> > The netfs would then be permitted to ask fscache to record or remove a > > writeback mark against a file. > > Does this have to have file granurality? Is pinning individual pages > not possible? The plan is to have records containing: uint64_t start; uint64_t end; uint32_t ino; uint8_t flags; /* including pinning bit */ uint8_t netfs_data[11]; Multiple records can be applied to a data file. The file inode will acquire counts of applied writeback records and applied pins. > > Actually, the reverse approach might be easier - allowing the > > swapper to make use of slack space in cachefs. > > That sounds good too. But optimally swap would have a higher > priority, so for example cached data should be thrown away when the > OOM killer would be invoked otherwise. Yes. David
