Messages in this thread | | | Date | Mon, 16 Dec 2013 17:34:56 -0800 | Subject | Re: [RFC][PATCH 3/3] staging: ion: Avoid using rt_mutexes directly. | From | Colin Cross <> |
| |
On Mon, Dec 16, 2013 at 5:22 PM, John Stultz <john.stultz@linaro.org> wrote: > On 12/16/2013 04:17 PM, Colin Cross wrote: >> The lock is only used to protect the heap's free_list and >> free_list_size members, and is not held for any long or sleeping >> operations. Converting to a spinlock should prevent priority >> inversion without using the rt_mutex. I'd also rename it to free_lock >> to so it doesn't get used as a general heap lock. > > Hrm.. So at least a trivial conversion to use spinlocks doesn't quite > work out, as we call ion_buffer_destroy() in ion_heap_freelist_drain() > while holding the lock, and that calls all sorts of not safe stuff. > > I'll spend some more time looking at it later tonight, but let me know > if you have an approach for this case in mind.
Drop and re-grab the lock around ion_buffer_destroy, it's not necessary during the destroy, and the list iteration is already using list_for_each_entry_safe, so it doesn't matter if another caller modifies the list during the destroy.
| |