On Mon 09-01-17 12:04:31, Nauman Rafique via Lsf-pc wrote: > I manage one of the storage teams at Google working on SSDs, iSCSI, SMR > etc. I have worked on cgroup support on block layer in the past, and whole > lot on different layers of SSD stack (drivers, userland daemons, etc). > > The topis that I am interested in include > - use of SSDs as slow memory devices. Do we have the right kind of > interfaces for those exposed by the kernel. So AFAIK SSDs are not directly addressable by the CPU so you can use them as a swap but I fail to see how you'd like to use them as "slow memory"... > - high efficiency, fine grained snapshots in kernel filesystems. Should it > be a facility in VFS. Are you speaking about the userspace API for snapshots (in that direction we already have clone_file_range()) or the actual implementation? Because the implementation requires support in each individual filesystem so I'm not sure what VFS could do about it... > - Can we apply DAX to SSDs, why shouldn't we? We cannot until SSDs are directly addressable by the CPU - and then they are essentially just a persistent memory. Honza -- Jan Kara <jack@xxxxxxxx> SUSE Labs, CR -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html