Hi all
In many files & functions of the kernel that use the tasklist_lock
it unlocks this lock before it stops using references to the task pointer
It has been get with the kernel lock (fs/proc/array.c is a file with many
examples). I think this way the task could exit on other processor and
then the references will be bogus.
An example this is the function linux-2.2.x/fs/proc/array.c:get_env(),
and even they notice it:
read_unlock(&tasklist_lock); /* FIXME!! This should be done after the last use */
I have made a propossal patch of how I think it should be (see below)
__please_reply_this_questions__:
- Is it a real bug?
- Have I fixed it correctly?
Please CC: your comments to uaca@alumni.uv.es
The example code patch follows:
--- array.c Wed Jun 7 23:26:44 2000
+++ array.c.modified Thu Jul 6 14:21:12 2000
@@ -483,14 +483,25 @@
static int get_env(int pid, char * buffer)
{
struct task_struct *p;
-
+ int retval;
+
read_lock(&tasklist_lock);
p = find_task_by_pid(pid);
- read_unlock(&tasklist_lock); /* FIXME!! This should be done after the last use */
- if (!p || !p->mm)
+ /* If unlock the tasklist_lock before use this, the task could exit
+ * on other processor and then the references be bogus, this also
+ * happens on other functions that use the tasklist_lock
+ */
+ if (!p && !p->mm) { /* We need both p and p->mm valid (non-NULL) pointers */
+ read_unlock(&tasklist_lock);
return 0;
- return get_array(p, p->mm->env_start, p->mm->env_end, buffer);
+ }
+
+ /* Seems get_array() can't sleep */
+ retval= get_array(p, p->mm->env_start, p->mm->env_end, buffer);
+ read_unlock(&tasklist_lock);
+
+ return retval;
}
static int get_arg(int pid, char * buffer)
Debian/GNU Linux: a dream come true
-----------------------------------------------------------------------------
"Computers are useless. They can only give answers." Pablo Picasso
---> Visita http://www.valux.org/ para saber acerca de la <---
---> Asociación Valenciana de Usuarios de Linux <---
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Jul 07 2000 - 21:00:18 EST