On Tue, 2019-02-19 at 10:07 -0500, Jeff Mahoney wrote: +AD4 That's more or less how the problem presents itself, but it's not that +AD4 we issue an fstrim on a block and it gets zeroed. That would be fine +AD4 from the file system perspective. It's that the fstrim is issued and +AD4 unrelated blocks are zeroed. Once I saw the storage layer was +AD4 responsible, I stopped digging and passed it to Hannes. How can issuing an fstrim cause unrelated blocks to be zeroed? I think the root cause needs to be identified before we continue with this patch. Thanks, Bart.
