Sorry, no. That considerably lessens the probability of a corruption,
but there still is a critical time between when you call flush() and
when the data is safely written to the CF memory.
You can mount the drive as synchronous (mount -o sync) which prevents
linux from buffering the data. Same argument applies: it lessens the
probability of corruption but doesn't eliminate it.
gvb
At 05:02 AM 9/27/00 -0700, peter johnson wrote:
>Puh-leez!
>Use flush(2) after each write of critical data and the imagined disaster
>will not occur. I wonder where "decision makers" get their
>disinformation???
>
>"Tkatchouk, Pavel" wrote:
>
> > Ralph,
> >
> > >As for actual CF cards, I've used
> >
> > <snip>
> >
> > What fs would you recommend for CF? My employer hesitates
> > to move from DOS, guessing ext2 would be more vulnerable to
> > data corruption and power cycles.
> >
> > Pavel.
--
To unsubscribe from this list, send a message to [EMAIL PROTECTED]
with the command "unsubscribe linux-embedded" in the message body.
For more information, see <http://waste.org/mail/linux-embedded>.
