On 5/17/2016 12:46 PM, Dan Williams wrote: > On Tue, May 17, 2016 at 9:28 AM, Vinod Koul <vinod.koul@xxxxxxxxx> wrote: >> On Mon, May 16, 2016 at 11:29:24AM -0400, Sinan Kaya wrote: >>> I realize that when I run DMA test kernel module, the performance numbers reported are not quite >>> accurate. The code is adding the compare and fill time to the IO performance achieved. I can see >>> the real numbers when I choose noverify option. >>> >>> Isn't this a bug? >> >> My suspicion is that it is a feature :) >>> >>> Why would you report a performance number including the source fill and destination compare? >> >> On this, I think I agree with you, Dan? > > It was just easier, i.e. deliberate laziness ;-), to not have a > special case. Yes, the performance numbers are only realistic in the > noverify case. I wouldn't say no to a patch that removes the > performance numbers in the !noverify case, but I personally don't see > the point... > I have testers and customers opening a bug case to me saying that we are not reaching to performance numbers advertised. I have to carefully explain them that the test is not correct and they are resisting. I want to remove this ambiguity. -- Sinan Kaya Qualcomm Technologies, Inc. on behalf of Qualcomm Innovation Center, Inc. Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project -- To unsubscribe from this list: send the line "unsubscribe dmaengine" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
