Return true from extend_thin_meta_device less #3648
Conversation
| @@ -1117,7 +1117,10 @@ impl ThinPool { | |||
| meta_growth, | |||
| ); | |||
|
|
|||
| (ext.is_ok(), ext) | |||
| match ext { | |||
| Ok(Sectors(0)) | Err(_) => (false, ext), | |||
There was a problem hiding this comment.
@jbaublitz You might be misreading the diff.
There was a problem hiding this comment.
Ah I see, so this is an optimization to not save the metadata if it's not extended at all? I am failing to see a case where this will ever be reached. Unreachable might be the wrong term, because it could be reached, but meta_growth is guaranteed to be larger than 0. I don't see a case where this will be reached with the current implementation. I was originally thinking that ext could be 0 but I don't see a case where meta_growth would be non-zero but ext would be zero.
There was a problem hiding this comment.
Let me dig into do_extend and see if maybe it's rounding down. That could potentially cause the behavior you're seeing.
There was a problem hiding this comment.
I understand your argument. I think we can leave this in pending for now and work on simplifying it later.
Return true only if the method returned without error AND the amount to extend is non-zero. This is the same condition already used in extend_thin_data_device. Signed-off-by: mulhern <amulhern@redhat.com>
mulkieran
force-pushed
the
avoid-changed-if-0-extension-meta
branch
from
c75bd89
to
b7aa992
Compare
alloc() return None if there is not enough space to satisfy the request. Signed-off-by: mulhern <amulhern@redhat.com>
|
Is this still relevant or can it be closed? |
It may be. I need to take another look. |
No description provided.