> I think we should use jffs (without compression) on /etc and
> /home as soon as possible. CRAMFS is a pain, but allows /usr
> and the rest to live compressed right now...Experimenting
> with /etc /home changes is a nightmare right now...JFFS
> w/compression for /usr and the rest should happen as soon as
> possible....
Could you test JFFS and tell us if is stable?? I will not have time until
Monday. I think that I could find 1MB for /home and /etc and still use the
existing cramfs system.
>
> I compiled busybox for the ipaq a while ago...I think I still
> have the binaries on my system....In the end I think it saved
> something like 400K...But that was on ext2 and I didn't do
> the loopback/copy trick to reclaim unused blocks...So it
> might save more (esp. under cramfs)...BTW: Does cramfs
> support links? I hope so...That is the main way busybox
> works....One exe...the rest are links...
>
There are lots of places that we could save space: For example use ash
instead of bash, etc... Do we have any volenteers to give it a try and run
some lengthy test??>
> Also remember, 16 MB of flash is not a lot of space overall
> and not everyone will run out to drop $40 on a CF sleeve plus
> $300+ for a decently sized CF device (Flash or HD) just to
> store stuff...I personally am more interested in getting a
> PCMCIA sleeve (which I am waiting for) and CDPD and 802.11 working....
>
> In the long run I think we need a combination jffs
> (w/compression) and dynamically sizeable ramdisk...BTW: This
> seems to be what CE does...All of the user documents are
> stored in a ramdisk...If you fully remove power on your CE
> device you will loose all of your files....We need the
> ramdisk to be dynamically sizeable because you want to give
> as much RAM as possible to running tasks, while allowing the
> user to use what ram they need for storage (Up to a limit of
> course)....Under CE you do have some control over the maximum
> sizes of RAM you will give to the system and to storage...I'd
> love to see that duplicated under linux...I think ramfs may
> give us some of that...But I'm not sure...
>
RAMFS is a dynamically sizeable ramdisk. I implemented it last week, it is
compiled into the v0.13 zImage. I have not had time for testing.
Volenteers????
> I'd love to impliment this stuff...But as I stated
> earlier...I am not a C guy....
> (Which I'm starting to think I should change...But in any
> case...It would be a while before I could contribute code to
> things like this)
>
No C experience need to test RAMFS.
Best Regards,
--George
George France, france@crl.dec.com
Cambridge Research Laboratory, Compaq Computer Corporation
One Cambridgecenter MS: CRL
Cambridge, MA 02142 USA
Phone: (617) 551-7665
FAX: (617) 551-7650
Received on Wed Sep 6 10:46:43 2000
This archive was generated by hypermail 2.1.8 : Tue May 04 2004 - 09:43:41 EDT