On Tue, Aug 24, 2021 at 1:56 PM Jeff Layton <jlayton@xxxxxxxxxx> wrote:
>
> On Tue, 2021-08-24 at 13:12 +0200, Lukas Bulwahn wrote:
> > Commit 3efee0567b4a ("fs: remove mandatory file locking support") removes
> > some operations in functions rw_verify_area() and remap_verify_area().
> >
> > As these functions are now simplified, do some syntactic clean-up as
> > follow-up to the removal as well, which was pointed out by compiler
> > warnings and static analysis.
> >
> > No functional change.
> >
> > Signed-off-by: Lukas Bulwahn <lukas.bulwahn@xxxxxxxxx>
> > ---
> > Jeff, please pick this clean-up patch on top of the commit above.
> >
> > fs/read_write.c | 10 +++-------
> > fs/remap_range.c | 2 --
> > 2 files changed, 3 insertions(+), 9 deletions(-)
> >
> > diff --git a/fs/read_write.c b/fs/read_write.c
> > index ffe821b8588e..af057c57bdc6 100644
> > --- a/fs/read_write.c
> > +++ b/fs/read_write.c
> > @@ -365,12 +365,8 @@ SYSCALL_DEFINE5(llseek, unsigned int, fd, unsigned long, offset_high,
> >
> > int rw_verify_area(int read_write, struct file *file, const loff_t *ppos, size_t count)
> > {
> > - struct inode *inode;
> > - int retval = -EINVAL;
> > -
> > - inode = file_inode(file);
> > if (unlikely((ssize_t) count < 0))
> > - return retval;
> > + return -EINVAL;
> >
> > /*
> > * ranged mandatory locking does not apply to streams - it makes sense
> > @@ -381,12 +377,12 @@ int rw_verify_area(int read_write, struct file *file, const loff_t *ppos, size_t
> >
> > if (unlikely(pos < 0)) {
> > if (!unsigned_offsets(file))
> > - return retval;
> > + return -EINVAL;
> > if (count >= -pos) /* both values are in 0..LLONG_MAX */
> > return -EOVERFLOW;
> > } else if (unlikely((loff_t) (pos + count) < 0)) {
> > if (!unsigned_offsets(file))
> > - return retval;
> > + return -EINVAL;
> > }
> > }
> >
> > diff --git a/fs/remap_range.c b/fs/remap_range.c
> > index ec6d26c526b3..6d4a9beaa097 100644
> > --- a/fs/remap_range.c
> > +++ b/fs/remap_range.c
> > @@ -99,8 +99,6 @@ static int generic_remap_checks(struct file *file_in, loff_t pos_in,
> > static int remap_verify_area(struct file *file, loff_t pos, loff_t len,
> > bool write)
> > {
> > - struct inode *inode = file_inode(file);
> > -
> > if (unlikely(pos < 0 || len < 0))
> > return -EINVAL;
> >
>
> Thanks Lukas,
>
> I had already removed the second hunk, but I merged read_write.c part
> into my queue for v5.15. It should show up in linux-next soon.
Yeah, I guess the issue on the second hunk caused a compiler warning
(so more obvious) and was reported otherwise; the issue around the
first hunk is only discovered with a bit more involved static
analysis, done by clang-analyzer, coccinelle scripts and friends.
Lukas
