Re: [PATCH 0/5] fuse: close file synchronously

[Maxim Patlasov]

On 4/17/13 1:53 PM, Miklos Szeredi wrote:
> On Mon, Apr 15, 2013 at 5:30 PM, Miklos Szeredi<miklos@xxxxxxxxxx>  wrote:
> > Let's approach it from the other direction:  what if you give back the
> > write lease on the first flush?  It will probably work fine for 99% of
> > cases, since no other writes are going to happen after the first
> > flush.  For the remaining cases you'll just have to reacquire the
> > lease when a write happens after the flush.  I guess performance-wise
> > that will not be an issue, but I may be wrong.
> What about this?

We'd like to do it, but we can't. Firstly because we rely on the fact 
that the file cannot be modified by someone else while we hold exclusive 
write lease. By the time we decide to reacquire the lease, the file can 
be re-used by someone else and become completely different comparatively 
with its state at the moment of first flush. Secondly, we can't sensibly 
handle a case when the lease is already acquired by someone else by the 
time of attempt to reacquire it.

Thanks,
Maxim
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/