On Wed, Oct 16, 2024 at 10:23:14AM -0400, Waiman Long wrote: > Another alternative that I have been thinking about is a down_read() variant > with intention to upgrade later. This will ensure that only one active > reader is allowed to upgrade later. With this, upgrade_read() will always > succeed, maybe with some sleeping, as long as the correct down_read() is > used. At least for the XFS use case where direct I/O takes a share lock that needs to be replaced with an exclusive one for certain kinds of I/O would be useless. But then again we've survived without this operation for a long time, despite the initial port bringing one over from IRIX.
