On Mon, 2015-02-23 at 10:06 +0100, Sebastian Andrzej Siewior wrote:
> - a patch to properly use the rtmutex deadlock detector in ww-mutex
> which seems to cure a nouveau deadlock (Gustavo Bittencourt)
How about the below instead. In 4.0.0-rt, i915 deadlocked, and the
below fixed that. DRM doesn't actually _work_ in 4.0-rt mind you,
there's something else lurking as well, but the locking is now happy,
and 3.18-rt continues to work just fine.
locking, ww_mutex: fix ww_mutex vs self-deadlock
If the caller already holds the mutex, task_blocks_on_rt_mutex()
returns -EDEADLK, we proceed directly to rt_mutex_handle_deadlock()
where it's instant game over.
Let ww_mutexes return EDEADLK/EALREADY as they want to instead.
Signed-off-by: Mike Galbraith <umgwanakikbuti@xxxxxxxxx>
---
kernel/locking/rtmutex.c | 17 +++++++++++------
1 file changed, 11 insertions(+), 6 deletions(-)
--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -1706,8 +1706,12 @@ rt_mutex_slowlock(struct rt_mutex *lock,
ret = task_blocks_on_rt_mutex(lock, &waiter, current, chwalk);
if (likely(!ret))
- ret = __rt_mutex_slowlock(lock, state, timeout, &waiter,
- ww_ctx);
+ ret = __rt_mutex_slowlock(lock, state, timeout, &waiter, ww_ctx);
+ else if (ww_ctx) {
+ /* ww_mutex received EDEADLK, let it become EALREADY */
+ ret = __mutex_lock_check_stamp(lock, ww_ctx);
+ BUG_ON(!ret);
+ }
set_current_state(TASK_RUNNING);
@@ -1715,6 +1719,9 @@ rt_mutex_slowlock(struct rt_mutex *lock,
if (rt_mutex_has_waiters(lock))
remove_waiter(lock, &waiter);
rt_mutex_handle_deadlock(ret, chwalk, &waiter);
+ /* ww_mutex want to report EDEADLK/EALREADY, let them */
+ if (!ww_ctx)
+ rt_mutex_handle_deadlock(ret, chwalk, &waiter);
} else if (ww_ctx) {
ww_mutex_account_lock(lock, ww_ctx);
}
@@ -2258,8 +2265,7 @@ __ww_mutex_lock_interruptible(struct ww_
might_sleep();
mutex_acquire_nest(&lock->base.dep_map, 0, 0, &ww_ctx->dep_map, _RET_IP_);
- ret = rt_mutex_slowlock(&lock->base.lock, TASK_INTERRUPTIBLE, NULL,
- RT_MUTEX_FULL_CHAINWALK, ww_ctx);
+ ret = rt_mutex_slowlock(&lock->base.lock, TASK_INTERRUPTIBLE, NULL, 0, ww_ctx);
if (ret)
mutex_release(&lock->base.dep_map, 1, _RET_IP_);
else if (!ret && ww_ctx->acquired > 1)
@@ -2277,8 +2283,7 @@ __ww_mutex_lock(struct ww_mutex *lock, s
might_sleep();
mutex_acquire_nest(&lock->base.dep_map, 0, 0, &ww_ctx->dep_map, _RET_IP_);
- ret = rt_mutex_slowlock(&lock->base.lock, TASK_UNINTERRUPTIBLE, NULL,
- RT_MUTEX_FULL_CHAINWALK, ww_ctx);
+ ret = rt_mutex_slowlock(&lock->base.lock, TASK_UNINTERRUPTIBLE, NULL, 0, ww_ctx);
if (ret)
mutex_release(&lock->base.dep_map, 1, _RET_IP_);
else if (!ret && ww_ctx->acquired > 1)
--
To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
