From: Arnd Bergmann <[email protected]>
The folio changes added a variable that is sometimes unused:
fs/9p/vfs_addr.c: In function 'v9fs_release_page':
fs/9p/vfs_addr.c:140:23: error: unused variable 'inode' [-Werror=unused-variable]
140 | struct inode *inode = folio_inode(folio);
| ^~~~~
Make this clearer to the compiler by replacing the #ifdef
with an equivalent if(IS_ENABLED()) check.
Fixes: 78525c74d9e7 ("netfs, 9p, afs, ceph: Use folios")
Signed-off-by: Arnd Bergmann <[email protected]>
---
fs/9p/vfs_addr.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/fs/9p/vfs_addr.c b/fs/9p/vfs_addr.c
index 1279970e9157..82ac2ceff06e 100644
--- a/fs/9p/vfs_addr.c
+++ b/fs/9p/vfs_addr.c
@@ -141,14 +141,17 @@ static int v9fs_release_page(struct page *page, gfp_t gfp)
if (folio_test_private(folio))
return 0;
-#ifdef CONFIG_9P_FSCACHE
+
+ if (!IS_ENABLED(CONFIG_9P_FSCACHE))
+ return 1;
+
if (folio_test_fscache(folio)) {
if (!(gfp & __GFP_DIRECT_RECLAIM) || !(gfp & __GFP_FS))
return 0;
folio_wait_fscache(folio);
}
fscache_note_page_release(v9fs_inode_cookie(V9FS_I(inode)));
-#endif
+
return 1;
}
--
2.29.2
Dominique Martinet wrote on Sat, Dec 04, 2021 at 09:04:04AM +0900:
> Arnd Bergmann wrote on Fri, Dec 03, 2021 at 07:58:06PM +0100:
> > From: Arnd Bergmann <[email protected]>
> >
> > The folio changes added a variable that is sometimes unused:
> >
> > fs/9p/vfs_addr.c: In function 'v9fs_release_page':
> > fs/9p/vfs_addr.c:140:23: error: unused variable 'inode' [-Werror=unused-variable]
> > 140 | struct inode *inode = folio_inode(folio);
> > | ^~~~~
> >
> > Make this clearer to the compiler by replacing the #ifdef
> > with an equivalent if(IS_ENABLED()) check.
> >
> > Fixes: 78525c74d9e7 ("netfs, 9p, afs, ceph: Use folios")
> > Signed-off-by: Arnd Bergmann <[email protected]>
>
> Looks good to me, picking it up
Sorry it took me a while to actually do the picking up part, but this
appears to have been a patch for linux-next back then and I didn't
notice because the Fixes tag is incorrect (78525c74d9e7 didn't introduce
the inode variable, it actually fixes a patch that never has been merged)
David since then fixed the warning differently in v2 of the patch (he
moved the fscache_note_page_release() out of the ifdef), so I won't do
anything with this even if in principle I tend to agree that
if(IS_ENABLED()) lead to better compiler coverage
Thanks though!
--
Dominique
Dominique Martinet <[email protected]> wrote:
> David since then fixed the warning differently in v2 of the patch (he
> moved the fscache_note_page_release() out of the ifdef), so I won't do
> anything with this even if in principle I tend to agree that
> if(IS_ENABLED()) lead to better compiler coverage
Yeah, fscache_note_page_release() compiles out in such a case because
v9fs_inode_cookie() becomes unconditionally NULL if it's disabled.
If you want me to do something different, can you give me an incremental patch
to merge into mine?
David