lkml.org 
[lkml]   [1999]   [Nov]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Linux Buffer Cache Does Not Support Mirroring

Gerard,

The Statement relates to increased parallelism on FS code paths (they
don't have to block), and allows multiple I/O requests to be handled as
a "bundle" rather than one at a time. The benefit is immediately
obvious since less time is spent calling semaphore and sync code in the
kernel (shorter code paths, less rendudant calling of sync code). There
are also benefits for remirroring since disk reqests can now be
coalesced into runs and ordered based on block number for elevator
seeking. Understand now?

Jeff

Gerard Roudier wrote:
>
> On Mon, 1 Nov 1999, Alan Cox wrote:
>
> > > separate (Hotfix and mirrroing ARE independent of NWFS even in Netware),
> > > however, there are fault tolerant dependencies that allow NWFS to handle
> > > some very complex data failure and recovery conditions not available
> > > with most other FS's. It is possible to implement without it, but then
> >
> > But can those be taught to another FS ?
> >
> > > drivers and the buffer cache. There's no obvious method of supporting
> > > async callbacks (which will increase file system performance by several
> > > hundred percent).
> >
> > This is callbacks on an I/O completion ? So you can execute some piece of fs
> > code in driver context ?
>
> Callbacks on completion may help to order. What help IO performances on
> heavily loaded system is likely to defer, since this allows coalescence
> and IO sorting at O/S level. Still I donnot follow completely Jeff's
> statements.
>
> Gérard.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:54    [W:0.872 / U:1.024 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site