Skip to contentSkip to conversation

ia64: ptrace_attach_sync_user_rbs: use ->sighand instead of ->signal to check the task is alive

PATCH In review Oleg Nesterov 2 participants 2 messages Feb 3, 2009 – Feb 5, 2009 ptrace support

lore
Share & tools
lore

Oleg Nesterov

author maintainer
#0 lore raw

PATCH 1/2 ia64: ptrace_attach_sync_user_rbs: use ->sighand instead of ->signal to check the task is alive

No functional changes.

It doesn't matter which pointer to check under tasklist to ensure the task was not released, ->signal or ->sighand. But we are going to make ->signal refcountable, change the code to use ->sighand.

Actually, this code doesn't need tasklist_lock and this check at all, it should be converted to use lock_task_sighand().

arch/ia64/kernel/ptrace.c renamed from arch/ia64/kernel/ptrace.c~1_IA64_PTRACE +2 -2
@@ -640,7 +640,7 @@ ptrace_attach_sync_user_rbs (struct task
 	 */
 
 	read_lock(&tasklist_lock);
-	if (child->signal) {
+	if (child->sighand) {
 		spin_lock_irq(&child->sighand->siglock);
 		if (child->state == TASK_STOPPED &&
 		    !test_and_set_tsk_thread_flag(child, TIF_RESTORE_RSE)) {
@@ -664,7 +664,7 @@ ptrace_attach_sync_user_rbs (struct task
 	 * job control stop, so that SIGCONT can be used to wake it up.
 	 */
 	read_lock(&tasklist_lock);
-	if (child->signal) {
+	if (child->sighand) {
 		spin_lock_irq(&child->sighand->siglock);
 		if (child->state == TASK_TRACED &&
 		    (child->signal->flags & SIGNAL_STOP_STOPPED)) {
#1 lore raw
Quoted 2 lines

Actually, this code doesn't need tasklist_lock and this check at all, it should be converted to use lock_task_sighand().

Might as well just do that, it's a harmless enough change. For other trivial no-op cleanup while you are there, task_is_stopped() and task_is_traced() are the norm now.

Thanks,
Roland