The write_latency parsed for pflush from nvmemul.ini should be a int rather than a bool.
And in my testing, the delay to insert in pflush sometimes may small than 0, which will cause a infinite loop.
After fixing the bug, quartz now can emulate the clflush latency correctly.
The
write_latencyparsed for pflush from nvmemul.ini should be aintrather than abool. And in my testing, the delay to insert in pflush sometimes may small than 0, which will cause a infinite loop. After fixing the bug, quartz now can emulate the clflush latency correctly.