linux-kernel - Re: [RFC PATCH V3] iomap: add support to track dirty state of sub pages

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:   Fri, 21 Aug 2020 07:17:52 +0100
From:   Christoph Hellwig <hch@...radead.org>
To:     Yu Kuai <yukuai3@...wei.com>
Cc:     hch@...radead.org, darrick.wong@...cle.com, willy@...radead.org,
        david@...morbit.com, linux-xfs@...r.kernel.org,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
        yi.zhang@...wei.com
Subject: Re: [RFC PATCH V3] iomap: add support to track dirty state of sub
 pages

> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index bcfc288dba3f..b6a7457d8581 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -21,15 +21,20 @@
>  
>  #include "../internal.h"
>  
> +#define DIRTY_BITS(x)	((x) + PAGE_SIZE / SECTOR_SIZE)
>  /*

Nit: please keep an empty line between a definition and a comment.

> +	 * The first half bits are used to track sub-page uptodate status,
> +	 * the second half bits are for dirty status.
> +	 */
> +	DECLARE_BITMAP(state, PAGE_SIZE * 2 / SECTOR_SIZE);
>  };
>  
>  static inline struct iomap_page *to_iomap_page(struct page *page)
> @@ -52,8 +57,8 @@ iomap_page_create(struct inode *inode, struct page *page)
>  	iop = kmalloc(sizeof(*iop), GFP_NOFS | __GFP_NOFAIL);
>  	atomic_set(&iop->read_count, 0);
>  	atomic_set(&iop->write_count, 0);
> -	spin_lock_init(&iop->uptodate_lock);
> -	bitmap_zero(iop->uptodate, PAGE_SIZE / SECTOR_SIZE);
> +	spin_lock_init(&iop->state_lock);
> +	bitmap_zero(iop->state, PAGE_SIZE * 2 / SECTOR_SIZE);

Maybe add a

#define IOMAP_STATE_ARRAY_SIZE	(PAGE_SIZE * 2 / SECTOR_SIZE)

and use?  That isn't much shorter, but a little easier to read at least.

> +	spin_lock_irqsave(&iop->state_lock, flags);
> +	for (i = first; i <= last; i++)
> +		set_bit(i, iop->state);

I think Matthew had some patches to move the these days pointlessly
atomic bitops to use the bitmap_* routines.  It might make sense to
start out that way for new code as well.

> +
> +	if (last >= first)
> +		iomap_set_page_dirty(page);
> +
> +	spin_unlock_irqrestore(&iop->state_lock, flags);

As aready pointed out, this probably needs to move out of the lock.

> +static void
> +iomap_set_range_dirty(struct page *page, unsigned int off,
> +		unsigned int len)
> +{
> +	if (PageError(page))
> +		return;
> +
> +	if (page_has_private(page))
> +		iomap_iop_set_range_dirty(page, off, len);

I'd be tempted to merge this function and iomap_iop_set_range_dirty,
and just return early if there is an error or no iomap_page structure,
relying on the fact that to_iomap_page returns NULL for that case.

> +static void
> +iomap_iop_clear_range_dirty(struct page *page, unsigned int off,
> +		unsigned int len)
> +{
> +	struct iomap_page *iop = to_iomap_page(page);
> +	struct inode *inode = page->mapping->host;
> +	unsigned int first = DIRTY_BITS(off >> inode->i_blkbits);
> +	unsigned int last = DIRTY_BITS((off + len - 1) >> inode->i_blkbits);
> +	unsigned long flags;
> +	unsigned int i;
> +
> +	spin_lock_irqsave(&iop->state_lock, flags);
> +	for (i = first; i <= last; i++)
> +		clear_bit(i, iop->state);

should probably use bitmap_clear().

> +static void
> +iomap_clear_range_dirty(struct page *page, unsigned int off,
> +		unsigned int len)
> +{
> +	if (PageError(page))
> +		return;
> +
> +	if (page_has_private(page))
> +		iomap_iop_clear_range_dirty(page, off, len);
> +}

Same comment about merging the two functions as above.