Re: [Qemu-devel] [PATCH 19/21] userfaultfd: remap_pages: UFFDIO_REMAP preparation

[Pavel Emelyanov <xemul@parallels.com>]

> +ssize_t remap_pages(struct mm_struct *dst_mm, struct mm_struct *src_mm,
> +		    unsigned long dst_start, unsigned long src_start,
> +		    unsigned long len, __u64 mode)
> +{
> +	struct vm_area_struct *src_vma, *dst_vma;
> +	long err = -EINVAL;
> +	pmd_t *src_pmd, *dst_pmd;
> +	pte_t *src_pte, *dst_pte;
> +	spinlock_t *dst_ptl, *src_ptl;
> +	unsigned long src_addr, dst_addr;
> +	int thp_aligned = -1;
> +	ssize_t moved = 0;
> +
> +	/*
> +	 * Sanitize the command parameters:
> +	 */
> +	BUG_ON(src_start & ~PAGE_MASK);
> +	BUG_ON(dst_start & ~PAGE_MASK);
> +	BUG_ON(len & ~PAGE_MASK);
> +
> +	/* Does the address range wrap, or is the span zero-sized? */
> +	BUG_ON(src_start + len <= src_start);
> +	BUG_ON(dst_start + len <= dst_start);
> +
> +	/*
> +	 * Because these are read sempahores there's no risk of lock
> +	 * inversion.
> +	 */
> +	down_read(&dst_mm->mmap_sem);
> +	if (dst_mm != src_mm)
> +		down_read(&src_mm->mmap_sem);
> +
> +	/*
> +	 * Make sure the vma is not shared, that the src and dst remap
> +	 * ranges are both valid and fully within a single existing
> +	 * vma.
> +	 */
> +	src_vma = find_vma(src_mm, src_start);
> +	if (!src_vma || (src_vma->vm_flags & VM_SHARED))
> +		goto out;
> +	if (src_start < src_vma->vm_start ||
> +	    src_start + len > src_vma->vm_end)
> +		goto out;
> +
> +	dst_vma = find_vma(dst_mm, dst_start);
> +	if (!dst_vma || (dst_vma->vm_flags & VM_SHARED))
> +		goto out;

I again have a concern about the case when one task monitors the VM of the
other one. If the target task (owning the mm) unmaps a VMA then the monitor
task (holding and operating on the ufd) will get plain EINVAL on UFFDIO_REMAP
request. This is not fatal, but still inconvenient as it will be hard to
find out the reason for failure -- dst VMA is removed and the monitor should
just drop the respective pages with data, or some other error has occurred
and some other actions should be taken.

Thanks,
Pavel