Re: [PATCH 2/2] block: create ioctl to discard-or-zeroout a range of blocks

[Martin K. Petersen]

>>>>> "Linus" == Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> writes:

Linus> On Thu, Mar 3, 2016 at 9:02 AM, Theodore Ts'o <tytso@xxxxxxx> wrote:
>> 
>> There is a massive bug in the SATA specs about trim, which is that it
>> is considered advisory.  So the storage device can throw it away
>> whenever it feels like it.  (In practice, when it's too busy doing
>> other things).

Linus> Ugh.

SCSI UNMAP provides similar semantics :(

Linus> But that essentially says that we shouldn't expose this interface
Linus> at all (unless we trust our white-lists - I'm sure they are
Linus> getting better, but if nobody has ever really _relied_ on the
Linus> zeroing behavior of trim, then I guess there could be tons of
Linus> bugs lurking).

We started out with the blacklist approach and it blew up. So now we're
down to a whitelist of drives that have been sanctioned by their
manufacturers to do the right thing.

Occasionally a new SSD model messes things up but we haven't updated it
since last summer. The drive vendors are much better at testing using
Linux than they used to be.

-- 
Martin K. Petersen	Oracle Linux Engineering