Re: [linux-lvm] lvm2: zeroing free space

["Martin K. Petersen" <martin petersen oracle com>]

>>>>> "Bryn" == Bryn M Reeves <bmr redhat com> writes:

Bryn> I suppose the ideal solution would be to detect whether the
Bryn> storage supports one or the other automatically but I don't know
Bryn> that there is a way to achieve this right now (TRIM/WRITE_SAME
Bryn> possibly but I think the zero page stuff is entirely transparent).

The problem with the zero-detection stuff is that it predates all the
thin provisioning goo in SCSI and we have no way of enabling
automatically. So it would have to be a manual per-device override.

However, all the vendors I have talked to prefer to use the
standards-based way now that it's there. And that's why I removed the
"write zeroes" code path before the original TP support went
upstream. It was there in the first few iterations of patches.

If people think it's a good idea I'll be happy to wire it back up. I
have a new set of TP patches going out shortly.

-- 
Martin K. Petersen	Oracle Linux Engineering