On Wed, May 21, 2025 at 03:41:38PM +0100, Robin Murphy wrote: > On 2025-05-20 11:39 pm, Luis Chamberlain wrote: > > Move statistics printing to its own routine, and while at it, put > > the test counters into the struct dmatest_thread for the streaming DMA > > API to allow us to later add IOVA DMA API support and be able to > > differentiate. > > > > While at it, use a mutex to serialize output so we don't get garbled > > messages between different threads. > > > > This makes no functional changes other than serializing the output > > and prepping us for IOVA DMA API support. > > Um, what about subtly changing the test timing and total runtime > calculation, and significantly changing the output format enough to almost > certainly break any scripts parsing it? What definition of "functional" are > we using here, exactly? :/ Sure, we can keep the old format if that is the preference. > Also the > mutex doesn't prevent *other* kernel messages from being interspersed, so > multi-line output still isn't really stable. *other* sure -- but for this test it makes things legible, otherwise its quite re-ordered. Luis
