On Tue, Jul 27, 2010 at 03:55:28PM +1000, Dave Chinner wrote:
> While XFS passes ranges to operate on from the core code, the
> functions being called ignore the either the entire range or the end
> of the range. This is historical because when the function were
> written linux didn't have the necessary range operations. Update the
> functions to use the correct operations.
Assuming you have actually tested this - given that we've ignore
these parameters so long that I'm really fearing some callers have
started to rely on that behaviour.
> if (mapping->nrpages) {
I'd drop this check ere as well - no other caller does it.
> xfs_iflags_clear(ip, XFS_ITRUNCATED);
> - ret = filemap_write_and_wait(mapping);
> + ret = filemap_write_and_wait_range(mapping, first,
> + last == -1 ? LLONG_MAX : last);
> if (!ret)
> - truncate_inode_pages(mapping, first);
> + truncate_inode_pages_range(mapping, first, last);
> }
> return -ret;
> }
> @@ -73,7 +71,8 @@ xfs_flush_pages(
>
> if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) {
Same for this check.
> xfs_iflags_clear(ip, XFS_ITRUNCATED);
> - ret = -filemap_fdatawrite(mapping);
> + ret = -filemap_fdatawrite_range(mapping, first,
> + last == -1 ? LLONG_MAX : last);
Also for the non-async case we should just use
filemap_write_and_wait_range, and kill off xfs_wait_on_pages.
|