Le 11/03/2020 à 22:18, Kamil Rytarowski a écrit : > On 11.03.2020 21:44, Chavdar Ivanov wrote: >>> The test program below shows the difference. On NetBSD-9 you have many >>> "resched", as expected. On NetBSD-current you have none. >> As you say, I don't get any output from your program on -current from >> yesterday. >> > > We definitely need ATF tests as this is a sensitive code. > > As a fallback it's possible to test HAXM.
I haven't looked closely at how HAXM handles scheduling, but if it's not using SPCF_SHOULDYIELD, then it probably has some bigger scheduling problem. Maxime
