Re: [RFC][PATCH] problem of cont_prepare_write()

[Andrew Morton <akpm@osdl.org>]

OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> wrote:
>
> 
>  		status = __block_prepare_write(inode, new_page, zerofrom,
>  						PAGE_CACHE_SIZE, get_block);
>  		if (status)
>  			goto out_unmap;
>  		kaddr = kmap_atomic(new_page, KM_USER0);
>  		memset(kaddr+zerofrom, 0, PAGE_CACHE_SIZE-zerofrom);
>  		flush_dcache_page(new_page);
>  		kunmap_atomic(kaddr, KM_USER0);
>  		__block_commit_write(inode, new_page,
>  				zerofrom, PAGE_CACHE_SIZE);
>  		unlock_page(new_page);
>  		page_cache_release(new_page);
>  	}
> 
>  But until ->commit_write(), kernel doesn't update the ->i_size. Then,
>  if kernel writes out that hole page before updates of ->i_size, dirty
>  flag of buffer_head is cleared in __block_write_full_page(). So hole
>  page was not writed to disk.

Oh I see.  After the above page is unlocked, it's temporarily outside
i_size.

Perhaps cont_prepare_write() should look to see if the zerofilled page is
outside the current i_size and if so, advance i_size to the end of the
zerofilled page prior to releasing the page lock.

We might need to run mark_inode_dirty() at some stage, or perhaps just rely
on the caller doing that in ->commit_write().