IB/rxe: Wait for tasklets to finish before tearing down QP

The system may crash when a malformed request is received and
the error is detected by the responder.

NodeA: $ ibv_rc_pingpong -g 0 -d rxe0 -i 1 -n 1 -s 50000
NodeB: $ ibv_rc_pingpong -g 0 -d rxe0 -i 1 -n 1 -s 1024 <NodeA_ip>

The responder generates a receive error on node B since the incoming
SEND is oversized. If the client tears down the QP before the responder
or the completer finish running, a page fault may occur.

The fix makes the destroy operation spin until the tasks complete, which
appears to be original intent of the design.

Signed-off-by: Andrew Boyer <andrew.boyer@dell.com>
Reviewed-by: Yuval Shaia <yuval.shaia@oracle.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
This commit is contained in:
Andrew Boyer 2016-12-05 08:43:20 -05:00 committed by Doug Ledford
parent 5407f53012
commit 07bf9627d5
2 changed files with 20 additions and 0 deletions

View file

@ -121,6 +121,7 @@ int rxe_init_task(void *obj, struct rxe_task *task,
task->arg = arg;
task->func = func;
snprintf(task->name, sizeof(task->name), "%s", name);
task->destroyed = false;
tasklet_init(&task->tasklet, rxe_do_task, (unsigned long)task);
@ -132,11 +133,29 @@ int rxe_init_task(void *obj, struct rxe_task *task,
void rxe_cleanup_task(struct rxe_task *task)
{
unsigned long flags;
bool idle;
/*
* Mark the task, then wait for it to finish. It might be
* running in a non-tasklet (direct call) context.
*/
task->destroyed = true;
do {
spin_lock_irqsave(&task->state_lock, flags);
idle = (task->state == TASK_STATE_START);
spin_unlock_irqrestore(&task->state_lock, flags);
} while (!idle);
tasklet_kill(&task->tasklet);
}
void rxe_run_task(struct rxe_task *task, int sched)
{
if (task->destroyed)
return;
if (sched)
tasklet_schedule(&task->tasklet);
else

View file

@ -54,6 +54,7 @@ struct rxe_task {
int (*func)(void *arg);
int ret;
char name[16];
bool destroyed;
};
/*