On Tue, Jun 06, 2017 at 12:35:30PM +0300, Sergei Shtylyov wrote:
> Hello!
>
> On 6/6/2017 1:08 AM, Eugeniu Rosca wrote:
>
> >Commit a47b70ea86bd ("ravb: unmap descriptors when freeing rings") has
> >introduced the issue seen in [1] reproduced on H3ULCB board.
> >
> >Fix this by relocating the RX skb ringbuffer free operation, so that
> >swiotlb page unmapping can be done first. Freeing of aligned TX buffers
> >is not relevant to the issue seen in [1]. Still, reposition TX free
> >calls as well, to have all kfree() operations performed consistently
> >_after_ dma_unmap_*()/dma_free_*().
>
> Perhaps it's a material of a separate cleanup patch?
Many thanks for feedback. For the moment, with a number of sanitizers
and debugging options enabled (UBSAN, KASAN, KMEMLEAK, DMA_API_DEBUG), I
couldn't find any other obvious ravb driver failures in basic usecases
(didn't stress-test it though).
Regarding the reordering of kfree vs dma_* API calls, which might be
needed in other parts of the driver, this possibly will be highlighted
by special usecases like repetitive suspend/resume or the like. I will
happily share any other fixes, if such are developed on our side.
Best regards,
Eugeniu.
