linux-kernel - Re: [RFC PATCH] Support map

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Sat, 15 Mar 2014 01:32:33 +0200
From:	"Kirill A. Shutemov" <kirill@...temov.name>
To:	Toshi Kani <toshi.kani@...com>
Cc:	willy@...ux.intel.com, kirill.shutemov@...ux.intel.com,
	david@...morbit.com, linux-fsdevel@...r.kernel.org,
	linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: [RFC PATCH] Support map_pages() for DAX

On Fri, Mar 14, 2014 at 05:03:19PM -0600, Toshi Kani wrote:
> +void dax_map_pages(struct vm_area_struct *vma, struct vm_fault *vmf,
> +		get_block_t get_block)
> +{
> +	struct file *file = vma->vm_file;
> +	struct inode *inode = file_inode(file);
> +	struct buffer_head bh;
> +	struct address_space *mapping = file->f_mapping;
> +	unsigned long vaddr = (unsigned long)vmf->virtual_address;
> +	pgoff_t pgoff = vmf->pgoff;
> +	sector_t block;
> +	pgoff_t size;
> +	unsigned long pfn;
> +	pte_t *pte = vmf->pte;
> +	int error;
> +
> +	while (pgoff < vmf->max_pgoff) {
> +		size = (i_size_read(inode) + PAGE_SIZE - 1) >> PAGE_SHIFT;
> +		if (pgoff >= size)
> +			return;
> +
> +		memset(&bh, 0, sizeof(bh));
> +		block = (sector_t)pgoff << (PAGE_SHIFT - inode->i_blkbits);
> +		bh.b_size = PAGE_SIZE;
> +		error = get_block(inode, block, &bh, 0);
> +		if (error || bh.b_size < PAGE_SIZE)
> +			goto next;
> +
> +		if (!buffer_mapped(&bh) || buffer_unwritten(&bh) ||
> +		    buffer_new(&bh))
> +			goto next;
> +
> +		/* Recheck i_size under i_mmap_mutex */
> +		mutex_lock(&mapping->i_mmap_mutex);

NAK. Have you tested this with lockdep enabled?

->map_pages() called with page table lock taken and ->i_mmap_mutex
should be taken before it. It seems we need to take ->i_mmap_mutex in
do_read_fault() before calling ->map_pages().

Side note: I'm sceptical about whole idea to use i_mmap_mutux to protect
against truncate. It will not scale good enough comparing lock_page()
with its granularity.

> +		size = (i_size_read(inode) + PAGE_SIZE - 1) >> PAGE_SHIFT;
> +		if (unlikely(pgoff >= size)) {
> +			mutex_unlock(&mapping->i_mmap_mutex);
> +			return;
> +		}
> +
> +		error = dax_get_pfn(inode, &bh, &pfn);
> +		if (error > 0)
> +			dax_set_pte(vma, vaddr, pfn, pte);
> +
> +		mutex_unlock(&mapping->i_mmap_mutex);
> +next:
> +		vaddr += PAGE_SIZE;
> +		pgoff++;
> +		pte++;
> +	}
> +}
> +EXPORT_SYMBOL_GPL(dax_map_pages);

-- 
 Kirill A. Shutemov
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/