]> www.pilppa.org Git - linux-2.6-omap-h63xx.git/commitdiff
[PATCH] copy_process: cosmetic ->ioprio tweak
authorOleg Nesterov <oleg@tv-sign.ru>
Fri, 29 Sep 2006 09:00:52 +0000 (02:00 -0700)
committerLinus Torvalds <torvalds@g5.osdl.org>
Fri, 29 Sep 2006 16:18:18 +0000 (09:18 -0700)
copy_process:
// holds tasklist_lock + ->siglock
       /*
        * inherit ioprio
        */
       p->ioprio = current->ioprio;

Why?  ->ioprio was already copied in dup_task_struct().  I guess this is
needed to ensure that the child can't escape
sys_ioprio_set(IOPRIO_WHO_{PGRP,USER}), yes?

In that case we don't need ->siglock held, and the comment should be
updated.

Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
kernel/fork.c

index bca6ce6d3ded7f2ef5c4048296ce13aa5d554bcc..1c999f3e0b47f02a4fd805daf90f9b914c44282b 100644 (file)
@@ -1150,7 +1150,6 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 
        /* Our parent execution domain becomes current domain
           These must match for thread signalling to apply */
-          
        p->parent_exec_id = p->self_exec_id;
 
        /* ok, now we should be set up.. */
@@ -1173,6 +1172,9 @@ static struct task_struct *copy_process(unsigned long clone_flags,
        /* Need tasklist lock for parent etc handling! */
        write_lock_irq(&tasklist_lock);
 
+       /* for sys_ioprio_set(IOPRIO_WHO_PGRP) */
+       p->ioprio = current->ioprio;
+
        /*
         * The task hasn't been attached yet, so its cpus_allowed mask will
         * not be changed, nor will its assigned CPU.
@@ -1232,11 +1234,6 @@ static struct task_struct *copy_process(unsigned long clone_flags,
                }
        }
 
-       /*
-        * inherit ioprio
-        */
-       p->ioprio = current->ioprio;
-
        if (likely(p->pid)) {
                add_parent(p);
                if (unlikely(p->ptrace & PT_PTRACED))