This is the mail archive of the ecos-devel@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [ECOS] load data from floppy


Jonathan Larmour <jifl@eCosCentric.com> writes:

> Gary Thomas wrote:
> > I think your idea of reasonable sized sectors as the basic
> > granularity
> > is the right approach.  (512 bytes is OK, but maybe this should be
> > configurable)  We also need to think about all the rest of the
> > "driver"
> > stuff like initialization, control, and maybe [shudder] a block cache.
> 
> Some FS implementations rely on it for sane performance (as in
> performance isn't just slow but orders of magnitude slow because they
> rely on ability to instantly read back what was written) so we'll have
> to support it. But I'm sure it would be possible to have a "null"
> layer to make it optional.
> 
> And similarly there's write buffering.

Caching should probably be the responsibility of the individual
filesystems. Trying to create a generic cache layer to suit every
filesystem is a big job. Just look at what they have in Linux or BSD
for this. And they have the luxury of having all the filesystems
written from scratch. Filesystems ported from elsewhere will already
have their own cache interface, which won't match ours. Also, as an
embedded system, we want to control the memory usage of individual
filesystems -- trading performance against footprint on a per-instance
basis.

-- 
Nick Garnett                    eCos Kernel Architect
http://www.ecoscentric.com      The eCos and RedBoot experts


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]