[ofa-general] Re: [PATCH 1 of 9] Lock the entire mm to prevent any mmu related operation to happen

[Rusty Russell]

On Wednesday 09 April 2008 01:44:04 Andrea Arcangeli wrote:
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1050,6 +1050,15 @@
>  				   unsigned long addr, unsigned long len,
>  				   unsigned long flags, struct page **pages);
>
> +struct mm_lock_data {
> +	spinlock_t **i_mmap_locks;
> +	spinlock_t **anon_vma_locks;
> +	unsigned long nr_i_mmap_locks;
> +	unsigned long nr_anon_vma_locks;
> +};
> +extern struct mm_lock_data *mm_lock(struct mm_struct * mm);
> +extern void mm_unlock(struct mm_struct *mm, struct mm_lock_data *data);

As far as I can tell you don't actually need to expose this struct at all?

> +		data->i_mmap_locks = vmalloc(nr_i_mmap_locks *
> +					     sizeof(spinlock_t));

This is why non-typesafe allocators suck.  You want 'sizeof(spinlock_t *)' 
here.

> +		data->anon_vma_locks = vmalloc(nr_anon_vma_locks *
> +					       sizeof(spinlock_t));

and here.

> +	err = -EINTR;
> +	i_mmap_lock_last = NULL;
> +	nr_i_mmap_locks = 0;
> +	for (;;) {
> +		spinlock_t *i_mmap_lock = (spinlock_t *) -1UL;
> +		for (vma = mm->mmap; vma; vma = vma->vm_next) {
...
> +		data->i_mmap_locks[nr_i_mmap_locks++] = i_mmap_lock;
> +	}
> +	data->nr_i_mmap_locks = nr_i_mmap_locks;

How about you track your running counter in data->nr_i_mmap_locks, leave 
nr_i_mmap_locks alone, and BUG_ON(data->nr_i_mmap_locks != nr_i_mmap_locks)?

Even nicer would be to wrap this in a "get_sorted_mmap_locks()" function.

Similarly for anon_vma locks.

Unfortunately, I just don't think we can fail locking like this.  In your next 
patch unregistering a notifier can fail because of it: that not usable.

I think it means you need to add a linked list element to the vma for the 
CONFIG_MMU_NOTIFIER case.  Or track the max number of vmas for any mm, and 
keep a pool to handle mm_lock for this number (ie. if you can't enlarge the 
pool, fail the vma allocation).  

Both have their problems though...
Rusty.