On 6/12/14, 10:19 PM, JP Abgrall wrote:

On Thu, Jun 12, 2014 at 8:12 PM, Eric Sandeen [off-list ref] wrote:

quoted

On 6/12/14, 10:02 PM, JP Abgrall wrote:

quoted

On Thu, Jun 12, 2014 at 7:36 PM, Eric Sandeen [off-list ref] wrote:

quoted

You're adding a new high-level IOCTL, so let's get a bit more
visibility than just linux-ext4; linux-fsdevel cc'd.

Wasn't too sure about that one, because I didn't feel like committing
to other FSes for now even if they do have FITRIM.
Will remember to include them.

You don't need to implement it for other filesystems, but whether or
not this should be a new fs-level ioctl is something worth airing
out on linux-fsdevel.

 FITRIM/SFITRIM
vs
 FITRIM/EXT4_IOC_SFITRIM ?

I don't mind either way.

Well, no, I'm not talking about making it another ext4-only implementation
- I'm talking about whether or not it's a good idea at all;
see for example Dave's reply.

If it's deemed a good idea, then yes, define SFITRIM, and implement
it for ext4, and let others follow.

If it's not a good idea then... well, perhaps don't do it at all.

This gets back to the "why" - what do you want to accomplish?
When would a user use this?  What if they have mounted -o discard,
or already issued an FITRIM?

What's the ultimate goal, here?

-Eric