On Wed, Mar 04, 2020 at 07:47:06AM -0800, Matthew Wilcox wrote:
> From: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
>
> The 'pgoff' displayed by the tracepoints wasn't a pgoff at all; it
> was a byte offset from the start of the file. We already emit that in
> the form of the 'offset', so we can just remove pgoff. That means we
> can remove 'page' as an argument to the tracepoint, and rename this
> type of tracepoint from being a page class to being a range class.
>
> Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
>
> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index 7057ef155a29..cab29ffb2b40 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -487,7 +487,7 @@ EXPORT_SYMBOL_GPL(iomap_is_partially_uptodate);
> int
> iomap_releasepage(struct page *page, gfp_t gfp_mask)
> {
> - trace_iomap_releasepage(page->mapping->host, page, 0, 0);
> + trace_iomap_releasepage(page->mapping->host, 0, 0);
I think we should pass page_offset() for the offset here now. Maybe
also PAGE_SIZE len for completeness while we're at it.
> @@ -1503,7 +1503,7 @@ iomap_do_writepage(struct page *page, struct writeback_control *wbc, void *data)
> u64 end_offset;
> loff_t offset;
>
> - trace_iomap_writepage(inode, page, 0, 0);
> + trace_iomap_writepage(inode, 0, 0);
Same here.
