> I'm not sure it's at all possible for us to address this in mountall safely
> (i.e., without breaking other cases). And since blkid knows the disk usage
> is ambiguous, blkid doesn't seem to be doing anything wrong here either.
Agreed -- altering mountall to try to work around mkswap's bug does seem risky.
> The only actionable bug here seems to be the mkswap one.
What about altering mountall to log the failure to activate swap partition? Then users would have something they could google.
> I'm not sure it's at all possible for us to address this in mountall safely
> (i.e., without breaking other cases). And since blkid knows the disk usage
> is ambiguous, blkid doesn't seem to be doing anything wrong here either.
Agreed -- altering mountall to try to work around mkswap's bug does seem risky.
> The only actionable bug here seems to be the mkswap one.
What about altering mountall to log the failure to activate swap partition? Then users would have something they could google.