Re: [Linaro-mm-sig] [PATCH 13/24] dma-buf: drop the DAG approach for the dma_resv object

On Tue, Dec 07, 2021 at 01:34:00PM +0100, Christian König wrote:

So far we had the approach of using a directed acyclic graph with the dma_resv obj.

This turned out to have many downsides, especially it means that every single driver and user of this interface needs to be aware of this restriction when adding fences. If the rules for the DAG are not followed then we end up with potential hard to debug memory corruption, information leaks or even elephant big security holes because we allow userspace to access freed up memory.

Since we already took a step back from that by always looking at all fences we now go a step further and stop dropping the shared fences when a new exclusive one is added.

Signed-off-by: Christian König christian.koenig@amd.com

drivers/dma-buf/dma-resv.c | 13 ------------- 1 file changed, 13 deletions(-)

diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c index 9acceabc9399..ecb2ff606bac 100644 --- a/drivers/dma-buf/dma-resv.c +++ b/drivers/dma-buf/dma-resv.c

No doc update at all!

I checked, we're not that shitty with docs, Minimally the DOC: section header and also the struct dma_resv kerneldoc. Also there's maybe more references and stuff I've missed on a quick look, please check for them (e.g. dma_buf.resv kerneldoc is rather important to keep correct too).

Code itself does what it says in the commit message, but we really need the most accurate docs we can get for this stuff, or the confusion will persist :-/

Cheers, Daniel

@@ -383,29 +383,16 @@ EXPORT_SYMBOL(dma_resv_replace_fences); void dma_resv_add_excl_fence(struct dma_resv *obj, struct dma_fence *fence) { struct dma_fence *old_fence = dma_resv_excl_fence(obj);

• struct dma_resv_list *old;
• u32 i = 0;

dma_resv_assert_held(obj);

• old = dma_resv_shared_list(obj);
• if (old)

• i = old->shared_count;
  

• dma_fence_get(fence);

write_seqcount_begin(&obj->seq); /* write_seqcount_begin provides the necessary memory barrier */ RCU_INIT_POINTER(obj->fence_excl, fence);

• if (old)

• old->shared_count = 0;
  

  write_seqcount_end(&obj->seq);

• /* inplace update, no shared fences */
• while (i--)

• dma_fence_put(rcu_dereference_protected(old->shared[i],
  

• 				dma_resv_held(obj)));
  

• dma_fence_put(old_fence);

} EXPORT_SYMBOL(dma_resv_add_excl_fence); -- 2.25.1