On Tue, Feb 24, 2015 at 06:08:54PM +0200, Stanimir Varbanov wrote: > yes, there is a potential race between atomic_inc and dma callback. I > reordered these calls to save few checks, and now it returns to me. > I imagine few options here: > - reorder the dmaengine calls and atomic operations, i.e. > call atomic_inc for rx and tx channels before corresponding > dmaengine_submit and dmaengine_issue_pending. > - have two different dma callbacks and two completions and waiting for > the two. > - manage to receive only one dma callback, i.e. the last transfer in > case of presence of the rx_buf and tx_buf at the same time. > - let me see for better solution. Any solution which doesn't make use of atomics is likely to be better, as I said they are enormously error prone. A more common approach is a single completion triggering on the RX (for RX only or bidirectional transfers) or TX if that's the only thing active. For most hardware you can just use the RX to manage completion since it must of necessity complete at the same time as or later than the transmit side, transmit often completes early since the DMA completes when the FIFO is full not when the data is on the wire.
Attachment:
signature.asc
Description: Digital signature
