Re: [PATCH] [RFC] dma-buf: implement vmap refcounting in the interface logic

2012-12-20 Thread Rob Clark
On Thu, Dec 20, 2012 at 10:50 AM, Rob Clark wrote: > On Thu, Dec 20, 2012 at 7:14 AM, Daniel Vetter wrote: >> All drivers which implement this need to have some sort of refcount to >> allow concurrent vmap usage. Hence implement this in the dma-buf core. >> >> To protect against concurrent calls

Re: [PATCH] [RFC] dma-buf: implement vmap refcounting in the interface logic

2012-12-20 Thread Rob Clark
On Thu, Dec 20, 2012 at 7:14 AM, Daniel Vetter wrote: > All drivers which implement this need to have some sort of refcount to > allow concurrent vmap usage. Hence implement this in the dma-buf core. > > To protect against concurrent calls we need a lock, which potentially > causes new funny locki

Re: [PATCH] [RFC] dma-buf: implement vmap refcounting in the interface logic

2012-12-20 Thread Aaron Plattner
On 12/20/2012 05:14 AM, Daniel Vetter wrote: All drivers which implement this need to have some sort of refcount to allow concurrent vmap usage. Hence implement this in the dma-buf core. To protect against concurrent calls we need a lock, which potentially causes new funny locking inversions. Bu

[PATCH] [RFC] dma-buf: implement vmap refcounting in the interface logic

2012-12-20 Thread Daniel Vetter
All drivers which implement this need to have some sort of refcount to allow concurrent vmap usage. Hence implement this in the dma-buf core. To protect against concurrent calls we need a lock, which potentially causes new funny locking inversions. But this shouldn't be a problem for exporters wit

Re: [PATCH] [RFC] dma-buf: implement vmap refcounting in the interface logic

2012-12-19 Thread Aaron Plattner
On 12/17/2012 03:58 PM, Daniel Vetter wrote: All drivers which implement this need to have some sort of refcount to allow concurrent vmap usage. Hence implement this in the dma-buf core. To protect against concurrent calls we need a lock, which potentially causes new funny locking inversions. Bu

[PATCH] [RFC] dma-buf: implement vmap refcounting in the interface logic

2012-12-17 Thread Daniel Vetter
All drivers which implement this need to have some sort of refcount to allow concurrent vmap usage. Hence implement this in the dma-buf core. To protect against concurrent calls we need a lock, which potentially causes new funny locking inversions. But this shouldn't be a problem for exporters wit

[PATCH] [RFC] dma-buf: implement vmap refcounting in the interface logic

2012-12-17 Thread Daniel Vetter
All drivers which implement this need to have some sort of refcount to allow concurrent vmap usage. Hence implement this in the dma-buf core. To protect against concurrent calls we need a lock, which potentially causes new funny locking inversions. But this shouldn't be a problem for exporters wit