On Thu, Apr 3, 2014 at 9:49 AM, Tejun Heo <tj@xxxxxxxxxx> wrote:
>
> C3. fed95bab8d29 ("sysfs: fix namespace refcnt leak") added an
> argument to kernfs_mount() and was routed through driver-core-next
> after cgroup pulled the tree. cgroup's usage needs to be updated
> accordingly.
This one I find very annoying. There's exactly *one* user of that
interface (back in fed95bab8d29 there were none), and it doesn't want
that dummy argument.
WTF? Why did the driver-core-next development add that argument at
all, rather than just pass in NULL directly? Do we expect to grow more
users of kernfs_mount?
And the "bool *new_sb_created" argument really makes *zero* sense to
kernfs_mount(). It was added to fix a namespace refcount leak, BUT
kernfs_mount() DOES NOT TAKE A NAMESPACE PARAMETER!
So quite frankly, that kernfs_mount() calling convention change looks
completely f*cked up.
I think I'm going to resolve that conflict differently from your
suggested one - I'm going to just remove the broken stupid
"new_sb_created" argument from the kernfs_mount() helper function.
(It obviously stays around for kernfs_mount_ns() that *does* take a
namespace pointer).
People, holler if you disagree, I'll hold off pushing out my
resolution for a while anyway (do all the normal build tests and
reboot into it).
Linus
_______________________________________________
Containers mailing list
Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linuxfoundation.org/mailman/listinfo/containers
