Re: [PATCH 0/6] freepgt: free_pgtables shakeup

[Nick Piggin <nickpiggin@yahoo.com.au>]

Russell King wrote:

> Ok.  What's happening is that the ARM pgd_alloc uses pte_alloc_map() to
> allocate the first L1 page table.  This sets mm->nr_ptes to be one.
> 
> The ARM free_pgd knows about this, and will free this PTE at the
> appropriate time.  However, exit_mmap() doesn't know about this itself,
> so in the ARM case, it should BUG_ON(mm->nr_ptes != 1) if we're using
> low vectors.
> 
> I guess we could hack it such that the ARM pgd_alloc decrements mm->nr_ptes
> itself to keep things balanced, since free_pte_range() won't be called
> for this pte.  However, I don't like that because its likely to break at
> some point in the future.
> 
> Any ideas how to cleanly support this with the new infrastructure?
> 

Hmm, in that case it could be just a problem with that BUG_ON() -
it wasn't there before... but it seems like a very useful test,
especially with all this new work going on, so it would be a shame
not to run it in releases.

But I don't quite understand (should really look at the code more),
how come you aren't leaking memory? Do _all_ processes share this
same first L1 page table? If so, can it be allocated with a more
specialised function? If not, can nr_ptes be decremented in the
place where it is freed?

/me goes to bed now - I'll have a bit of a look tomorrow.

Nick