On Wed, May 13, 2020 at 08:17:54AM -0700, Darrick J. Wong wrote:
> On Wed, May 13, 2020 at 04:39:05PM +0300, Dan Carpenter wrote:
> > Oh yeah. You're right. This patch isn't correct. Sorry about that.
> >
> > I worry that there are several static analyzer's which will warn about
> > this code...
>
> /me wonders if this particular instance ought to have a breadcrumb to
> remind future readers that we can handle the lack of memory, e.g.
>
> cilist = kmem_alloc(..., KM_MAYFAIL...);
> if (!cilist) {
> /* memory is tight, so defer the inode cluster flush */
> goto out_put;
> }
I'm working on patches that make this memory allocation go away
altogether, so I'd suggest just ignoring it for now.
Cheers,
Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
