[Crash-utility] Test patch to fix kmem -s bug

[Dave Anderson]

Hi Sharyathi,

With the "limit/readmem()" fix to max_cpudata_limit(), an additional
check in gather_cpudata_list_v2_nodes() to make sure that the
multiple-index list doesn't get overflowed, and a few cosmetic mods,
this patch looks good to go.  (No further patch is required.)

Nice work -- this part of the crash code has always been a huge
PITA to keep functional...

We're overdue for a release -- but it would be nice if you can
come up with a reason for your arithmetic bug in "kmem -i".
Like I said, it works OK for me on x86 and x86_64.  Let me know
what you find.

And again, multiple thanks for undertaking this "kmem -s" effort.

Dave