Kernel panic on divide-by-zero in get_dirty_limits

Issue

• Kernel panicked on divide-by-zero error in get_dirty_limits. Call trace reads similar to:

divide error: 0000 [#1] SMP 
[...]
RIP  [<ffffffff81125211>] get_dirty_limits+0x1d1/0x2f0
 RSP <ffff880187c83a08>

or

 #5 [ffff880c1cd67950] do_divide_error at ffffffff8100cfff
 #6 [ffff880c1cd679f0] divide_error at ffffffff8100be7b
    [exception RIP: get_dirty_limits+193]
 #7 [ffff880c1cd67b00] balance_dirty_pages at ffffffff8112568f
 #8 [ffff880c1cd67c20] balance_dirty_pages_ratelimited_nr at ffffffff81125a64
 #9 [ffff880c1cd67c30] do_wp_page at ffffffff8113ae2b