[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: JFFS2: powerfailtesting again



good point on file names - for historical reasons, we are
using dos lenght file names, did not try with (very) long names!
that could dramatically increase overhead, on JFFS1 anyway (kinda
like java).

re Vpp, are you saying that the low batter situation will cause
constant failures because each reboot will cause a flash write and
hence another failure?  If so, might not be so bad - if you are 
line powered, the it's not likely you'd get so many power failures 
happening in this worse-case scenario.  If you are battery powered,
you will normally have low battery detection that can be set to 
prevent this kind of thing from happening (modern battery management
will prevent the unit from running the charge down below a certain
level to maximize battery life.



-----Original Message-----
From: David Woodhouse [mailto:dwmw2@xxxxxxx.org]
Sent: Wednesday, June 06, 2001 11:39 PM
To: Lance Nakamura
Cc: 'Vladimir Serov'; jffs-dev@xxxxxxx.com
Subject: Re: JFFS2: powerfailtesting again 



LANCE_N@xxxxxxx.COM said:
>  we arrived at roughly 2.5 sectors for JFFS1 empircally  - it's hard
> to tell from the documentation what it really needs to be  beyond 2,
> and agree it is not something easy to prove formally.

It's not bounded if you rename files to larger names. Actually a horrible
thought occurred to me on the way into the office this morning. 

The Linux flash drivers assert Vpp just before writing to the flash. 
Consider a situation where you have a low battery, and few milliseconds after 
you assert Vpp, the box dies. It reboots OK, right up to the time when it 
next asserts Vpp.

JFFS2 was asserting Vpp because it was writing to the flash. And it managed 
to write something (even one bit) before it died.

Watch it eat _all_ your reserved free space.

--
dwmw2


To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo@xxxxxxx.com