Re: soft-update/async write file systems

Miguel de Icaza (miguel@nuclecu.unam.mx)
Tue, 24 Feb 1998 17:59:09 -0600


> However, doesn't this have the problem that while the buffer is blocked,
> other processes that need to access that want to read from that buffer
> will block until the buffer head is unlocked?

Ok: small hack: do not lock the buffer, set another flag that will
indicate that the block is just "held". That fixes this problem.

The transaction routines already detect this scenario and they call a
transaction-merge routine (which is not coded, been busy with the
gnome-beta-1 release :-).

> Or if you're not letting the original filesystem transaction
> complete until the log records are flushed to disk, that would seem
> to have some really nasty performance problems.

those will have to be measured once the implementation is finished.
At that point we can work on other work arounds. But the thing is:
implementing any ordering constraints that might be required using
this hack should be enough to build the foundation for the logging
file system extension. If later the block device layer gets revamped
to provide nicer featuers, we can use those.

Miguel.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu