| Message ID | 20220307142121.GB19660@kili (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ceph: uninitialized variable in debug output | expand |
On 3/7/22 10:21 PM, Dan Carpenter wrote: > If read_mapping_folio() fails then "inline_version" is printed without > being initialized. > > Fixes: 083db6fd3e73 ("ceph: uninline the data on a file opened for writing") > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> > --- > fs/ceph/addr.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c > index 3c1257b09775..0d4120297ede 100644 > --- a/fs/ceph/addr.c > +++ b/fs/ceph/addr.c > @@ -1632,9 +1632,10 @@ int ceph_uninline_data(struct file *file) > struct ceph_osd_request *req; > struct ceph_cap_flush *prealloc_cf; > struct folio *folio = NULL; > + u64 inline_version = -1; > struct page *pages[1]; > - u64 len, inline_version; > int err = 0; > + u64 len; > > prealloc_cf = ceph_alloc_cap_flush(); > if (!prealloc_cf) Possibly we'd better format the 'inline_version' in hexadecimal ? Reviewed-by: Xiubo Li<xiubli@redhat.com>
Dan Carpenter <dan.carpenter@oracle.com> wrote:
> + u64 inline_version = -1;
ULLONG_MAX?
David
On Tue, 2022-03-08 at 09:36 +0000, David Howells wrote: > Dan Carpenter <dan.carpenter@oracle.com> wrote: > > > + u64 inline_version = -1; > > ULLONG_MAX? > > David > ...or maybe CEPH_INLINE_NONE, which is: #define CEPH_INLINE_NONE ((__u64)-1)
On Tue, 2022-03-08 at 08:24 -0500, Jeff Layton wrote: > On Tue, 2022-03-08 at 09:36 +0000, David Howells wrote: > > Dan Carpenter <dan.carpenter@oracle.com> wrote: > > > > > + u64 inline_version = -1; > > > > ULLONG_MAX? > > > > David > > > > ...or maybe CEPH_INLINE_NONE, which is: > > #define CEPH_INLINE_NONE ((__u64)-1) > I went ahead and merged the patch into our testing branch, with the change to use CEPH_INLINE_NONE instead. Thanks!
On Tue, Mar 08, 2022 at 08:28:21AM -0500, Jeff Layton wrote: > On Tue, 2022-03-08 at 08:24 -0500, Jeff Layton wrote: > > On Tue, 2022-03-08 at 09:36 +0000, David Howells wrote: > > > Dan Carpenter <dan.carpenter@oracle.com> wrote: > > > > > > > + u64 inline_version = -1; > > > > > > ULLONG_MAX? > > > > > > David > > > > > > > ...or maybe CEPH_INLINE_NONE, which is: > > > > #define CEPH_INLINE_NONE ((__u64)-1) > > > > I went ahead and merged the patch into our testing branch, with the > change to use CEPH_INLINE_NONE instead. I liked CEPH_INLINE_NONE so I was resending it right now... But thanks! regards, dan carpenter
diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 3c1257b09775..0d4120297ede 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1632,9 +1632,10 @@ int ceph_uninline_data(struct file *file) struct ceph_osd_request *req; struct ceph_cap_flush *prealloc_cf; struct folio *folio = NULL; + u64 inline_version = -1; struct page *pages[1]; - u64 len, inline_version; int err = 0; + u64 len; prealloc_cf = ceph_alloc_cap_flush(); if (!prealloc_cf)
If read_mapping_folio() fails then "inline_version" is printed without being initialized. Fixes: 083db6fd3e73 ("ceph: uninline the data on a file opened for writing") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> --- fs/ceph/addr.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)