Re: [PATCH] protect remove_proc_entry

[Mitchell Blank Jr]

Steven Rostedt wrote:
> > I guess the lock_kernel() approach is the way to go.  Fixing a race and
> > de-BKLing procfs are separate exercises...
> >
> > Did the patch work?
> 
> Sorry, I forgot to respond, because the test is still running ;)
> 
> So yes, it not only ran for three days, it ran for six. I just killed it.

Have you looked at the other paths that touch ->subdir?  Namely:
  proc_devtree.c:
    proc_device_tree_add_node() -- plays games with ->subdir directly
  generic.c:
    proc_create() -- calls xlate_proc_name() which touches ->subdir and
      should therefore probably be called under BKL
    proc_register() -- both the call to xlate_proc_name() and the
      following accesses to ->subdir should be under BKL

-Mitch
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/