tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Proposal: B_ARRIER (addresses wapbl performance?)



On Tue, Dec 09, 2008 at 12:12:32PM -0800, Jason Thorpe wrote:
> 
> On Dec 9, 2008, at 11:00 AM, Manuel Bouyer wrote:
> 
> >But I don't get why we want to have the journal on stable storage  
> >*now*.
> >>From what I understand, what we want is to have the journal entry for
> >this transaction to stable storage *before* the metadata changes  
> >start to hit
> >stable storage, and the journal cleanup hits stable storage *after*
> >all metadata changes are on stable storage. I can't see what FUA  
> >brings
> >us here, as long as writes to stable storage are properly ordered.
> 
> You want your journal to remain self-consistent, otherwise you can't  
> trust it to replay it.

Sure. But if the write occurs in the right order, the journal will
stay self-consistent isn't it ? even if part of the writes are on
disk and part of them are in cache. Otherwise, even with FUA you'll
have an inconsistent journal between 2 writes.

> 
> Also, without explicit cache management, you can't be sure when the  
> data gets written out of the drive's cache.

But your "when" is only relative to other queued writes, isn't it ?
the software doesn't care if it happened at this time or not, it just
wants it to happen in order. In my view, ordered tag guarantees this,
while FUA does not.

> Again, command completion  
> has nothing to do with how the writes are ordered to the oxide.

Sure. And you don't care much about command completion either, as
long as the write to disk happen in order.

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--


Home | Main Index | Thread Index | Old Index