On Tue, Aug 16, 2022 at 03:28:50PM +0900, Hector Martin wrote:
This is the same reason I argued queue_work() itself needs to have a similar guarantee, even when it doesn't queue work (and I updated the doc to match). If test_and_set_bit() is used in this kind of context often in the kernel, clearly the current implementation/doc clashes with that.
Kernel code all over the place rely on the fact that test_and_set_bit provides a memory barrier. So this bug that you've discovered is not at all isolated to the workqeueue system. It'll break the kernel in lots of places in exactly the same way.
As I said, I don't have any particular beef in this fight, but this is horribly broken on M1/2 right now, so I'll send a patch to change the bitops instead and you all can fight it out over which way is correct :)
Please do.
Thanks,