[CRIU] [RFC 0/2] Proc parsing update and slab for shared memory

[Cyrill Gorcunov]

When we do restore we use shared memory slabs pretty heavy,
so instead of tracking them open-coded i think introduce own
slab facility might be worth.

This is rfc siply because there is no slab users out of
my repo (I use it for inotify restore procedure, which
I'm still working on) but I would like to hear what
you guys think about such approach in general.

Note -- the 'slab' has nothing to to with slab concept
used in kernel, this slab only allocates memory checking
for overflow and align pointers provided back but never
release objects, the whole slab can be released only.