Re: [Tails-dev] Please review feature/hugetlb_mem_wipe

anonym wrote (30 Nov 2012 13:41:21 GMT) :
> I didn't experience the "wipe_page: Cannot allocate memory" or
> "spawn_new failed (1)" errors in either test.

Good news. I'm glad the bugs were fixed.

> Well, while this branch makes the wipe much faster and better looking,
> it's not as efficient as Tails' current parallel sdmem approach when
> using a PAE kernel (which usually results in 0 occurrences in my tests,
> if not it's just a few hundred occurences), which arguably is what most
> users will use. For the non-PAE kernel I believe this branch is better,
> though.

> Are we sure we want this?

OK, not in the current state of things. I summed up the current state
of things in a new ticket so that it's not forgotten again:
todo/hugetlb_mem_wipe. If things don't move forward on this side until
0.16 is out, we should not merge the branch again in experimental when
we reset it post-0.16.

I must say I don't feel competent to finish this work... but well, now
that we ship a PAE kernel and a working implementation for it, the
problem is actually limited to non-PAE systems with big amounts of
memory as I understand it, so it's quite less of an emergency than it
used to be.