[PATCH] Assert that pages being written back are locked

[PATCH] Assert that pages being written back are locked

[Matthew Wilcox]

From: Matthew Wilcox (Oracle) <willy@infradead.org>

In write_cache_pages(), we rely on PageLock to prevent writeback from
starting on locked pages.  I'm not sure if we document anywhere that pages
that we're starting writes on must be locked, but having an assertion
will make it clear to users that this is required.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>

diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index 3671568d433f..f83dd855594d 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -2774,6 +2774,8 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
 	struct address_space *mapping = page_mapping(page);
 	int ret, access_ret;
 
+	VM_BUG_ON_PGFLAGS(!PageLocked(page), page);
+
 	lock_page_memcg(page);
 	if (mapping && mapping_use_writeback_tags(mapping)) {
 		XA_STATE(xas, &mapping->i_pages, page_index(page));

Re: [PATCH] Assert that pages being written back are locked

[Matthew Wilcox]

Anyone want to review this for me?

On Tue, Oct 27, 2020 at 10:46:45PM +0000, Matthew Wilcox wrote:
> From: Matthew Wilcox (Oracle) <willy@infradead.org>
> 
> In write_cache_pages(), we rely on PageLock to prevent writeback from
> starting on locked pages.  I'm not sure if we document anywhere that pages
> that we're starting writes on must be locked, but having an assertion
> will make it clear to users that this is required.
> 
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> 
> diff --git a/mm/page-writeback.c b/mm/page-writeback.c
> index 3671568d433f..f83dd855594d 100644
> --- a/mm/page-writeback.c
> +++ b/mm/page-writeback.c
> @@ -2774,6 +2774,8 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
>  	struct address_space *mapping = page_mapping(page);
>  	int ret, access_ret;
>  
> +	VM_BUG_ON_PGFLAGS(!PageLocked(page), page);
> +
>  	lock_page_memcg(page);
>  	if (mapping && mapping_use_writeback_tags(mapping)) {
>  		XA_STATE(xas, &mapping->i_pages, page_index(page));
> 

Re: [PATCH] Assert that pages being written back are locked

[William Kucharski]

Looks good.

Reviewed-by: William Kucharski <william.kucharski@oracle.com>

> On Oct 27, 2020, at 4:46 PM, Matthew Wilcox <willy@infradead.org> wrote:
> 
> From: Matthew Wilcox (Oracle) <willy@infradead.org>
> 
> In write_cache_pages(), we rely on PageLock to prevent writeback from
> starting on locked pages.  I'm not sure if we document anywhere that pages
> that we're starting writes on must be locked, but having an assertion
> will make it clear to users that this is required.
> 
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> 
> diff --git a/mm/page-writeback.c b/mm/page-writeback.c
> index 3671568d433f..f83dd855594d 100644
> --- a/mm/page-writeback.c
> +++ b/mm/page-writeback.c
> @@ -2774,6 +2774,8 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
> 	struct address_space *mapping = page_mapping(page);
> 	int ret, access_ret;
> 
> +	VM_BUG_ON_PGFLAGS(!PageLocked(page), page);
> +
> 	lock_page_memcg(page);
> 	if (mapping && mapping_use_writeback_tags(mapping)) {
> 		XA_STATE(xas, &mapping->i_pages, page_index(page));
>