On Wed, Feb 19, 2025 at 11:24:43AM +0800, Ming Lei wrote: > On Tue, Feb 18, 2025 at 05:29:53PM +0100, Christoph Hellwig wrote: > > On Tue, Feb 18, 2025 at 09:45:02PM +0800, Ming Lei wrote: > > > IMO, this RO attributes needn't protection from q->limits_lock: > > > > > > - no lifetime issue > > > > > > - in-tree code needn't limits_lock. > > > > > > - all are scalar variable, so the attribute itself is updated atomically > > > > Except in the memory model they aren't without READ_ONCE/WRITE_ONCE. > > RW_ONCE is supposed for avoiding compiler optimization, and scalar > variable atomic update should be decided by hardware. Nothing force the compiler to update atomically without that. Yes, it is very unlikely to get thing wrong, but I'd rather be explicit. And make sure static and dynamic analysis knows what we are doing here.
