Btrfs Ready For Linux 5.6 With Async Discard For Better Efficiency + Performance

[phoronix]

Phoronix: Btrfs Ready For Linux 5.6 With Async Discard For Better Efficiency + Performance

Btrfs in the now-stable Linux 5.5 kernel is exciting for its new RAID1C3/RAID1C4 capability allowing three/four copies of data rather than just two while looking ahead to Linux 5.6 is further feature work on this Linux file-system...

http://www.phoronix.com/scan.php?page=news_item&px=Btrfs-Changes-For-Linux-5.6

[milkylainen]

I don't get why TRIM was a non-queued command to begin with. Or why anything needs to wait for it at all.
Also. Isn't queued TRIM buggy as hell?

[S.Pam]

I think this is Queued inside Btrfs, instead of the disk queue.

[intelfx]

I don't get why TRIM was a non-queued command to begin with. Or why anything needs to wait for it at all.
Also. Isn't queued TRIM buggy as hell?

As I said in the previous thread on this matter — async discard != queued TRIM. Async discard is not a protocol, but a purely algorithmic change within Btrfs.