Re: [PATCH 6/7] vfio: Add vfio_device_io_remap_mapping_range()

From: Christoph Hellwig <hch@infradead.org>
To: Alex Williamson <alex.williamson@redhat.com>
Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
	jgg@nvidia.com, peterx@redhat.com
Subject: Re: [PATCH 6/7] vfio: Add vfio_device_io_remap_mapping_range()
Date: Tue, 10 Aug 2021 11:04:03 +0200	[thread overview]
Message-ID: <YRJBA0A7vIF6PKMB@infradead.org> (raw)
In-Reply-To: <162818329235.1511194.15804833796430403640.stgit@omen>

> +int vfio_device_io_remap_mapping_range(struct vfio_device *device,
> +				       loff_t start, loff_t len)
> +{
> +	struct address_space *mapping = device->inode->i_mapping;
> +	int ret = 0;
> +
> +	i_mmap_lock_write(mapping);
> +	if (mapping_mapped(mapping)) {
> +		struct rb_root_cached *root = &mapping->i_mmap;
> +		pgoff_t pgstart = start >> PAGE_SHIFT;
> +		pgoff_t pgend = (start + len - 1) >> PAGE_SHIFT;
> +		struct vm_area_struct *vma;
> +
> +		vma_interval_tree_foreach(vma, root, pgstart, pgend) {

There is no need for the mapping_mapped check here,
vma_interval_tree_foreach will the right thing for an empty tree.
That also allows to move a few more instructions out of the lock.

> +			/*
> +			 * Force NOFS memory allocation context to avoid
> +			 * deadlock while we hold i_mmap_rwsem.
> +			 */
> +			flags = memalloc_nofs_save();

Please move this out of the loop.

> +extern int vfio_device_io_remap_mapping_range(struct vfio_device *device,
> +					      loff_t start, loff_t len);

No need for the extern.