JFFS2 & NAND failure

[David Woodhouse]

On Fri, 2004-11-19 at 17:22 +0100, Estelle HAMMACHE wrote:
> I feel that writing wbuf immediately adds code without necessity
> (or maybe I'm just too lazy to test it).

The code is there already; it's just a case of making it use wbuf
instead of buf. That shouldn't be so hard.

> Maybe I ought to test nextblock == 0 in addition to the
> obsolete flag to make the exclusion case more precise in
> jffs2_add_physical_node_ref ?

Ah, OK. Yes, that might be appropriate.

> Or is the "dirty" node really necessary ? what happens if we
> don't list it in the raw node refs ?

We'll try writing to that same offset again, and will fail.

-- 
dwmw2