_cpu_start, __per_cpu_end - __per_cpu_start);
     }
}
 

 Instead of depending on paca[#].data_offset, my fix will use cpu_online_map to read the valid cpu paca structs. But, if the kernel sets 0xeeeeeeeeeeeeeeeeULL, yes, using this value will be the better solution.

Dave, the attached patch contains both fixes that Rachita reported.
        - Report an error message if the thread is running in user space and print regs.
        - Finding the valid per_cpu_offset values

Thanks
Haren
 

Haren,

Is this right?  NR_CPUS is hardcoded (in crash) to 32 for ppc64,
and BITS_FOR_LONG is 64, making it "ulong cpu_online_map[0]":

@@ -2282,21 +2298,22 @@ parse_cmdline_arg(void)
 static void
 ppc64_paca_init(void)
 {
+#define BITS_FOR_LONG sizeof(ulong)*8
        int i, cpus, nr_paca;
        char *cpu_paca_buf;
        ulong data_offset;
-       ulong per_cpu_offset;
-       ulong cpuid_offset;
-       signed short cpuid;
+       ulong cpu_online_map[NR_CPUS/BITS_FOR_LONG];

Or am I missing something obvious?

Also, I get a little nervous about depending upon the NR_CPUS
value compiled into crash being the same as the configured
NR_CPUS in the kernel.  As far as crash is concerned, NR_CPUS
means the "largest possible" cpu configuration that this version
of *crash* will support; it's not meant to be aligned with
the kernel's value.

But if you guys are happy with that, it's fine with me.

Dave