Bless your heart, you missed the point. Most of the stuff in NT is geared
towards parallelism, not disk vs. cache access. FYI. Nice to hear from you
though. I know I'm in trouble when I get an answer from you.
Your friend,
Jeff
----- Original Message -----
From: Stephen C. Tweedie <sct@redhat.com>
To: Jeff Merkey <jmerkey@timpanogas.com>
Cc: Ingo Molnar <mingo@chiara.csoma.elte.hu>;
<linux-kernel@vger.rutgers.edu>; Stephen Tweedie <sct@redhat.com>
Sent: Wednesday, June 23, 1999 8:24 AM
Subject: Re: Windows NT File Systems and Linux File Systems
> Hi,
>
> On Tue, 22 Jun 1999 12:15:13 -0600, "Jeff Merkey"
> <jmerkey@timpanogas.com> said:
>
> > I've reviewed the code sections you are refering to. This is along the
same
> > lines to what Windows NT is doing, but not identical. From what I've
seen,
> > this could really help Linux. The next step would be to create file
runs
> > for SMP read/write.
>
> We can do that on Linux. The io_request spinlock protects the block
> device request structures, but otherwise the actual IO to/from can
> progress in parallel on multiple devices and (in 2.3) copies between
> user space and the caches can also run in parallel.
>
> > MS actually stores the file allocations in the VM Cache, and then
> > transacts them by completely bypassing the file system driver. In
> > essence, the file system driver is reduced to simply being an on-disk
> > decode agent for the VM Cache Manager.
>
> Just like Linux. The "decode agent" for existing data is the bmap()
> function. 2.3 now uses a similar fs-supplied agent for writes too,
> which encapsulates both overwrite and allocate in a transparent manner.
>
> > In the Windows NT Fenris, only certain requests ever make it into an
> > IORP request. Files are read and written directly by the VM Cache
> > Manager and the File System driver only maintains meta-data fior the
> > file system (FAT's, Dirs, etc.). This is how NTFS, FASTFAT, CDFS, and
> > FENRIS are all structured on Windows NT. WARNING!!!! There are some
> > very complex unwind cases associated with the methods used by MS in
> > their file systems. On SMP, however, their model is unequalled from
> > structly a parallelism standpoint.
>
> Linux behaves very similarly. Only requests which have to hit disk
> actually get a "struct request" generated; everything else is done in
> cache.
>
> --Stephen
>
-
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/