Re: [PATCH] dmaengine: qcom_hidma: release the descriptor before the callback

[Sinan Kaya]

On 8/4/2016 10:40 AM, Russell King - ARM Linux wrote:
> On Thu, Aug 04, 2016 at 10:17:24AM -0400, Sinan Kaya wrote:
>> On 8/4/2016 8:55 AM, Vinod Koul wrote:
>>> Dmaengine tells transaction is complete. It does not say if the txn is
>>> success or failure. It can transfer data and not say if data was
>>> correct. A successful transaction implies data integrity as well, which
>>> dmaengine can't provide.
>>
>> Thanks for describing this. I was confused about DMA_SUCCESS and DMA_COMPLETE.
>> I now understand that tx_success API just returns information that the request
>> was executed whether the result is error or not. This makes sense now.
>>
>> However, if the txn is failure; then we should never call the client callback
>> since DMA engine cannot provide such feedback to the client without Dave's patch.
>> You are saying that the calling the callback is optional.
>>
>> Then, the callback cannot be optional in the error case for old behavior.
>>
>> How does the client know if memcpy executed or not? The client got its callback
>> and tx_status is also DMA_COMPLETE.
> 
> If an error occurred, then dma_async_is_tx_complete() is supposed to
> return DMA_ERROR.  It's up to the DMA engine driver to ensure that
> this happens if it has error detection abilities.
> 

Well, that's not what I'm getting from Vinod and also from the current usage
in most of the drivers that I looked.

The current drivers implement tx_status as follows.

static enum dma_status xyz_tx_status(struct dma_chan *chan,
	dma_cookie_t cookie, struct dma_tx_state *state) 
{
...
	ret = dma_cookie_status(&c->vc.chan, cookie, state);
	if (ret == DMA_COMPLETE)
		return ret; 
...
}

What Vinod is telling me that I need to set the cookie to complete 
whether the transaction is successful or not if the request was accepted
by HW. xyz_tx_status is just an indication that the transaction was accepted
by HW. An error can happen as a result of transaction execution.

If I call dma_cookie_complete for all transactions regardless of transaction
success or not, then the xyz_tx_status returns DMA_COMPLETE. 

This also matches what Vinod is saying. The transaction is complete but
it may not be success. I'm saying that if we follow this scheme, then
we should not call the callback.

I also made the argument that the driver should not call dma_cookie_complete
for failure cases and somehow return an error for transactions that failed. 
This is your suggestion.

I don't think it matches Vinod's expectation.


> Most of the helpers in drivers/dma/dmaengine.h are there to _assist_
> the driver writer - they can't do magic.  dma_cookie_status() will
> return from the point of view of the generic DMA code what the status
> of a particular cookie is, and the cookie state.  It doesn't take
> care of whether a particular transaction associated with a cookie
> failed or not - that's up to the driver.
> 
> So, if dma_cookie_status() says that a cookie has DMA_COMPLETED
> status, and the DMA engine is able to detect errors on individual
> transfers, then the driver needs to do further status lookup to
> determine whether the particular transaction referred to by the
> cookie did fail, and modify the returned status appropriately.
> 
> If dma_cookie_status() says that the cookie is DMA_IN_PROGRESS,
> then the driver is expected to calculate and report the residue
> (the remaining number of bytes) of the referred to transaction.
> 

This part is fine. I'm worried about transactions that are failing.



-- 
Sinan Kaya
Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.