]> git.baikalelectronics.ru Git - kernel.git/commit
do_wait: fix security checks
authorOleg Nesterov <oleg@tv-sign.ru>
Fri, 8 Feb 2008 12:19:06 +0000 (04:19 -0800)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Fri, 8 Feb 2008 17:22:26 +0000 (09:22 -0800)
commitbb9f29b7fb74fa27e82c049db075aab958914cc3
tree37b08158ee5296e79a61aad086be36c742825e3b
parentcd98a19131f7b460698821a8bf9cf3706f96ff45
do_wait: fix security checks

Imho, the current usage of security_task_wait() is not logical.

Suppose we have the single child p, and security_task_wait(p) return
-EANY.  In that case waitpid(-1) returns this error.  Why? Isn't it
better to return ECHLD? We don't really have reapable children.

Now suppose that child was stolen by gdb.  In that case we find this
child on ->ptrace_children and set flag = 1, but we don't check that the
child was denied.  So, do_wait(..., WNOHANG) returns 0, this doesn't
match the behaviour above.  Without WNOHANG do_wait() blocks only to
return the error later, when the child will be untraced.  Inho, really
strange.

I think eligible_child() should return the error only if the child's pid
was requested explicitly, otherwise we should silently ignore the tasks
which were nacked by security_task_wait().

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Roland McGrath <roland@redhat.com>
Cc: Chris Wright <chrisw@sous-sol.org>
Cc: Eric Paris <eparis@redhat.com>
Cc: James Morris <jmorris@namei.org>
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/exit.c