On 10/05/2016 17:38, Alex Bligh wrote: > > and are at the > > mercy of however the kernel currently decides to split large requests). > > I am surprised TRIM doesn't get broken up the same way READ and WRITE > do.
The payload of TRIM has constant size, so it makes sense to do that. The kernel then self-imposes a 2GB limit in blkdev_issue_discard. On the other hand, READ and WRITE of size n can possibly require O(n) memory. Paolo
