On Wed, 20 Apr 2011, Geert Uytterhoeven wrote:
On Tue, Apr 19, 2011 at 23:29, James Bottomley
<James.Bottomley@xxxxxxxxxxxxxxxxxxxxx> wrote:
Slub makes assumptions about page_to_nid() which are violated by
DISCONTIGMEM and !NUMA. ÂThis violation results in a panic because
page_to_nid() can be non-zero for pages in the discontiguous ranges and
this leads to a null return by get_node(). ÂThe assertion by the
maintainer is that DISCONTIGMEM should only be allowed when NUMA is also
defined. ÂHowever, at least six architectures: alpha, ia64, m32r, m68k,
mips, parisc violate this. ÂThe panic is a regression against slab, so
just mark slub broken in the problem configuration to prevent users
reporting these panics.
How does the problem manifest itself? We're having a problem on m68k, which
seems to go away when switching from SLUB to SLAB, or when reverting a commit
in [2] (probably this was never reported upstream).
The NULL pointer dereference is typical. If SLUB breaks in this way then
other things are also not functioning in the core.
Guess you are also using DISCONTIG in !NUMA?
