[Suspend2-devel] Re: disk consistency

[Johannes Berg]

Jens Gustedt wrote:
> But here come the semantics of `mount read-only' into play. We see two
> different interpretations of what that should be. Ext3, as you say,
> replays the journal and touches the device. Ext2 leaves the device
> untouched.

I don't even know if it does.

> BTW, I think the third option I mentioned would be much, much saner:
>
> o mount the fs strictly read-only as ext2 does
> o then replay the journal *in memory* to give the user acces to the
>   state he thought he left the device.
> o replay the journal for real to the device, only when remounting
>   read-write

 o exhaust all your available memory on a given size and dirty filesystem.

Seriously, this mailing list isn't the place to discuss file system
read-only mount semantics. Due to it being the way it is (and I believe
there are good reasons for that) you cannot mount your filesystem
read-only the way you want to, and hence can't mount it at all.

johannes