Thanks Alex.
-mm also needs s/page_cache_release/read_cache_pages_release_page/ on patch
--
From: Alexander Zarochentsev <[email protected]>
Don't let fuse_readpages leave the @pages list not empty when exiting
on error.
Signed-off-by: Alexander Zarochentsev <[email protected]>
Signed-off-by: Miklos Szeredi <[email protected]>
---
Index: linux/fs/fuse/file.c
===================================================================
--- linux.orig/fs/fuse/file.c 2006-08-10 09:41:12.000000000 +0200
+++ linux/fs/fuse/file.c 2006-08-10 09:43:36.000000000 +0200
@@ -395,14 +395,16 @@ static int fuse_readpages(struct file *f
struct fuse_readpages_data data;
int err;
+ err = -EIO;
if (is_bad_inode(inode))
- return -EIO;
+ goto clean_pages_up;
data.file = file;
data.inode = inode;
data.req = fuse_get_req(fc);
+ err = PTR_ERR(data.req);
if (IS_ERR(data.req))
- return PTR_ERR(data.req);
+ goto clean_pages_up;
err = read_cache_pages(mapping, pages, fuse_readpages_fill, &data);
if (!err) {
@@ -412,6 +414,10 @@ static int fuse_readpages(struct file *f
fuse_put_request(fc, data.req);
}
return err;
+
+clean_pages_up:
+ readpages_cleanup_helper(pages);
+ return err;
}
static size_t fuse_send_write(struct fuse_req *req, struct file *file,
Index: linux/include/linux/pagemap.h
===================================================================
--- linux.orig/include/linux/pagemap.h 2006-08-10 09:41:14.000000000 +0200
+++ linux/include/linux/pagemap.h 2006-08-10 09:42:33.000000000 +0200
@@ -98,6 +98,7 @@ extern struct page * read_cache_page(str
void *data);
extern int read_cache_pages(struct address_space *mapping,
struct list_head *pages, filler_t *filler, void *data);
+extern void readpages_cleanup_helper(struct list_head *);
static inline struct page *read_mapping_page(struct address_space *mapping,
unsigned long index, void *data)
Index: linux/mm/readahead.c
===================================================================
--- linux.orig/mm/readahead.c 2006-08-10 09:41:14.000000000 +0200
+++ linux/mm/readahead.c 2006-08-10 09:42:33.000000000 +0200
@@ -118,6 +118,22 @@ static inline unsigned long get_next_ra_
#define list_to_page(head) (list_entry((head)->prev, struct page, lru))
/**
+ * may be useful in foo_fs_readpages method for the page pool cleanup
+ * before exiting on error.
+ */
+void readpages_cleanup_helper(struct list_head *pages)
+{
+ while (!list_empty(pages)) {
+ struct page *victim;
+
+ victim = list_to_page(pages);
+ list_del(&victim->lru);
+ page_cache_release(victim);
+ }
+}
+EXPORT_SYMBOL(readpages_cleanup_helper);
+
+/**
* read_cache_pages - populate an address space with some pages & start reads against them
* @mapping: the address_space
* @pages: The address of a list_head which contains the target pages. These
@@ -147,13 +163,7 @@ int read_cache_pages(struct address_spac
if (!pagevec_add(&lru_pvec, page))
__pagevec_lru_add(&lru_pvec);
if (ret) {
- while (!list_empty(pages)) {
- struct page *victim;
-
- victim = list_to_page(pages);
- list_del(&victim->lru);
- page_cache_release(victim);
- }
+ readpages_cleanup_helper(pages);
break;
}
}
On Thu, 10 Aug 2006 09:57:22 +0200
Miklos Szeredi <[email protected]> wrote:
> Thanks Alex.
>
> -mm also needs s/page_cache_release/read_cache_pages_release_page/ on patch
>
> --
> From: Alexander Zarochentsev <[email protected]>
>
> Don't let fuse_readpages leave the @pages list not empty when exiting
> on error.
>
hm, OK.
> +extern void readpages_cleanup_helper(struct list_head *);
Can we think of a better name? I mean, all it does is throw away a list of
pages. The caller doesn't _have_ to be using it within the context of
read_cache_pages(). put_pages()? put_pages_list()?
And if we do make this a more generic thing, it'd be better off living in
mm/swap.c, which is really mm/mm-library-stuff.c nowadays.
> /**
> + * may be useful in foo_fs_readpages method for the page pool cleanup
> + * before exiting on error.
> + */
"/**" introduces a kerneldoc comment, but that isn't a kerneldoc comment.
It would be best to turn it into a kerneldoc comment.
> > Thanks Alex.
> >
> > -mm also needs s/page_cache_release/read_cache_pages_release_page/ on patch
> >
> > --
> > From: Alexander Zarochentsev <[email protected]>
> >
> > Don't let fuse_readpages leave the @pages list not empty when exiting
> > on error.
> >
>
> hm, OK.
>
> > +extern void readpages_cleanup_helper(struct list_head *);
>
> Can we think of a better name? I mean, all it does is throw away a list of
> pages. The caller doesn't _have_ to be using it within the context of
> read_cache_pages(). put_pages()? put_pages_list()?
So what's up with read_cache_pages_release_page() in -mm, which seems
to be rather specific to read_cache_pages()?
Miklos
On Thu, 10 Aug 2006 10:45:59 +0200
Miklos Szeredi <[email protected]> wrote:
> > > Thanks Alex.
> > >
> > > -mm also needs s/page_cache_release/read_cache_pages_release_page/ on patch
> > >
> > > --
> > > From: Alexander Zarochentsev <[email protected]>
> > >
> > > Don't let fuse_readpages leave the @pages list not empty when exiting
> > > on error.
> > >
> >
> > hm, OK.
> >
> > > +extern void readpages_cleanup_helper(struct list_head *);
> >
> > Can we think of a better name? I mean, all it does is throw away a list of
> > pages. The caller doesn't _have_ to be using it within the context of
> > read_cache_pages(). put_pages()? put_pages_list()?
>
> So what's up with read_cache_pages_release_page() in -mm, which seems
> to be rather specific to read_cache_pages()?
>
That's part of the readahead patches. Don't look at that ;)