Detecting missing memory barriers with KCSAN

Posted Dec 10, 2021 2:02 UTC (Fri) by kpfleming (subscriber, #23250)
Parent article: Detecting missing memory barriers with KCSAN

This isn't a typo, but maybe a clarification suggestion:

"the write to x in T1() could, in fact, be visible to the rest of the system after the write to flag"

As I understand it, the write to x will *always* be visible after the write to flag, the issue is when it is not visible *before* the write to flag as it needs to be. A possible change to:

"the write to x in T1() could, in fact, not be visible to the rest of the system until after the write to flag"