Re: [PATCH] fs/ceph/file: fix memory leaks in __ceph_sync_read()

From: Max Kellermann
Date: Thu Nov 28 2024 - 07:28:43 EST


On Thu, Nov 28, 2024 at 1:18 PM Alex Markuze <amarkuze@xxxxxxxxxx> wrote:
> Pages are freed in `ceph_osdc_put_request`, trying to release them
> this way will end badly.

I don't get it. If this ends badly, why does the other
ceph_release_page_vector() call after ceph_osdc_put_request() in that
function not end badly?